2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* Simple NUMA memory policy for the Linux kernel.
|
|
|
|
*
|
|
|
|
* Copyright 2003,2004 Andi Kleen, SuSE Labs.
|
2005-10-30 08:16:59 +07:00
|
|
|
* (C) Copyright 2005 Christoph Lameter, Silicon Graphics, Inc.
|
2005-04-17 05:20:36 +07:00
|
|
|
* Subject to the GNU Public License, version 2.
|
|
|
|
*
|
|
|
|
* NUMA policy allows the user to give hints in which node(s) memory should
|
|
|
|
* be allocated.
|
|
|
|
*
|
|
|
|
* Support four policies per VMA and per process:
|
|
|
|
*
|
|
|
|
* The VMA policy has priority over the process policy for a page fault.
|
|
|
|
*
|
|
|
|
* interleave Allocate memory interleaved over a set of nodes,
|
|
|
|
* with normal fallback if it fails.
|
|
|
|
* For VMA based allocations this interleaves based on the
|
|
|
|
* offset into the backing object or offset into the mapping
|
|
|
|
* for anonymous memory. For process policy an process counter
|
|
|
|
* is used.
|
2005-10-30 08:16:59 +07:00
|
|
|
*
|
2005-04-17 05:20:36 +07:00
|
|
|
* bind Only allocate memory on a specific set of nodes,
|
|
|
|
* no fallback.
|
2005-10-30 08:16:59 +07:00
|
|
|
* FIXME: memory is allocated starting with the first node
|
|
|
|
* to the last. It would be better if bind would truly restrict
|
|
|
|
* the allocation to memory nodes instead
|
|
|
|
*
|
2005-04-17 05:20:36 +07:00
|
|
|
* preferred Try a specific node first before normal fallback.
|
|
|
|
* As a special case node -1 here means do the allocation
|
|
|
|
* on the local CPU. This is normally identical to default,
|
|
|
|
* but useful to set in a VMA when you have a non default
|
|
|
|
* process policy.
|
2005-10-30 08:16:59 +07:00
|
|
|
*
|
2005-04-17 05:20:36 +07:00
|
|
|
* default Allocate on the local node first, or when on a VMA
|
|
|
|
* use the process policy. This is what Linux always did
|
|
|
|
* in a NUMA aware kernel and still does by, ahem, default.
|
|
|
|
*
|
|
|
|
* The process policy is applied for most non interrupt memory allocations
|
|
|
|
* in that process' context. Interrupts ignore the policies and always
|
|
|
|
* try to allocate on the local CPU. The VMA policy is only applied for memory
|
|
|
|
* allocations for a VMA in the VM.
|
|
|
|
*
|
|
|
|
* Currently there are a few corner cases in swapping where the policy
|
|
|
|
* is not applied, but the majority should be handled. When process policy
|
|
|
|
* is used it is not remembered over swap outs/swap ins.
|
|
|
|
*
|
|
|
|
* Only the highest zone in the zone hierarchy gets policied. Allocations
|
|
|
|
* requesting a lower zone just use default policy. This implies that
|
|
|
|
* on systems with highmem kernel lowmem allocation don't get policied.
|
|
|
|
* Same with GFP_DMA allocations.
|
|
|
|
*
|
|
|
|
* For shmfs/tmpfs/hugetlbfs shared memory the policy is shared between
|
|
|
|
* all users and remembered even when nobody has memory mapped.
|
|
|
|
*/
|
|
|
|
|
|
|
|
/* Notebook:
|
|
|
|
fix mmap readahead to honour policy and enable policy for any page cache
|
|
|
|
object
|
|
|
|
statistics for bigpages
|
|
|
|
global policy for page cache? currently it uses process policy. Requires
|
|
|
|
first item above.
|
|
|
|
handle mremap for shared memory (currently ignored for the policy)
|
|
|
|
grows down?
|
|
|
|
make bind policy root only? It can trigger oom much faster and the
|
|
|
|
kernel is not always grateful with that.
|
|
|
|
could replace all the switch()es with a mempolicy_ops structure.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <linux/mempolicy.h>
|
|
|
|
#include <linux/mm.h>
|
|
|
|
#include <linux/highmem.h>
|
|
|
|
#include <linux/hugetlb.h>
|
|
|
|
#include <linux/kernel.h>
|
|
|
|
#include <linux/sched.h>
|
|
|
|
#include <linux/nodemask.h>
|
|
|
|
#include <linux/cpuset.h>
|
|
|
|
#include <linux/gfp.h>
|
|
|
|
#include <linux/slab.h>
|
|
|
|
#include <linux/string.h>
|
|
|
|
#include <linux/module.h>
|
2007-10-19 13:40:14 +07:00
|
|
|
#include <linux/nsproxy.h>
|
2005-04-17 05:20:36 +07:00
|
|
|
#include <linux/interrupt.h>
|
|
|
|
#include <linux/init.h>
|
|
|
|
#include <linux/compat.h>
|
2006-01-08 16:00:50 +07:00
|
|
|
#include <linux/swap.h>
|
2006-01-08 16:01:02 +07:00
|
|
|
#include <linux/seq_file.h>
|
|
|
|
#include <linux/proc_fs.h>
|
2006-03-22 15:09:12 +07:00
|
|
|
#include <linux/migrate.h>
|
2006-06-23 16:03:53 +07:00
|
|
|
#include <linux/rmap.h>
|
2006-06-23 16:04:02 +07:00
|
|
|
#include <linux/security.h>
|
2007-10-16 15:26:26 +07:00
|
|
|
#include <linux/syscalls.h>
|
2006-01-08 16:00:50 +07:00
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
#include <asm/tlbflush.h>
|
|
|
|
#include <asm/uaccess.h>
|
|
|
|
|
2006-01-08 16:01:01 +07:00
|
|
|
/* Internal flags */
|
2006-01-08 16:00:50 +07:00
|
|
|
#define MPOL_MF_DISCONTIG_OK (MPOL_MF_INTERNAL << 0) /* Skip checks for continuous vmas */
|
2006-01-08 16:01:01 +07:00
|
|
|
#define MPOL_MF_INVERT (MPOL_MF_INTERNAL << 1) /* Invert check for nodemask */
|
2006-01-08 16:01:02 +07:00
|
|
|
#define MPOL_MF_STATS (MPOL_MF_INTERNAL << 2) /* Gather statistics */
|
2006-01-08 16:00:50 +07:00
|
|
|
|
2006-03-22 15:08:13 +07:00
|
|
|
static struct kmem_cache *policy_cache;
|
|
|
|
static struct kmem_cache *sn_cache;
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
/* Highest zone. An specific allocation for a zone below that is not
|
|
|
|
policied. */
|
2007-02-10 16:43:07 +07:00
|
|
|
enum zone_type policy_zone = 0;
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2005-07-07 00:56:03 +07:00
|
|
|
struct mempolicy default_policy = {
|
2005-04-17 05:20:36 +07:00
|
|
|
.refcnt = ATOMIC_INIT(1), /* never free it */
|
|
|
|
.policy = MPOL_DEFAULT,
|
|
|
|
};
|
|
|
|
|
2007-10-16 15:26:26 +07:00
|
|
|
static void mpol_rebind_policy(struct mempolicy *pol,
|
|
|
|
const nodemask_t *newmask);
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/* Do sanity checking on a policy */
|
2005-10-30 08:15:48 +07:00
|
|
|
static int mpol_check_policy(int mode, nodemask_t *nodes)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
2005-10-30 08:15:48 +07:00
|
|
|
int empty = nodes_empty(*nodes);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
switch (mode) {
|
|
|
|
case MPOL_DEFAULT:
|
|
|
|
if (!empty)
|
|
|
|
return -EINVAL;
|
|
|
|
break;
|
|
|
|
case MPOL_BIND:
|
|
|
|
case MPOL_INTERLEAVE:
|
|
|
|
/* Preferred will only use the first bit, but allow
|
|
|
|
more for now. */
|
|
|
|
if (empty)
|
|
|
|
return -EINVAL;
|
|
|
|
break;
|
|
|
|
}
|
2007-10-16 15:25:39 +07:00
|
|
|
return nodes_subset(*nodes, node_states[N_HIGH_MEMORY]) ? 0 : -EINVAL;
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
2006-02-17 07:39:16 +07:00
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/* Generate a custom zonelist for the BIND policy. */
|
2005-10-30 08:15:48 +07:00
|
|
|
static struct zonelist *bind_zonelist(nodemask_t *nodes)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
struct zonelist *zl;
|
2006-09-26 13:31:18 +07:00
|
|
|
int num, max, nd;
|
|
|
|
enum zone_type k;
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2005-10-30 08:15:48 +07:00
|
|
|
max = 1 + MAX_NR_ZONES * nodes_weight(*nodes);
|
[PATCH] memory page_alloc zonelist caching speedup
Optimize the critical zonelist scanning for free pages in the kernel memory
allocator by caching the zones that were found to be full recently, and
skipping them.
Remembers the zones in a zonelist that were short of free memory in the
last second. And it stashes a zone-to-node table in the zonelist struct,
to optimize that conversion (minimize its cache footprint.)
Recent changes:
This differs in a significant way from a similar patch that I
posted a week ago. Now, instead of having a nodemask_t of
recently full nodes, I have a bitmask of recently full zones.
This solves a problem that last weeks patch had, which on
systems with multiple zones per node (such as DMA zone) would
take seeing any of these zones full as meaning that all zones
on that node were full.
Also I changed names - from "zonelist faster" to "zonelist cache",
as that seemed to better convey what we're doing here - caching
some of the key zonelist state (for faster access.)
See below for some performance benchmark results. After all that
discussion with David on why I didn't need them, I went and got
some ;). I wanted to verify that I had not hurt the normal case
of memory allocation noticeably. At least for my one little
microbenchmark, I found (1) the normal case wasn't affected, and
(2) workloads that forced scanning across multiple nodes for
memory improved up to 10% fewer System CPU cycles and lower
elapsed clock time ('sys' and 'real'). Good. See details, below.
I didn't have the logic in get_page_from_freelist() for various
full nodes and zone reclaim failures correct. That should be
fixed up now - notice the new goto labels zonelist_scan,
this_zone_full, and try_next_zone, in get_page_from_freelist().
There are two reasons I persued this alternative, over some earlier
proposals that would have focused on optimizing the fake numa
emulation case by caching the last useful zone:
1) Contrary to what I said before, we (SGI, on large ia64 sn2 systems)
have seen real customer loads where the cost to scan the zonelist
was a problem, due to many nodes being full of memory before
we got to a node we could use. Or at least, I think we have.
This was related to me by another engineer, based on experiences
from some time past. So this is not guaranteed. Most likely, though.
The following approach should help such real numa systems just as
much as it helps fake numa systems, or any combination thereof.
2) The effort to distinguish fake from real numa, using node_distance,
so that we could cache a fake numa node and optimize choosing
it over equivalent distance fake nodes, while continuing to
properly scan all real nodes in distance order, was going to
require a nasty blob of zonelist and node distance munging.
The following approach has no new dependency on node distances or
zone sorting.
See comment in the patch below for a description of what it actually does.
Technical details of note (or controversy):
- See the use of "zlc_active" and "did_zlc_setup" below, to delay
adding any work for this new mechanism until we've looked at the
first zone in zonelist. I figured the odds of the first zone
having the memory we needed were high enough that we should just
look there, first, then get fancy only if we need to keep looking.
- Some odd hackery was needed to add items to struct zonelist, while
not tripping up the custom zonelists built by the mm/mempolicy.c
code for MPOL_BIND. My usual wordy comments below explain this.
Search for "MPOL_BIND".
- Some per-node data in the struct zonelist is now modified frequently,
with no locking. Multiple CPU cores on a node could hit and mangle
this data. The theory is that this is just performance hint data,
and the memory allocator will work just fine despite any such mangling.
The fields at risk are the struct 'zonelist_cache' fields 'fullzones'
(a bitmask) and 'last_full_zap' (unsigned long jiffies). It should
all be self correcting after at most a one second delay.
- This still does a linear scan of the same lengths as before. All
I've optimized is making the scan faster, not algorithmically
shorter. It is now able to scan a compact array of 'unsigned
short' in the case of many full nodes, so one cache line should
cover quite a few nodes, rather than each node hitting another
one or two new and distinct cache lines.
- If both Andi and Nick don't find this too complicated, I will be
(pleasantly) flabbergasted.
- I removed the comment claiming we only use one cachline's worth of
zonelist. We seem, at least in the fake numa case, to have put the
lie to that claim.
- I pay no attention to the various watermarks and such in this performance
hint. A node could be marked full for one watermark, and then skipped
over when searching for a page using a different watermark. I think
that's actually quite ok, as it will tend to slightly increase the
spreading of memory over other nodes, away from a memory stressed node.
===============
Performance - some benchmark results and analysis:
This benchmark runs a memory hog program that uses multiple
threads to touch alot of memory as quickly as it can.
Multiple runs were made, touching 12, 38, 64 or 90 GBytes out of
the total 96 GBytes on the system, and using 1, 19, 37, or 55
threads (on a 56 CPU system.) System, user and real (elapsed)
timings were recorded for each run, shown in units of seconds,
in the table below.
Two kernels were tested - 2.6.18-mm3 and the same kernel with
this zonelist caching patch added. The table also shows the
percentage improvement the zonelist caching sys time is over
(lower than) the stock *-mm kernel.
number 2.6.18-mm3 zonelist-cache delta (< 0 good) percent
GBs N ------------ -------------- ---------------- systime
mem threads sys user real sys user real sys user real better
12 1 153 24 177 151 24 176 -2 0 -1 1%
12 19 99 22 8 99 22 8 0 0 0 0%
12 37 111 25 6 112 25 6 1 0 0 -0%
12 55 115 25 5 110 23 5 -5 -2 0 4%
38 1 502 74 576 497 73 570 -5 -1 -6 0%
38 19 426 78 48 373 76 39 -53 -2 -9 12%
38 37 544 83 36 547 82 36 3 -1 0 -0%
38 55 501 77 23 511 80 24 10 3 1 -1%
64 1 917 125 1042 890 124 1014 -27 -1 -28 2%
64 19 1118 138 119 965 141 103 -153 3 -16 13%
64 37 1202 151 94 1136 150 81 -66 -1 -13 5%
64 55 1118 141 61 1072 140 58 -46 -1 -3 4%
90 1 1342 177 1519 1275 174 1450 -67 -3 -69 4%
90 19 2392 199 192 2116 189 176 -276 -10 -16 11%
90 37 3313 238 175 2972 225 145 -341 -13 -30 10%
90 55 1948 210 104 1843 213 100 -105 3 -4 5%
Notes:
1) This test ran a memory hog program that started a specified number N of
threads, and had each thread allocate and touch 1/N'th of
the total memory to be used in the test run in a single loop,
writing a constant word to memory, one store every 4096 bytes.
Watching this test during some earlier trial runs, I would see
each of these threads sit down on one CPU and stay there, for
the remainder of the pass, a different CPU for each thread.
2) The 'real' column is not comparable to the 'sys' or 'user' columns.
The 'real' column is seconds wall clock time elapsed, from beginning
to end of that test pass. The 'sys' and 'user' columns are total
CPU seconds spent on that test pass. For a 19 thread test run,
for example, the sum of 'sys' and 'user' could be up to 19 times the
number of 'real' elapsed wall clock seconds.
3) Tests were run on a fresh, single-user boot, to minimize the amount
of memory already in use at the start of the test, and to minimize
the amount of background activity that might interfere.
4) Tests were done on a 56 CPU, 28 Node system with 96 GBytes of RAM.
5) Notice that the 'real' time gets large for the single thread runs, even
though the measured 'sys' and 'user' times are modest. I'm not sure what
that means - probably something to do with it being slow for one thread to
be accessing memory along ways away. Perhaps the fake numa system, running
ostensibly the same workload, would not show this substantial degradation
of 'real' time for one thread on many nodes -- lets hope not.
6) The high thread count passes (one thread per CPU - on 55 of 56 CPUs)
ran quite efficiently, as one might expect. Each pair of threads needed
to allocate and touch the memory on the node the two threads shared, a
pleasantly parallizable workload.
7) The intermediate thread count passes, when asking for alot of memory forcing
them to go to a few neighboring nodes, improved the most with this zonelist
caching patch.
Conclusions:
* This zonelist cache patch probably makes little difference one way or the
other for most workloads on real numa hardware, if those workloads avoid
heavy off node allocations.
* For memory intensive workloads requiring substantial off-node allocations
on real numa hardware, this patch improves both kernel and elapsed timings
up to ten per-cent.
* For fake numa systems, I'm optimistic, but will have to leave that up to
Rohit Seth to actually test (once I get him a 2.6.18 backport.)
Signed-off-by: Paul Jackson <pj@sgi.com>
Cc: Rohit Seth <rohitseth@google.com>
Cc: Christoph Lameter <clameter@engr.sgi.com>
Cc: David Rientjes <rientjes@cs.washington.edu>
Cc: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-07 11:31:48 +07:00
|
|
|
max++; /* space for zlcache_ptr (see mmzone.h) */
|
2006-02-17 07:39:16 +07:00
|
|
|
zl = kmalloc(sizeof(struct zone *) * max, GFP_KERNEL);
|
2005-04-17 05:20:36 +07:00
|
|
|
if (!zl)
|
2007-02-21 04:57:49 +07:00
|
|
|
return ERR_PTR(-ENOMEM);
|
[PATCH] memory page_alloc zonelist caching speedup
Optimize the critical zonelist scanning for free pages in the kernel memory
allocator by caching the zones that were found to be full recently, and
skipping them.
Remembers the zones in a zonelist that were short of free memory in the
last second. And it stashes a zone-to-node table in the zonelist struct,
to optimize that conversion (minimize its cache footprint.)
Recent changes:
This differs in a significant way from a similar patch that I
posted a week ago. Now, instead of having a nodemask_t of
recently full nodes, I have a bitmask of recently full zones.
This solves a problem that last weeks patch had, which on
systems with multiple zones per node (such as DMA zone) would
take seeing any of these zones full as meaning that all zones
on that node were full.
Also I changed names - from "zonelist faster" to "zonelist cache",
as that seemed to better convey what we're doing here - caching
some of the key zonelist state (for faster access.)
See below for some performance benchmark results. After all that
discussion with David on why I didn't need them, I went and got
some ;). I wanted to verify that I had not hurt the normal case
of memory allocation noticeably. At least for my one little
microbenchmark, I found (1) the normal case wasn't affected, and
(2) workloads that forced scanning across multiple nodes for
memory improved up to 10% fewer System CPU cycles and lower
elapsed clock time ('sys' and 'real'). Good. See details, below.
I didn't have the logic in get_page_from_freelist() for various
full nodes and zone reclaim failures correct. That should be
fixed up now - notice the new goto labels zonelist_scan,
this_zone_full, and try_next_zone, in get_page_from_freelist().
There are two reasons I persued this alternative, over some earlier
proposals that would have focused on optimizing the fake numa
emulation case by caching the last useful zone:
1) Contrary to what I said before, we (SGI, on large ia64 sn2 systems)
have seen real customer loads where the cost to scan the zonelist
was a problem, due to many nodes being full of memory before
we got to a node we could use. Or at least, I think we have.
This was related to me by another engineer, based on experiences
from some time past. So this is not guaranteed. Most likely, though.
The following approach should help such real numa systems just as
much as it helps fake numa systems, or any combination thereof.
2) The effort to distinguish fake from real numa, using node_distance,
so that we could cache a fake numa node and optimize choosing
it over equivalent distance fake nodes, while continuing to
properly scan all real nodes in distance order, was going to
require a nasty blob of zonelist and node distance munging.
The following approach has no new dependency on node distances or
zone sorting.
See comment in the patch below for a description of what it actually does.
Technical details of note (or controversy):
- See the use of "zlc_active" and "did_zlc_setup" below, to delay
adding any work for this new mechanism until we've looked at the
first zone in zonelist. I figured the odds of the first zone
having the memory we needed were high enough that we should just
look there, first, then get fancy only if we need to keep looking.
- Some odd hackery was needed to add items to struct zonelist, while
not tripping up the custom zonelists built by the mm/mempolicy.c
code for MPOL_BIND. My usual wordy comments below explain this.
Search for "MPOL_BIND".
- Some per-node data in the struct zonelist is now modified frequently,
with no locking. Multiple CPU cores on a node could hit and mangle
this data. The theory is that this is just performance hint data,
and the memory allocator will work just fine despite any such mangling.
The fields at risk are the struct 'zonelist_cache' fields 'fullzones'
(a bitmask) and 'last_full_zap' (unsigned long jiffies). It should
all be self correcting after at most a one second delay.
- This still does a linear scan of the same lengths as before. All
I've optimized is making the scan faster, not algorithmically
shorter. It is now able to scan a compact array of 'unsigned
short' in the case of many full nodes, so one cache line should
cover quite a few nodes, rather than each node hitting another
one or two new and distinct cache lines.
- If both Andi and Nick don't find this too complicated, I will be
(pleasantly) flabbergasted.
- I removed the comment claiming we only use one cachline's worth of
zonelist. We seem, at least in the fake numa case, to have put the
lie to that claim.
- I pay no attention to the various watermarks and such in this performance
hint. A node could be marked full for one watermark, and then skipped
over when searching for a page using a different watermark. I think
that's actually quite ok, as it will tend to slightly increase the
spreading of memory over other nodes, away from a memory stressed node.
===============
Performance - some benchmark results and analysis:
This benchmark runs a memory hog program that uses multiple
threads to touch alot of memory as quickly as it can.
Multiple runs were made, touching 12, 38, 64 or 90 GBytes out of
the total 96 GBytes on the system, and using 1, 19, 37, or 55
threads (on a 56 CPU system.) System, user and real (elapsed)
timings were recorded for each run, shown in units of seconds,
in the table below.
Two kernels were tested - 2.6.18-mm3 and the same kernel with
this zonelist caching patch added. The table also shows the
percentage improvement the zonelist caching sys time is over
(lower than) the stock *-mm kernel.
number 2.6.18-mm3 zonelist-cache delta (< 0 good) percent
GBs N ------------ -------------- ---------------- systime
mem threads sys user real sys user real sys user real better
12 1 153 24 177 151 24 176 -2 0 -1 1%
12 19 99 22 8 99 22 8 0 0 0 0%
12 37 111 25 6 112 25 6 1 0 0 -0%
12 55 115 25 5 110 23 5 -5 -2 0 4%
38 1 502 74 576 497 73 570 -5 -1 -6 0%
38 19 426 78 48 373 76 39 -53 -2 -9 12%
38 37 544 83 36 547 82 36 3 -1 0 -0%
38 55 501 77 23 511 80 24 10 3 1 -1%
64 1 917 125 1042 890 124 1014 -27 -1 -28 2%
64 19 1118 138 119 965 141 103 -153 3 -16 13%
64 37 1202 151 94 1136 150 81 -66 -1 -13 5%
64 55 1118 141 61 1072 140 58 -46 -1 -3 4%
90 1 1342 177 1519 1275 174 1450 -67 -3 -69 4%
90 19 2392 199 192 2116 189 176 -276 -10 -16 11%
90 37 3313 238 175 2972 225 145 -341 -13 -30 10%
90 55 1948 210 104 1843 213 100 -105 3 -4 5%
Notes:
1) This test ran a memory hog program that started a specified number N of
threads, and had each thread allocate and touch 1/N'th of
the total memory to be used in the test run in a single loop,
writing a constant word to memory, one store every 4096 bytes.
Watching this test during some earlier trial runs, I would see
each of these threads sit down on one CPU and stay there, for
the remainder of the pass, a different CPU for each thread.
2) The 'real' column is not comparable to the 'sys' or 'user' columns.
The 'real' column is seconds wall clock time elapsed, from beginning
to end of that test pass. The 'sys' and 'user' columns are total
CPU seconds spent on that test pass. For a 19 thread test run,
for example, the sum of 'sys' and 'user' could be up to 19 times the
number of 'real' elapsed wall clock seconds.
3) Tests were run on a fresh, single-user boot, to minimize the amount
of memory already in use at the start of the test, and to minimize
the amount of background activity that might interfere.
4) Tests were done on a 56 CPU, 28 Node system with 96 GBytes of RAM.
5) Notice that the 'real' time gets large for the single thread runs, even
though the measured 'sys' and 'user' times are modest. I'm not sure what
that means - probably something to do with it being slow for one thread to
be accessing memory along ways away. Perhaps the fake numa system, running
ostensibly the same workload, would not show this substantial degradation
of 'real' time for one thread on many nodes -- lets hope not.
6) The high thread count passes (one thread per CPU - on 55 of 56 CPUs)
ran quite efficiently, as one might expect. Each pair of threads needed
to allocate and touch the memory on the node the two threads shared, a
pleasantly parallizable workload.
7) The intermediate thread count passes, when asking for alot of memory forcing
them to go to a few neighboring nodes, improved the most with this zonelist
caching patch.
Conclusions:
* This zonelist cache patch probably makes little difference one way or the
other for most workloads on real numa hardware, if those workloads avoid
heavy off node allocations.
* For memory intensive workloads requiring substantial off-node allocations
on real numa hardware, this patch improves both kernel and elapsed timings
up to ten per-cent.
* For fake numa systems, I'm optimistic, but will have to leave that up to
Rohit Seth to actually test (once I get him a 2.6.18 backport.)
Signed-off-by: Paul Jackson <pj@sgi.com>
Cc: Rohit Seth <rohitseth@google.com>
Cc: Christoph Lameter <clameter@engr.sgi.com>
Cc: David Rientjes <rientjes@cs.washington.edu>
Cc: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-07 11:31:48 +07:00
|
|
|
zl->zlcache_ptr = NULL;
|
2005-04-17 05:20:36 +07:00
|
|
|
num = 0;
|
2006-02-17 07:39:16 +07:00
|
|
|
/* First put in the highest zones from all nodes, then all the next
|
|
|
|
lower zones etc. Avoid empty zones because the memory allocator
|
|
|
|
doesn't like them. If you implement node hot removal you
|
|
|
|
have to fix that. */
|
2007-08-23 04:02:05 +07:00
|
|
|
k = MAX_NR_ZONES - 1;
|
2006-09-26 13:31:18 +07:00
|
|
|
while (1) {
|
2006-02-17 07:39:16 +07:00
|
|
|
for_each_node_mask(nd, *nodes) {
|
|
|
|
struct zone *z = &NODE_DATA(nd)->node_zones[k];
|
|
|
|
if (z->present_pages > 0)
|
|
|
|
zl->zones[num++] = z;
|
|
|
|
}
|
2006-09-26 13:31:18 +07:00
|
|
|
if (k == 0)
|
|
|
|
break;
|
|
|
|
k--;
|
2006-02-17 07:39:16 +07:00
|
|
|
}
|
2007-02-21 04:57:49 +07:00
|
|
|
if (num == 0) {
|
|
|
|
kfree(zl);
|
|
|
|
return ERR_PTR(-EINVAL);
|
|
|
|
}
|
2005-04-17 05:20:36 +07:00
|
|
|
zl->zones[num] = NULL;
|
|
|
|
return zl;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Create a new policy */
|
2005-10-30 08:15:48 +07:00
|
|
|
static struct mempolicy *mpol_new(int mode, nodemask_t *nodes)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
struct mempolicy *policy;
|
|
|
|
|
2007-07-16 13:38:16 +07:00
|
|
|
pr_debug("setting mode %d nodes[0] %lx\n",
|
|
|
|
mode, nodes ? nodes_addr(*nodes)[0] : -1);
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
if (mode == MPOL_DEFAULT)
|
|
|
|
return NULL;
|
|
|
|
policy = kmem_cache_alloc(policy_cache, GFP_KERNEL);
|
|
|
|
if (!policy)
|
|
|
|
return ERR_PTR(-ENOMEM);
|
|
|
|
atomic_set(&policy->refcnt, 1);
|
|
|
|
switch (mode) {
|
|
|
|
case MPOL_INTERLEAVE:
|
2005-10-30 08:15:48 +07:00
|
|
|
policy->v.nodes = *nodes;
|
2007-10-16 15:25:30 +07:00
|
|
|
nodes_and(policy->v.nodes, policy->v.nodes,
|
|
|
|
node_states[N_HIGH_MEMORY]);
|
|
|
|
if (nodes_weight(policy->v.nodes) == 0) {
|
2006-01-03 06:07:28 +07:00
|
|
|
kmem_cache_free(policy_cache, policy);
|
|
|
|
return ERR_PTR(-EINVAL);
|
|
|
|
}
|
2005-04-17 05:20:36 +07:00
|
|
|
break;
|
|
|
|
case MPOL_PREFERRED:
|
2005-10-30 08:15:48 +07:00
|
|
|
policy->v.preferred_node = first_node(*nodes);
|
2005-04-17 05:20:36 +07:00
|
|
|
if (policy->v.preferred_node >= MAX_NUMNODES)
|
|
|
|
policy->v.preferred_node = -1;
|
|
|
|
break;
|
|
|
|
case MPOL_BIND:
|
|
|
|
policy->v.zonelist = bind_zonelist(nodes);
|
2007-02-21 04:57:49 +07:00
|
|
|
if (IS_ERR(policy->v.zonelist)) {
|
|
|
|
void *error_code = policy->v.zonelist;
|
2005-04-17 05:20:36 +07:00
|
|
|
kmem_cache_free(policy_cache, policy);
|
2007-02-21 04:57:49 +07:00
|
|
|
return error_code;
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
policy->policy = mode;
|
[PATCH] cpuset: numa_policy_rebind cleanup
Cleanup, reorganize and make more robust the mempolicy.c code to rebind
mempolicies relative to the containing cpuset after a tasks memory placement
changes.
The real motivator for this cleanup patch is to lay more groundwork for the
upcoming patch to correctly rebind NUMA mempolicies that are attached to vma's
after the containing cpuset memory placement changes.
NUMA mempolicies are constrained by the cpuset their task is a member of.
When either (1) a task is moved to a different cpuset, or (2) the 'mems'
mems_allowed of a cpuset is changed, then the NUMA mempolicies have embedded
node numbers (for MPOL_BIND, MPOL_INTERLEAVE and MPOL_PREFERRED) that need to
be recalculated, relative to their new cpuset placement.
The old code used an unreliable method of determining what was the old
mems_allowed constraining the mempolicy. It just looked at the tasks
mems_allowed value. This sort of worked with the present code, that just
rebinds the -task- mempolicy, and leaves any -vma- mempolicies broken,
referring to the old nodes. But in an upcoming patch, the vma mempolicies
will be rebound as well. Then the order in which the various task and vma
mempolicies are updated will no longer be deterministic, and one can no longer
count on the task->mems_allowed holding the old value for as long as needed.
It's not even clear if the current code was guaranteed to work reliably for
task mempolicies.
So I added a mems_allowed field to each mempolicy, stating exactly what
mems_allowed the policy is relative to, and updated synchronously and reliably
anytime that the mempolicy is rebound.
Also removed a useless wrapper routine, numa_policy_rebind(), and had its
caller, cpuset_update_task_memory_state(), call directly to the rewritten
policy_rebind() routine, and made that rebind routine extern instead of
static, and added a "mpol_" prefix to its name, making it
mpol_rebind_policy().
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-08 16:01:56 +07:00
|
|
|
policy->cpuset_mems_allowed = cpuset_mems_allowed(current);
|
2005-04-17 05:20:36 +07:00
|
|
|
return policy;
|
|
|
|
}
|
|
|
|
|
2006-03-07 06:42:53 +07:00
|
|
|
static void gather_stats(struct page *, void *, int pte_dirty);
|
2006-01-19 08:42:29 +07:00
|
|
|
static void migrate_page_add(struct page *page, struct list_head *pagelist,
|
|
|
|
unsigned long flags);
|
2006-01-08 16:01:02 +07:00
|
|
|
|
2006-01-08 16:01:01 +07:00
|
|
|
/* Scan through pages checking if pages follow certain conditions. */
|
2005-10-30 08:16:12 +07:00
|
|
|
static int check_pte_range(struct vm_area_struct *vma, pmd_t *pmd,
|
2006-01-08 16:00:50 +07:00
|
|
|
unsigned long addr, unsigned long end,
|
|
|
|
const nodemask_t *nodes, unsigned long flags,
|
2006-01-08 16:01:01 +07:00
|
|
|
void *private)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
2005-06-22 07:15:07 +07:00
|
|
|
pte_t *orig_pte;
|
|
|
|
pte_t *pte;
|
2005-10-30 08:16:27 +07:00
|
|
|
spinlock_t *ptl;
|
2005-06-22 07:15:06 +07:00
|
|
|
|
2005-10-30 08:16:27 +07:00
|
|
|
orig_pte = pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
|
2005-06-22 07:15:07 +07:00
|
|
|
do {
|
2005-11-29 05:34:23 +07:00
|
|
|
struct page *page;
|
2006-12-07 11:33:03 +07:00
|
|
|
int nid;
|
2005-06-22 07:15:07 +07:00
|
|
|
|
|
|
|
if (!pte_present(*pte))
|
2005-04-17 05:20:36 +07:00
|
|
|
continue;
|
2005-11-29 05:34:23 +07:00
|
|
|
page = vm_normal_page(vma, addr, *pte);
|
|
|
|
if (!page)
|
2005-04-17 05:20:36 +07:00
|
|
|
continue;
|
2006-01-19 08:42:27 +07:00
|
|
|
/*
|
|
|
|
* The check for PageReserved here is important to avoid
|
|
|
|
* handling zero pages and other pages that may have been
|
|
|
|
* marked special by the system.
|
|
|
|
*
|
|
|
|
* If the PageReserved would not be checked here then f.e.
|
|
|
|
* the location of the zero page could have an influence
|
|
|
|
* on MPOL_MF_STRICT, zero pages would be counted for
|
|
|
|
* the per node stats, and there would be useless attempts
|
|
|
|
* to put zero pages on the migration list.
|
|
|
|
*/
|
2006-01-12 16:05:20 +07:00
|
|
|
if (PageReserved(page))
|
|
|
|
continue;
|
2005-11-29 05:34:23 +07:00
|
|
|
nid = page_to_nid(page);
|
2006-01-08 16:01:01 +07:00
|
|
|
if (node_isset(nid, *nodes) == !!(flags & MPOL_MF_INVERT))
|
|
|
|
continue;
|
|
|
|
|
2006-01-08 16:01:02 +07:00
|
|
|
if (flags & MPOL_MF_STATS)
|
2006-03-07 06:42:53 +07:00
|
|
|
gather_stats(page, private, pte_dirty(*pte));
|
2006-01-19 08:42:27 +07:00
|
|
|
else if (flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL))
|
2006-01-19 08:42:29 +07:00
|
|
|
migrate_page_add(page, private, flags);
|
2006-01-08 16:01:01 +07:00
|
|
|
else
|
|
|
|
break;
|
2005-06-22 07:15:07 +07:00
|
|
|
} while (pte++, addr += PAGE_SIZE, addr != end);
|
2005-10-30 08:16:27 +07:00
|
|
|
pte_unmap_unlock(orig_pte, ptl);
|
2005-06-22 07:15:07 +07:00
|
|
|
return addr != end;
|
|
|
|
}
|
|
|
|
|
2005-10-30 08:16:12 +07:00
|
|
|
static inline int check_pmd_range(struct vm_area_struct *vma, pud_t *pud,
|
2006-01-08 16:00:50 +07:00
|
|
|
unsigned long addr, unsigned long end,
|
|
|
|
const nodemask_t *nodes, unsigned long flags,
|
2006-01-08 16:01:01 +07:00
|
|
|
void *private)
|
2005-06-22 07:15:07 +07:00
|
|
|
{
|
|
|
|
pmd_t *pmd;
|
|
|
|
unsigned long next;
|
|
|
|
|
|
|
|
pmd = pmd_offset(pud, addr);
|
|
|
|
do {
|
|
|
|
next = pmd_addr_end(addr, end);
|
|
|
|
if (pmd_none_or_clear_bad(pmd))
|
|
|
|
continue;
|
2006-01-08 16:00:50 +07:00
|
|
|
if (check_pte_range(vma, pmd, addr, next, nodes,
|
2006-01-08 16:01:01 +07:00
|
|
|
flags, private))
|
2005-06-22 07:15:07 +07:00
|
|
|
return -EIO;
|
|
|
|
} while (pmd++, addr = next, addr != end);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2005-10-30 08:16:12 +07:00
|
|
|
static inline int check_pud_range(struct vm_area_struct *vma, pgd_t *pgd,
|
2006-01-08 16:00:50 +07:00
|
|
|
unsigned long addr, unsigned long end,
|
|
|
|
const nodemask_t *nodes, unsigned long flags,
|
2006-01-08 16:01:01 +07:00
|
|
|
void *private)
|
2005-06-22 07:15:07 +07:00
|
|
|
{
|
|
|
|
pud_t *pud;
|
|
|
|
unsigned long next;
|
|
|
|
|
|
|
|
pud = pud_offset(pgd, addr);
|
|
|
|
do {
|
|
|
|
next = pud_addr_end(addr, end);
|
|
|
|
if (pud_none_or_clear_bad(pud))
|
|
|
|
continue;
|
2006-01-08 16:00:50 +07:00
|
|
|
if (check_pmd_range(vma, pud, addr, next, nodes,
|
2006-01-08 16:01:01 +07:00
|
|
|
flags, private))
|
2005-06-22 07:15:07 +07:00
|
|
|
return -EIO;
|
|
|
|
} while (pud++, addr = next, addr != end);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2005-10-30 08:16:12 +07:00
|
|
|
static inline int check_pgd_range(struct vm_area_struct *vma,
|
2006-01-08 16:00:50 +07:00
|
|
|
unsigned long addr, unsigned long end,
|
|
|
|
const nodemask_t *nodes, unsigned long flags,
|
2006-01-08 16:01:01 +07:00
|
|
|
void *private)
|
2005-06-22 07:15:07 +07:00
|
|
|
{
|
|
|
|
pgd_t *pgd;
|
|
|
|
unsigned long next;
|
|
|
|
|
2005-10-30 08:16:12 +07:00
|
|
|
pgd = pgd_offset(vma->vm_mm, addr);
|
2005-06-22 07:15:07 +07:00
|
|
|
do {
|
|
|
|
next = pgd_addr_end(addr, end);
|
|
|
|
if (pgd_none_or_clear_bad(pgd))
|
|
|
|
continue;
|
2006-01-08 16:00:50 +07:00
|
|
|
if (check_pud_range(vma, pgd, addr, next, nodes,
|
2006-01-08 16:01:01 +07:00
|
|
|
flags, private))
|
2005-06-22 07:15:07 +07:00
|
|
|
return -EIO;
|
|
|
|
} while (pgd++, addr = next, addr != end);
|
|
|
|
return 0;
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
|
2006-01-08 16:00:50 +07:00
|
|
|
/*
|
|
|
|
* Check if all pages in a range are on a set of nodes.
|
|
|
|
* If pagelist != NULL then isolate pages from the LRU and
|
|
|
|
* put them on the pagelist.
|
|
|
|
*/
|
2005-04-17 05:20:36 +07:00
|
|
|
static struct vm_area_struct *
|
|
|
|
check_range(struct mm_struct *mm, unsigned long start, unsigned long end,
|
2006-01-08 16:01:01 +07:00
|
|
|
const nodemask_t *nodes, unsigned long flags, void *private)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
int err;
|
|
|
|
struct vm_area_struct *first, *vma, *prev;
|
|
|
|
|
2006-03-17 14:03:59 +07:00
|
|
|
if (flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL)) {
|
|
|
|
|
2006-03-22 15:09:12 +07:00
|
|
|
err = migrate_prep();
|
|
|
|
if (err)
|
|
|
|
return ERR_PTR(err);
|
2006-03-17 14:03:59 +07:00
|
|
|
}
|
2006-01-19 08:42:27 +07:00
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
first = find_vma(mm, start);
|
|
|
|
if (!first)
|
|
|
|
return ERR_PTR(-EFAULT);
|
|
|
|
prev = NULL;
|
|
|
|
for (vma = first; vma && vma->vm_start < end; vma = vma->vm_next) {
|
2006-01-08 16:00:50 +07:00
|
|
|
if (!(flags & MPOL_MF_DISCONTIG_OK)) {
|
|
|
|
if (!vma->vm_next && vma->vm_end < end)
|
|
|
|
return ERR_PTR(-EFAULT);
|
|
|
|
if (prev && prev->vm_end < vma->vm_start)
|
|
|
|
return ERR_PTR(-EFAULT);
|
|
|
|
}
|
|
|
|
if (!is_vm_hugetlb_page(vma) &&
|
|
|
|
((flags & MPOL_MF_STRICT) ||
|
|
|
|
((flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL)) &&
|
|
|
|
vma_migratable(vma)))) {
|
2005-09-13 15:25:08 +07:00
|
|
|
unsigned long endvma = vma->vm_end;
|
2006-01-08 16:00:50 +07:00
|
|
|
|
2005-09-13 15:25:08 +07:00
|
|
|
if (endvma > end)
|
|
|
|
endvma = end;
|
|
|
|
if (vma->vm_start > start)
|
|
|
|
start = vma->vm_start;
|
2006-01-08 16:00:50 +07:00
|
|
|
err = check_pgd_range(vma, start, endvma, nodes,
|
2006-01-08 16:01:01 +07:00
|
|
|
flags, private);
|
2005-04-17 05:20:36 +07:00
|
|
|
if (err) {
|
|
|
|
first = ERR_PTR(err);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
prev = vma;
|
|
|
|
}
|
|
|
|
return first;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Apply policy to a single VMA */
|
|
|
|
static int policy_vma(struct vm_area_struct *vma, struct mempolicy *new)
|
|
|
|
{
|
|
|
|
int err = 0;
|
|
|
|
struct mempolicy *old = vma->vm_policy;
|
|
|
|
|
2007-07-16 13:38:16 +07:00
|
|
|
pr_debug("vma %lx-%lx/%lx vm_ops %p vm_file %p set_policy %p\n",
|
2005-04-17 05:20:36 +07:00
|
|
|
vma->vm_start, vma->vm_end, vma->vm_pgoff,
|
|
|
|
vma->vm_ops, vma->vm_file,
|
|
|
|
vma->vm_ops ? vma->vm_ops->set_policy : NULL);
|
|
|
|
|
|
|
|
if (vma->vm_ops && vma->vm_ops->set_policy)
|
|
|
|
err = vma->vm_ops->set_policy(vma, new);
|
|
|
|
if (!err) {
|
|
|
|
mpol_get(new);
|
|
|
|
vma->vm_policy = new;
|
|
|
|
mpol_free(old);
|
|
|
|
}
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Step 2: apply policy to a range and do splits. */
|
|
|
|
static int mbind_range(struct vm_area_struct *vma, unsigned long start,
|
|
|
|
unsigned long end, struct mempolicy *new)
|
|
|
|
{
|
|
|
|
struct vm_area_struct *next;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
err = 0;
|
|
|
|
for (; vma && vma->vm_start < end; vma = next) {
|
|
|
|
next = vma->vm_next;
|
|
|
|
if (vma->vm_start < start)
|
|
|
|
err = split_vma(vma->vm_mm, vma, start, 1);
|
|
|
|
if (!err && vma->vm_end > end)
|
|
|
|
err = split_vma(vma->vm_mm, vma, end, 0);
|
|
|
|
if (!err)
|
|
|
|
err = policy_vma(vma, new);
|
|
|
|
if (err)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2005-10-30 08:16:59 +07:00
|
|
|
static int contextualize_policy(int mode, nodemask_t *nodes)
|
|
|
|
{
|
|
|
|
if (!nodes)
|
|
|
|
return 0;
|
|
|
|
|
2006-01-08 16:01:54 +07:00
|
|
|
cpuset_update_task_memory_state();
|
2006-01-08 16:01:47 +07:00
|
|
|
if (!cpuset_nodes_subset_current_mems_allowed(*nodes))
|
|
|
|
return -EINVAL;
|
2005-10-30 08:16:59 +07:00
|
|
|
return mpol_check_policy(mode, nodes);
|
|
|
|
}
|
|
|
|
|
2006-03-24 18:16:08 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Update task->flags PF_MEMPOLICY bit: set iff non-default
|
|
|
|
* mempolicy. Allows more rapid checking of this (combined perhaps
|
|
|
|
* with other PF_* flag bits) on memory allocation hot code paths.
|
|
|
|
*
|
|
|
|
* If called from outside this file, the task 'p' should -only- be
|
|
|
|
* a newly forked child not yet visible on the task list, because
|
|
|
|
* manipulating the task flags of a visible task is not safe.
|
|
|
|
*
|
|
|
|
* The above limitation is why this routine has the funny name
|
|
|
|
* mpol_fix_fork_child_flag().
|
|
|
|
*
|
|
|
|
* It is also safe to call this with a task pointer of current,
|
|
|
|
* which the static wrapper mpol_set_task_struct_flag() does,
|
|
|
|
* for use within this file.
|
|
|
|
*/
|
|
|
|
|
|
|
|
void mpol_fix_fork_child_flag(struct task_struct *p)
|
|
|
|
{
|
|
|
|
if (p->mempolicy)
|
|
|
|
p->flags |= PF_MEMPOLICY;
|
|
|
|
else
|
|
|
|
p->flags &= ~PF_MEMPOLICY;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void mpol_set_task_struct_flag(void)
|
|
|
|
{
|
|
|
|
mpol_fix_fork_child_flag(current);
|
|
|
|
}
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/* Set the process memory policy */
|
2007-10-16 15:26:26 +07:00
|
|
|
static long do_set_mempolicy(int mode, nodemask_t *nodes)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
struct mempolicy *new;
|
|
|
|
|
2005-10-30 08:16:59 +07:00
|
|
|
if (contextualize_policy(mode, nodes))
|
2005-04-17 05:20:36 +07:00
|
|
|
return -EINVAL;
|
2005-10-30 08:16:59 +07:00
|
|
|
new = mpol_new(mode, nodes);
|
2005-04-17 05:20:36 +07:00
|
|
|
if (IS_ERR(new))
|
|
|
|
return PTR_ERR(new);
|
|
|
|
mpol_free(current->mempolicy);
|
|
|
|
current->mempolicy = new;
|
2006-03-24 18:16:08 +07:00
|
|
|
mpol_set_task_struct_flag();
|
2005-04-17 05:20:36 +07:00
|
|
|
if (new && new->policy == MPOL_INTERLEAVE)
|
2005-10-30 08:15:48 +07:00
|
|
|
current->il_next = first_node(new->v.nodes);
|
2005-04-17 05:20:36 +07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Fill a zone bitmap for a policy */
|
2005-10-30 08:15:48 +07:00
|
|
|
static void get_zonemask(struct mempolicy *p, nodemask_t *nodes)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
2005-10-30 08:15:48 +07:00
|
|
|
nodes_clear(*nodes);
|
2005-04-17 05:20:36 +07:00
|
|
|
switch (p->policy) {
|
|
|
|
case MPOL_BIND:
|
|
|
|
for (i = 0; p->v.zonelist->zones[i]; i++)
|
2006-09-26 13:31:55 +07:00
|
|
|
node_set(zone_to_nid(p->v.zonelist->zones[i]),
|
2005-10-30 08:16:59 +07:00
|
|
|
*nodes);
|
2005-04-17 05:20:36 +07:00
|
|
|
break;
|
|
|
|
case MPOL_DEFAULT:
|
|
|
|
break;
|
|
|
|
case MPOL_INTERLEAVE:
|
2005-10-30 08:15:48 +07:00
|
|
|
*nodes = p->v.nodes;
|
2005-04-17 05:20:36 +07:00
|
|
|
break;
|
|
|
|
case MPOL_PREFERRED:
|
2007-10-16 15:25:35 +07:00
|
|
|
/* or use current node instead of memory_map? */
|
2005-04-17 05:20:36 +07:00
|
|
|
if (p->v.preferred_node < 0)
|
2007-10-16 15:25:35 +07:00
|
|
|
*nodes = node_states[N_HIGH_MEMORY];
|
2005-04-17 05:20:36 +07:00
|
|
|
else
|
2005-10-30 08:15:48 +07:00
|
|
|
node_set(p->v.preferred_node, *nodes);
|
2005-04-17 05:20:36 +07:00
|
|
|
break;
|
|
|
|
default:
|
|
|
|
BUG();
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static int lookup_node(struct mm_struct *mm, unsigned long addr)
|
|
|
|
{
|
|
|
|
struct page *p;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
err = get_user_pages(current, mm, addr & PAGE_MASK, 1, 0, 0, &p, NULL);
|
|
|
|
if (err >= 0) {
|
|
|
|
err = page_to_nid(p);
|
|
|
|
put_page(p);
|
|
|
|
}
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Retrieve NUMA policy */
|
2007-10-16 15:26:26 +07:00
|
|
|
static long do_get_mempolicy(int *policy, nodemask_t *nmask,
|
|
|
|
unsigned long addr, unsigned long flags)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
2005-10-30 08:16:59 +07:00
|
|
|
int err;
|
2005-04-17 05:20:36 +07:00
|
|
|
struct mm_struct *mm = current->mm;
|
|
|
|
struct vm_area_struct *vma = NULL;
|
|
|
|
struct mempolicy *pol = current->mempolicy;
|
|
|
|
|
2006-01-08 16:01:54 +07:00
|
|
|
cpuset_update_task_memory_state();
|
2007-10-16 15:24:51 +07:00
|
|
|
if (flags &
|
|
|
|
~(unsigned long)(MPOL_F_NODE|MPOL_F_ADDR|MPOL_F_MEMS_ALLOWED))
|
2005-04-17 05:20:36 +07:00
|
|
|
return -EINVAL;
|
2007-10-16 15:24:51 +07:00
|
|
|
|
|
|
|
if (flags & MPOL_F_MEMS_ALLOWED) {
|
|
|
|
if (flags & (MPOL_F_NODE|MPOL_F_ADDR))
|
|
|
|
return -EINVAL;
|
|
|
|
*policy = 0; /* just so it's initialized */
|
|
|
|
*nmask = cpuset_current_mems_allowed;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
if (flags & MPOL_F_ADDR) {
|
|
|
|
down_read(&mm->mmap_sem);
|
|
|
|
vma = find_vma_intersection(mm, addr, addr+1);
|
|
|
|
if (!vma) {
|
|
|
|
up_read(&mm->mmap_sem);
|
|
|
|
return -EFAULT;
|
|
|
|
}
|
|
|
|
if (vma->vm_ops && vma->vm_ops->get_policy)
|
|
|
|
pol = vma->vm_ops->get_policy(vma, addr);
|
|
|
|
else
|
|
|
|
pol = vma->vm_policy;
|
|
|
|
} else if (addr)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
if (!pol)
|
|
|
|
pol = &default_policy;
|
|
|
|
|
|
|
|
if (flags & MPOL_F_NODE) {
|
|
|
|
if (flags & MPOL_F_ADDR) {
|
|
|
|
err = lookup_node(mm, addr);
|
|
|
|
if (err < 0)
|
|
|
|
goto out;
|
2005-10-30 08:16:59 +07:00
|
|
|
*policy = err;
|
2005-04-17 05:20:36 +07:00
|
|
|
} else if (pol == current->mempolicy &&
|
|
|
|
pol->policy == MPOL_INTERLEAVE) {
|
2005-10-30 08:16:59 +07:00
|
|
|
*policy = current->il_next;
|
2005-04-17 05:20:36 +07:00
|
|
|
} else {
|
|
|
|
err = -EINVAL;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
} else
|
2005-10-30 08:16:59 +07:00
|
|
|
*policy = pol->policy;
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
if (vma) {
|
|
|
|
up_read(¤t->mm->mmap_sem);
|
|
|
|
vma = NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
err = 0;
|
2005-10-30 08:16:59 +07:00
|
|
|
if (nmask)
|
|
|
|
get_zonemask(pol, nmask);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
out:
|
|
|
|
if (vma)
|
|
|
|
up_read(¤t->mm->mmap_sem);
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2006-03-22 15:09:12 +07:00
|
|
|
#ifdef CONFIG_MIGRATION
|
2006-01-08 16:01:04 +07:00
|
|
|
/*
|
|
|
|
* page migration
|
|
|
|
*/
|
2006-01-19 08:42:29 +07:00
|
|
|
static void migrate_page_add(struct page *page, struct list_head *pagelist,
|
|
|
|
unsigned long flags)
|
2006-01-08 16:01:04 +07:00
|
|
|
{
|
|
|
|
/*
|
2006-01-19 08:42:29 +07:00
|
|
|
* Avoid migrating a page that is shared with others.
|
2006-01-08 16:01:04 +07:00
|
|
|
*/
|
2006-03-22 15:09:12 +07:00
|
|
|
if ((flags & MPOL_MF_MOVE_ALL) || page_mapcount(page) == 1)
|
|
|
|
isolate_lru_page(page, pagelist);
|
2006-02-01 18:05:40 +07:00
|
|
|
}
|
2006-01-08 16:01:04 +07:00
|
|
|
|
2006-06-23 16:03:55 +07:00
|
|
|
static struct page *new_node_page(struct page *page, unsigned long node, int **x)
|
2006-06-23 16:03:53 +07:00
|
|
|
{
|
2007-07-17 18:03:05 +07:00
|
|
|
return alloc_pages_node(node, GFP_HIGHUSER_MOVABLE, 0);
|
2006-06-23 16:03:53 +07:00
|
|
|
}
|
|
|
|
|
2006-02-01 18:05:40 +07:00
|
|
|
/*
|
|
|
|
* Migrate pages from one node to a target node.
|
|
|
|
* Returns error or the number of pages not migrated.
|
|
|
|
*/
|
2007-10-16 15:26:26 +07:00
|
|
|
static int migrate_to_node(struct mm_struct *mm, int source, int dest,
|
|
|
|
int flags)
|
2006-02-01 18:05:40 +07:00
|
|
|
{
|
|
|
|
nodemask_t nmask;
|
|
|
|
LIST_HEAD(pagelist);
|
|
|
|
int err = 0;
|
|
|
|
|
|
|
|
nodes_clear(nmask);
|
|
|
|
node_set(source, nmask);
|
2006-01-08 16:01:04 +07:00
|
|
|
|
2006-02-01 18:05:40 +07:00
|
|
|
check_range(mm, mm->mmap->vm_start, TASK_SIZE, &nmask,
|
|
|
|
flags | MPOL_MF_DISCONTIG_OK, &pagelist);
|
|
|
|
|
2006-06-23 16:03:52 +07:00
|
|
|
if (!list_empty(&pagelist))
|
2006-06-23 16:03:53 +07:00
|
|
|
err = migrate_pages(&pagelist, new_node_page, dest);
|
|
|
|
|
2006-02-01 18:05:40 +07:00
|
|
|
return err;
|
2006-01-08 16:01:04 +07:00
|
|
|
}
|
|
|
|
|
2006-01-08 16:00:51 +07:00
|
|
|
/*
|
2006-02-01 18:05:40 +07:00
|
|
|
* Move pages between the two nodesets so as to preserve the physical
|
|
|
|
* layout as much as possible.
|
2006-01-08 16:00:51 +07:00
|
|
|
*
|
|
|
|
* Returns the number of page that could not be moved.
|
|
|
|
*/
|
|
|
|
int do_migrate_pages(struct mm_struct *mm,
|
|
|
|
const nodemask_t *from_nodes, const nodemask_t *to_nodes, int flags)
|
|
|
|
{
|
|
|
|
LIST_HEAD(pagelist);
|
2006-02-01 18:05:40 +07:00
|
|
|
int busy = 0;
|
|
|
|
int err = 0;
|
|
|
|
nodemask_t tmp;
|
2006-01-08 16:00:51 +07:00
|
|
|
|
2006-02-01 18:05:40 +07:00
|
|
|
down_read(&mm->mmap_sem);
|
2006-01-08 16:00:51 +07:00
|
|
|
|
2006-06-25 19:46:48 +07:00
|
|
|
err = migrate_vmas(mm, from_nodes, to_nodes, flags);
|
|
|
|
if (err)
|
|
|
|
goto out;
|
|
|
|
|
2006-02-01 18:05:40 +07:00
|
|
|
/*
|
|
|
|
* Find a 'source' bit set in 'tmp' whose corresponding 'dest'
|
|
|
|
* bit in 'to' is not also set in 'tmp'. Clear the found 'source'
|
|
|
|
* bit in 'tmp', and return that <source, dest> pair for migration.
|
|
|
|
* The pair of nodemasks 'to' and 'from' define the map.
|
|
|
|
*
|
|
|
|
* If no pair of bits is found that way, fallback to picking some
|
|
|
|
* pair of 'source' and 'dest' bits that are not the same. If the
|
|
|
|
* 'source' and 'dest' bits are the same, this represents a node
|
|
|
|
* that will be migrating to itself, so no pages need move.
|
|
|
|
*
|
|
|
|
* If no bits are left in 'tmp', or if all remaining bits left
|
|
|
|
* in 'tmp' correspond to the same bit in 'to', return false
|
|
|
|
* (nothing left to migrate).
|
|
|
|
*
|
|
|
|
* This lets us pick a pair of nodes to migrate between, such that
|
|
|
|
* if possible the dest node is not already occupied by some other
|
|
|
|
* source node, minimizing the risk of overloading the memory on a
|
|
|
|
* node that would happen if we migrated incoming memory to a node
|
|
|
|
* before migrating outgoing memory source that same node.
|
|
|
|
*
|
|
|
|
* A single scan of tmp is sufficient. As we go, we remember the
|
|
|
|
* most recent <s, d> pair that moved (s != d). If we find a pair
|
|
|
|
* that not only moved, but what's better, moved to an empty slot
|
|
|
|
* (d is not set in tmp), then we break out then, with that pair.
|
|
|
|
* Otherwise when we finish scannng from_tmp, we at least have the
|
|
|
|
* most recent <s, d> pair that moved. If we get all the way through
|
|
|
|
* the scan of tmp without finding any node that moved, much less
|
|
|
|
* moved to an empty node, then there is nothing left worth migrating.
|
|
|
|
*/
|
2006-01-08 16:00:55 +07:00
|
|
|
|
2006-02-01 18:05:40 +07:00
|
|
|
tmp = *from_nodes;
|
|
|
|
while (!nodes_empty(tmp)) {
|
|
|
|
int s,d;
|
|
|
|
int source = -1;
|
|
|
|
int dest = 0;
|
|
|
|
|
|
|
|
for_each_node_mask(s, tmp) {
|
|
|
|
d = node_remap(s, *from_nodes, *to_nodes);
|
|
|
|
if (s == d)
|
|
|
|
continue;
|
|
|
|
|
|
|
|
source = s; /* Node moved. Memorize */
|
|
|
|
dest = d;
|
|
|
|
|
|
|
|
/* dest not in remaining from nodes? */
|
|
|
|
if (!node_isset(dest, tmp))
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (source == -1)
|
|
|
|
break;
|
|
|
|
|
|
|
|
node_clear(source, tmp);
|
|
|
|
err = migrate_to_node(mm, source, dest, flags);
|
|
|
|
if (err > 0)
|
|
|
|
busy += err;
|
|
|
|
if (err < 0)
|
|
|
|
break;
|
2006-01-08 16:00:51 +07:00
|
|
|
}
|
2006-06-25 19:46:48 +07:00
|
|
|
out:
|
2006-01-08 16:00:51 +07:00
|
|
|
up_read(&mm->mmap_sem);
|
2006-02-01 18:05:40 +07:00
|
|
|
if (err < 0)
|
|
|
|
return err;
|
|
|
|
return busy;
|
2006-03-22 15:09:12 +07:00
|
|
|
|
|
|
|
}
|
|
|
|
|
2006-06-23 16:03:55 +07:00
|
|
|
static struct page *new_vma_page(struct page *page, unsigned long private, int **x)
|
2006-06-23 16:03:53 +07:00
|
|
|
{
|
|
|
|
struct vm_area_struct *vma = (struct vm_area_struct *)private;
|
|
|
|
|
2007-07-17 18:03:05 +07:00
|
|
|
return alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma,
|
|
|
|
page_address_in_vma(page, vma));
|
2006-06-23 16:03:53 +07:00
|
|
|
}
|
2006-03-22 15:09:12 +07:00
|
|
|
#else
|
|
|
|
|
|
|
|
static void migrate_page_add(struct page *page, struct list_head *pagelist,
|
|
|
|
unsigned long flags)
|
|
|
|
{
|
2006-01-08 16:00:51 +07:00
|
|
|
}
|
|
|
|
|
2006-03-22 15:09:12 +07:00
|
|
|
int do_migrate_pages(struct mm_struct *mm,
|
|
|
|
const nodemask_t *from_nodes, const nodemask_t *to_nodes, int flags)
|
|
|
|
{
|
|
|
|
return -ENOSYS;
|
|
|
|
}
|
2006-06-23 16:03:53 +07:00
|
|
|
|
2006-10-11 15:21:28 +07:00
|
|
|
static struct page *new_vma_page(struct page *page, unsigned long private, int **x)
|
2006-06-23 16:03:53 +07:00
|
|
|
{
|
|
|
|
return NULL;
|
|
|
|
}
|
2006-03-22 15:09:12 +07:00
|
|
|
#endif
|
|
|
|
|
2007-10-16 15:26:26 +07:00
|
|
|
static long do_mbind(unsigned long start, unsigned long len,
|
|
|
|
unsigned long mode, nodemask_t *nmask,
|
|
|
|
unsigned long flags)
|
2006-01-08 16:01:04 +07:00
|
|
|
{
|
|
|
|
struct vm_area_struct *vma;
|
|
|
|
struct mm_struct *mm = current->mm;
|
|
|
|
struct mempolicy *new;
|
|
|
|
unsigned long end;
|
|
|
|
int err;
|
|
|
|
LIST_HEAD(pagelist);
|
|
|
|
|
|
|
|
if ((flags & ~(unsigned long)(MPOL_MF_STRICT |
|
|
|
|
MPOL_MF_MOVE | MPOL_MF_MOVE_ALL))
|
|
|
|
|| mode > MPOL_MAX)
|
|
|
|
return -EINVAL;
|
2006-03-15 10:50:21 +07:00
|
|
|
if ((flags & MPOL_MF_MOVE_ALL) && !capable(CAP_SYS_NICE))
|
2006-01-08 16:01:04 +07:00
|
|
|
return -EPERM;
|
|
|
|
|
|
|
|
if (start & ~PAGE_MASK)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
if (mode == MPOL_DEFAULT)
|
|
|
|
flags &= ~MPOL_MF_STRICT;
|
|
|
|
|
|
|
|
len = (len + PAGE_SIZE - 1) & PAGE_MASK;
|
|
|
|
end = start + len;
|
|
|
|
|
|
|
|
if (end < start)
|
|
|
|
return -EINVAL;
|
|
|
|
if (end == start)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
if (mpol_check_policy(mode, nmask))
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
new = mpol_new(mode, nmask);
|
|
|
|
if (IS_ERR(new))
|
|
|
|
return PTR_ERR(new);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If we are using the default policy then operation
|
|
|
|
* on discontinuous address spaces is okay after all
|
|
|
|
*/
|
|
|
|
if (!new)
|
|
|
|
flags |= MPOL_MF_DISCONTIG_OK;
|
|
|
|
|
2007-07-16 13:38:16 +07:00
|
|
|
pr_debug("mbind %lx-%lx mode:%ld nodes:%lx\n",start,start+len,
|
|
|
|
mode, nmask ? nodes_addr(*nmask)[0] : -1);
|
2006-01-08 16:01:04 +07:00
|
|
|
|
|
|
|
down_write(&mm->mmap_sem);
|
|
|
|
vma = check_range(mm, start, end, nmask,
|
|
|
|
flags | MPOL_MF_INVERT, &pagelist);
|
|
|
|
|
|
|
|
err = PTR_ERR(vma);
|
|
|
|
if (!IS_ERR(vma)) {
|
|
|
|
int nr_failed = 0;
|
|
|
|
|
|
|
|
err = mbind_range(vma, start, end, new);
|
2006-02-01 18:05:40 +07:00
|
|
|
|
2006-01-08 16:01:04 +07:00
|
|
|
if (!list_empty(&pagelist))
|
2006-06-23 16:03:53 +07:00
|
|
|
nr_failed = migrate_pages(&pagelist, new_vma_page,
|
|
|
|
(unsigned long)vma);
|
2006-01-08 16:01:04 +07:00
|
|
|
|
|
|
|
if (!err && nr_failed && (flags & MPOL_MF_STRICT))
|
|
|
|
err = -EIO;
|
|
|
|
}
|
2006-03-22 15:09:12 +07:00
|
|
|
|
2006-01-08 16:01:04 +07:00
|
|
|
up_write(&mm->mmap_sem);
|
|
|
|
mpol_free(new);
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2005-10-30 08:16:59 +07:00
|
|
|
/*
|
|
|
|
* User space interface with variable sized bitmaps for nodelists.
|
|
|
|
*/
|
|
|
|
|
|
|
|
/* Copy a node mask from user space. */
|
2006-01-08 16:00:51 +07:00
|
|
|
static int get_nodes(nodemask_t *nodes, const unsigned long __user *nmask,
|
2005-10-30 08:16:59 +07:00
|
|
|
unsigned long maxnode)
|
|
|
|
{
|
|
|
|
unsigned long k;
|
|
|
|
unsigned long nlongs;
|
|
|
|
unsigned long endmask;
|
|
|
|
|
|
|
|
--maxnode;
|
|
|
|
nodes_clear(*nodes);
|
|
|
|
if (maxnode == 0 || !nmask)
|
|
|
|
return 0;
|
2006-02-21 09:27:59 +07:00
|
|
|
if (maxnode > PAGE_SIZE*BITS_PER_BYTE)
|
2006-02-18 04:59:36 +07:00
|
|
|
return -EINVAL;
|
2005-10-30 08:16:59 +07:00
|
|
|
|
|
|
|
nlongs = BITS_TO_LONGS(maxnode);
|
|
|
|
if ((maxnode % BITS_PER_LONG) == 0)
|
|
|
|
endmask = ~0UL;
|
|
|
|
else
|
|
|
|
endmask = (1UL << (maxnode % BITS_PER_LONG)) - 1;
|
|
|
|
|
|
|
|
/* When the user specified more nodes than supported just check
|
|
|
|
if the non supported part is all zero. */
|
|
|
|
if (nlongs > BITS_TO_LONGS(MAX_NUMNODES)) {
|
|
|
|
if (nlongs > PAGE_SIZE/sizeof(long))
|
|
|
|
return -EINVAL;
|
|
|
|
for (k = BITS_TO_LONGS(MAX_NUMNODES); k < nlongs; k++) {
|
|
|
|
unsigned long t;
|
|
|
|
if (get_user(t, nmask + k))
|
|
|
|
return -EFAULT;
|
|
|
|
if (k == nlongs - 1) {
|
|
|
|
if (t & endmask)
|
|
|
|
return -EINVAL;
|
|
|
|
} else if (t)
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
nlongs = BITS_TO_LONGS(MAX_NUMNODES);
|
|
|
|
endmask = ~0UL;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (copy_from_user(nodes_addr(*nodes), nmask, nlongs*sizeof(unsigned long)))
|
|
|
|
return -EFAULT;
|
|
|
|
nodes_addr(*nodes)[nlongs-1] &= endmask;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Copy a kernel node mask to user space */
|
|
|
|
static int copy_nodes_to_user(unsigned long __user *mask, unsigned long maxnode,
|
|
|
|
nodemask_t *nodes)
|
|
|
|
{
|
|
|
|
unsigned long copy = ALIGN(maxnode-1, 64) / 8;
|
|
|
|
const int nbytes = BITS_TO_LONGS(MAX_NUMNODES) * sizeof(long);
|
|
|
|
|
|
|
|
if (copy > nbytes) {
|
|
|
|
if (copy > PAGE_SIZE)
|
|
|
|
return -EINVAL;
|
|
|
|
if (clear_user((char __user *)mask + nbytes, copy - nbytes))
|
|
|
|
return -EFAULT;
|
|
|
|
copy = nbytes;
|
|
|
|
}
|
|
|
|
return copy_to_user(mask, nodes_addr(*nodes), copy) ? -EFAULT : 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
asmlinkage long sys_mbind(unsigned long start, unsigned long len,
|
|
|
|
unsigned long mode,
|
|
|
|
unsigned long __user *nmask, unsigned long maxnode,
|
|
|
|
unsigned flags)
|
|
|
|
{
|
|
|
|
nodemask_t nodes;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
err = get_nodes(&nodes, nmask, maxnode);
|
|
|
|
if (err)
|
|
|
|
return err;
|
2007-01-23 11:40:45 +07:00
|
|
|
#ifdef CONFIG_CPUSETS
|
|
|
|
/* Restrict the nodes to the allowed nodes in the cpuset */
|
|
|
|
nodes_and(nodes, nodes, current->mems_allowed);
|
|
|
|
#endif
|
2005-10-30 08:16:59 +07:00
|
|
|
return do_mbind(start, len, mode, &nodes, flags);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Set the process memory policy */
|
|
|
|
asmlinkage long sys_set_mempolicy(int mode, unsigned long __user *nmask,
|
|
|
|
unsigned long maxnode)
|
|
|
|
{
|
|
|
|
int err;
|
|
|
|
nodemask_t nodes;
|
|
|
|
|
|
|
|
if (mode < 0 || mode > MPOL_MAX)
|
|
|
|
return -EINVAL;
|
|
|
|
err = get_nodes(&nodes, nmask, maxnode);
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
return do_set_mempolicy(mode, &nodes);
|
|
|
|
}
|
|
|
|
|
2006-01-08 16:00:51 +07:00
|
|
|
asmlinkage long sys_migrate_pages(pid_t pid, unsigned long maxnode,
|
|
|
|
const unsigned long __user *old_nodes,
|
|
|
|
const unsigned long __user *new_nodes)
|
|
|
|
{
|
|
|
|
struct mm_struct *mm;
|
|
|
|
struct task_struct *task;
|
|
|
|
nodemask_t old;
|
|
|
|
nodemask_t new;
|
|
|
|
nodemask_t task_nodes;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
err = get_nodes(&old, old_nodes, maxnode);
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
|
|
|
|
err = get_nodes(&new, new_nodes, maxnode);
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
|
|
|
|
/* Find the mm_struct */
|
|
|
|
read_lock(&tasklist_lock);
|
2007-10-19 13:40:16 +07:00
|
|
|
task = pid ? find_task_by_vpid(pid) : current;
|
2006-01-08 16:00:51 +07:00
|
|
|
if (!task) {
|
|
|
|
read_unlock(&tasklist_lock);
|
|
|
|
return -ESRCH;
|
|
|
|
}
|
|
|
|
mm = get_task_mm(task);
|
|
|
|
read_unlock(&tasklist_lock);
|
|
|
|
|
|
|
|
if (!mm)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Check if this process has the right to modify the specified
|
|
|
|
* process. The right exists if the process has administrative
|
2006-03-28 16:56:53 +07:00
|
|
|
* capabilities, superuser privileges or the same
|
2006-01-08 16:00:51 +07:00
|
|
|
* userid as the target process.
|
|
|
|
*/
|
|
|
|
if ((current->euid != task->suid) && (current->euid != task->uid) &&
|
|
|
|
(current->uid != task->suid) && (current->uid != task->uid) &&
|
2006-03-15 10:50:21 +07:00
|
|
|
!capable(CAP_SYS_NICE)) {
|
2006-01-08 16:00:51 +07:00
|
|
|
err = -EPERM;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
task_nodes = cpuset_mems_allowed(task);
|
|
|
|
/* Is the user allowed to access the target nodes? */
|
2006-03-15 10:50:21 +07:00
|
|
|
if (!nodes_subset(new, task_nodes) && !capable(CAP_SYS_NICE)) {
|
2006-01-08 16:00:51 +07:00
|
|
|
err = -EPERM;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2007-10-16 15:25:39 +07:00
|
|
|
if (!nodes_subset(new, node_states[N_HIGH_MEMORY])) {
|
2007-08-31 14:12:08 +07:00
|
|
|
err = -EINVAL;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2006-06-23 16:04:02 +07:00
|
|
|
err = security_task_movememory(task);
|
|
|
|
if (err)
|
|
|
|
goto out;
|
|
|
|
|
2006-03-01 07:58:57 +07:00
|
|
|
err = do_migrate_pages(mm, &old, &new,
|
2006-03-15 10:50:21 +07:00
|
|
|
capable(CAP_SYS_NICE) ? MPOL_MF_MOVE_ALL : MPOL_MF_MOVE);
|
2006-01-08 16:00:51 +07:00
|
|
|
out:
|
|
|
|
mmput(mm);
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2005-10-30 08:16:59 +07:00
|
|
|
/* Retrieve NUMA policy */
|
|
|
|
asmlinkage long sys_get_mempolicy(int __user *policy,
|
|
|
|
unsigned long __user *nmask,
|
|
|
|
unsigned long maxnode,
|
|
|
|
unsigned long addr, unsigned long flags)
|
|
|
|
{
|
2007-10-16 15:26:26 +07:00
|
|
|
int err;
|
|
|
|
int uninitialized_var(pval);
|
2005-10-30 08:16:59 +07:00
|
|
|
nodemask_t nodes;
|
|
|
|
|
|
|
|
if (nmask != NULL && maxnode < MAX_NUMNODES)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
err = do_get_mempolicy(&pval, &nodes, addr, flags);
|
|
|
|
|
|
|
|
if (err)
|
|
|
|
return err;
|
|
|
|
|
|
|
|
if (policy && put_user(pval, policy))
|
|
|
|
return -EFAULT;
|
|
|
|
|
|
|
|
if (nmask)
|
|
|
|
err = copy_nodes_to_user(nmask, maxnode, &nodes);
|
|
|
|
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
|
|
|
|
asmlinkage long compat_sys_get_mempolicy(int __user *policy,
|
|
|
|
compat_ulong_t __user *nmask,
|
|
|
|
compat_ulong_t maxnode,
|
|
|
|
compat_ulong_t addr, compat_ulong_t flags)
|
|
|
|
{
|
|
|
|
long err;
|
|
|
|
unsigned long __user *nm = NULL;
|
|
|
|
unsigned long nr_bits, alloc_size;
|
|
|
|
DECLARE_BITMAP(bm, MAX_NUMNODES);
|
|
|
|
|
|
|
|
nr_bits = min_t(unsigned long, maxnode-1, MAX_NUMNODES);
|
|
|
|
alloc_size = ALIGN(nr_bits, BITS_PER_LONG) / 8;
|
|
|
|
|
|
|
|
if (nmask)
|
|
|
|
nm = compat_alloc_user_space(alloc_size);
|
|
|
|
|
|
|
|
err = sys_get_mempolicy(policy, nm, nr_bits+1, addr, flags);
|
|
|
|
|
|
|
|
if (!err && nmask) {
|
|
|
|
err = copy_from_user(bm, nm, alloc_size);
|
|
|
|
/* ensure entire bitmap is zeroed */
|
|
|
|
err |= clear_user(nmask, ALIGN(maxnode-1, 8) / 8);
|
|
|
|
err |= compat_put_bitmap(nmask, bm, nr_bits);
|
|
|
|
}
|
|
|
|
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
asmlinkage long compat_sys_set_mempolicy(int mode, compat_ulong_t __user *nmask,
|
|
|
|
compat_ulong_t maxnode)
|
|
|
|
{
|
|
|
|
long err = 0;
|
|
|
|
unsigned long __user *nm = NULL;
|
|
|
|
unsigned long nr_bits, alloc_size;
|
|
|
|
DECLARE_BITMAP(bm, MAX_NUMNODES);
|
|
|
|
|
|
|
|
nr_bits = min_t(unsigned long, maxnode-1, MAX_NUMNODES);
|
|
|
|
alloc_size = ALIGN(nr_bits, BITS_PER_LONG) / 8;
|
|
|
|
|
|
|
|
if (nmask) {
|
|
|
|
err = compat_get_bitmap(bm, nmask, nr_bits);
|
|
|
|
nm = compat_alloc_user_space(alloc_size);
|
|
|
|
err |= copy_to_user(nm, bm, alloc_size);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (err)
|
|
|
|
return -EFAULT;
|
|
|
|
|
|
|
|
return sys_set_mempolicy(mode, nm, nr_bits+1);
|
|
|
|
}
|
|
|
|
|
|
|
|
asmlinkage long compat_sys_mbind(compat_ulong_t start, compat_ulong_t len,
|
|
|
|
compat_ulong_t mode, compat_ulong_t __user *nmask,
|
|
|
|
compat_ulong_t maxnode, compat_ulong_t flags)
|
|
|
|
{
|
|
|
|
long err = 0;
|
|
|
|
unsigned long __user *nm = NULL;
|
|
|
|
unsigned long nr_bits, alloc_size;
|
2005-10-30 08:15:48 +07:00
|
|
|
nodemask_t bm;
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
nr_bits = min_t(unsigned long, maxnode-1, MAX_NUMNODES);
|
|
|
|
alloc_size = ALIGN(nr_bits, BITS_PER_LONG) / 8;
|
|
|
|
|
|
|
|
if (nmask) {
|
2005-10-30 08:15:48 +07:00
|
|
|
err = compat_get_bitmap(nodes_addr(bm), nmask, nr_bits);
|
2005-04-17 05:20:36 +07:00
|
|
|
nm = compat_alloc_user_space(alloc_size);
|
2005-10-30 08:15:48 +07:00
|
|
|
err |= copy_to_user(nm, nodes_addr(bm), alloc_size);
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
if (err)
|
|
|
|
return -EFAULT;
|
|
|
|
|
|
|
|
return sys_mbind(start, len, mode, nm, nr_bits+1, flags);
|
|
|
|
}
|
|
|
|
|
|
|
|
#endif
|
|
|
|
|
Fix NUMA Memory Policy Reference Counting
This patch proposes fixes to the reference counting of memory policy in the
page allocation paths and in show_numa_map(). Extracted from my "Memory
Policy Cleanups and Enhancements" series as stand-alone.
Shared policy lookup [shmem] has always added a reference to the policy,
but this was never unrefed after page allocation or after formatting the
numa map data.
Default system policy should not require additional ref counting, nor
should the current task's task policy. However, show_numa_map() calls
get_vma_policy() to examine what may be [likely is] another task's policy.
The latter case needs protection against freeing of the policy.
This patch adds a reference count to a mempolicy returned by
get_vma_policy() when the policy is a vma policy or another task's
mempolicy. Again, shared policy is already reference counted on lookup. A
matching "unref" [__mpol_free()] is performed in alloc_page_vma() for
shared and vma policies, and in show_numa_map() for shared and another
task's mempolicy. We can call __mpol_free() directly, saving an admittedly
inexpensive inline NULL test, because we know we have a non-NULL policy.
Handling policy ref counts for hugepages is a bit trickier.
huge_zonelist() returns a zone list that might come from a shared or vma
'BIND policy. In this case, we should hold the reference until after the
huge page allocation in dequeue_hugepage(). The patch modifies
huge_zonelist() to return a pointer to the mempolicy if it needs to be
unref'd after allocation.
Kernel Build [16cpu, 32GB, ia64] - average of 10 runs:
w/o patch w/ refcount patch
Avg Std Devn Avg Std Devn
Real: 100.59 0.38 100.63 0.43
User: 1209.60 0.37 1209.91 0.31
System: 81.52 0.42 81.64 0.34
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Acked-by: Andi Kleen <ak@suse.de>
Cc: Christoph Lameter <clameter@sgi.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-09-19 12:46:47 +07:00
|
|
|
/*
|
|
|
|
* get_vma_policy(@task, @vma, @addr)
|
|
|
|
* @task - task for fallback if vma policy == default
|
|
|
|
* @vma - virtual memory area whose policy is sought
|
|
|
|
* @addr - address in @vma for shared policy lookup
|
|
|
|
*
|
|
|
|
* Returns effective policy for a VMA at specified address.
|
|
|
|
* Falls back to @task or system default policy, as necessary.
|
|
|
|
* Returned policy has extra reference count if shared, vma,
|
|
|
|
* or some other task's policy [show_numa_maps() can pass
|
|
|
|
* @task != current]. It is the caller's responsibility to
|
|
|
|
* free the reference in these cases.
|
|
|
|
*/
|
2006-01-08 16:01:03 +07:00
|
|
|
static struct mempolicy * get_vma_policy(struct task_struct *task,
|
|
|
|
struct vm_area_struct *vma, unsigned long addr)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
2005-09-04 05:54:45 +07:00
|
|
|
struct mempolicy *pol = task->mempolicy;
|
Fix NUMA Memory Policy Reference Counting
This patch proposes fixes to the reference counting of memory policy in the
page allocation paths and in show_numa_map(). Extracted from my "Memory
Policy Cleanups and Enhancements" series as stand-alone.
Shared policy lookup [shmem] has always added a reference to the policy,
but this was never unrefed after page allocation or after formatting the
numa map data.
Default system policy should not require additional ref counting, nor
should the current task's task policy. However, show_numa_map() calls
get_vma_policy() to examine what may be [likely is] another task's policy.
The latter case needs protection against freeing of the policy.
This patch adds a reference count to a mempolicy returned by
get_vma_policy() when the policy is a vma policy or another task's
mempolicy. Again, shared policy is already reference counted on lookup. A
matching "unref" [__mpol_free()] is performed in alloc_page_vma() for
shared and vma policies, and in show_numa_map() for shared and another
task's mempolicy. We can call __mpol_free() directly, saving an admittedly
inexpensive inline NULL test, because we know we have a non-NULL policy.
Handling policy ref counts for hugepages is a bit trickier.
huge_zonelist() returns a zone list that might come from a shared or vma
'BIND policy. In this case, we should hold the reference until after the
huge page allocation in dequeue_hugepage(). The patch modifies
huge_zonelist() to return a pointer to the mempolicy if it needs to be
unref'd after allocation.
Kernel Build [16cpu, 32GB, ia64] - average of 10 runs:
w/o patch w/ refcount patch
Avg Std Devn Avg Std Devn
Real: 100.59 0.38 100.63 0.43
User: 1209.60 0.37 1209.91 0.31
System: 81.52 0.42 81.64 0.34
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Acked-by: Andi Kleen <ak@suse.de>
Cc: Christoph Lameter <clameter@sgi.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-09-19 12:46:47 +07:00
|
|
|
int shared_pol = 0;
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
if (vma) {
|
Fix NUMA Memory Policy Reference Counting
This patch proposes fixes to the reference counting of memory policy in the
page allocation paths and in show_numa_map(). Extracted from my "Memory
Policy Cleanups and Enhancements" series as stand-alone.
Shared policy lookup [shmem] has always added a reference to the policy,
but this was never unrefed after page allocation or after formatting the
numa map data.
Default system policy should not require additional ref counting, nor
should the current task's task policy. However, show_numa_map() calls
get_vma_policy() to examine what may be [likely is] another task's policy.
The latter case needs protection against freeing of the policy.
This patch adds a reference count to a mempolicy returned by
get_vma_policy() when the policy is a vma policy or another task's
mempolicy. Again, shared policy is already reference counted on lookup. A
matching "unref" [__mpol_free()] is performed in alloc_page_vma() for
shared and vma policies, and in show_numa_map() for shared and another
task's mempolicy. We can call __mpol_free() directly, saving an admittedly
inexpensive inline NULL test, because we know we have a non-NULL policy.
Handling policy ref counts for hugepages is a bit trickier.
huge_zonelist() returns a zone list that might come from a shared or vma
'BIND policy. In this case, we should hold the reference until after the
huge page allocation in dequeue_hugepage(). The patch modifies
huge_zonelist() to return a pointer to the mempolicy if it needs to be
unref'd after allocation.
Kernel Build [16cpu, 32GB, ia64] - average of 10 runs:
w/o patch w/ refcount patch
Avg Std Devn Avg Std Devn
Real: 100.59 0.38 100.63 0.43
User: 1209.60 0.37 1209.91 0.31
System: 81.52 0.42 81.64 0.34
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Acked-by: Andi Kleen <ak@suse.de>
Cc: Christoph Lameter <clameter@sgi.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-09-19 12:46:47 +07:00
|
|
|
if (vma->vm_ops && vma->vm_ops->get_policy) {
|
2005-10-30 08:16:59 +07:00
|
|
|
pol = vma->vm_ops->get_policy(vma, addr);
|
Fix NUMA Memory Policy Reference Counting
This patch proposes fixes to the reference counting of memory policy in the
page allocation paths and in show_numa_map(). Extracted from my "Memory
Policy Cleanups and Enhancements" series as stand-alone.
Shared policy lookup [shmem] has always added a reference to the policy,
but this was never unrefed after page allocation or after formatting the
numa map data.
Default system policy should not require additional ref counting, nor
should the current task's task policy. However, show_numa_map() calls
get_vma_policy() to examine what may be [likely is] another task's policy.
The latter case needs protection against freeing of the policy.
This patch adds a reference count to a mempolicy returned by
get_vma_policy() when the policy is a vma policy or another task's
mempolicy. Again, shared policy is already reference counted on lookup. A
matching "unref" [__mpol_free()] is performed in alloc_page_vma() for
shared and vma policies, and in show_numa_map() for shared and another
task's mempolicy. We can call __mpol_free() directly, saving an admittedly
inexpensive inline NULL test, because we know we have a non-NULL policy.
Handling policy ref counts for hugepages is a bit trickier.
huge_zonelist() returns a zone list that might come from a shared or vma
'BIND policy. In this case, we should hold the reference until after the
huge page allocation in dequeue_hugepage(). The patch modifies
huge_zonelist() to return a pointer to the mempolicy if it needs to be
unref'd after allocation.
Kernel Build [16cpu, 32GB, ia64] - average of 10 runs:
w/o patch w/ refcount patch
Avg Std Devn Avg Std Devn
Real: 100.59 0.38 100.63 0.43
User: 1209.60 0.37 1209.91 0.31
System: 81.52 0.42 81.64 0.34
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Acked-by: Andi Kleen <ak@suse.de>
Cc: Christoph Lameter <clameter@sgi.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-09-19 12:46:47 +07:00
|
|
|
shared_pol = 1; /* if pol non-NULL, add ref below */
|
|
|
|
} else if (vma->vm_policy &&
|
2005-04-17 05:20:36 +07:00
|
|
|
vma->vm_policy->policy != MPOL_DEFAULT)
|
|
|
|
pol = vma->vm_policy;
|
|
|
|
}
|
|
|
|
if (!pol)
|
|
|
|
pol = &default_policy;
|
Fix NUMA Memory Policy Reference Counting
This patch proposes fixes to the reference counting of memory policy in the
page allocation paths and in show_numa_map(). Extracted from my "Memory
Policy Cleanups and Enhancements" series as stand-alone.
Shared policy lookup [shmem] has always added a reference to the policy,
but this was never unrefed after page allocation or after formatting the
numa map data.
Default system policy should not require additional ref counting, nor
should the current task's task policy. However, show_numa_map() calls
get_vma_policy() to examine what may be [likely is] another task's policy.
The latter case needs protection against freeing of the policy.
This patch adds a reference count to a mempolicy returned by
get_vma_policy() when the policy is a vma policy or another task's
mempolicy. Again, shared policy is already reference counted on lookup. A
matching "unref" [__mpol_free()] is performed in alloc_page_vma() for
shared and vma policies, and in show_numa_map() for shared and another
task's mempolicy. We can call __mpol_free() directly, saving an admittedly
inexpensive inline NULL test, because we know we have a non-NULL policy.
Handling policy ref counts for hugepages is a bit trickier.
huge_zonelist() returns a zone list that might come from a shared or vma
'BIND policy. In this case, we should hold the reference until after the
huge page allocation in dequeue_hugepage(). The patch modifies
huge_zonelist() to return a pointer to the mempolicy if it needs to be
unref'd after allocation.
Kernel Build [16cpu, 32GB, ia64] - average of 10 runs:
w/o patch w/ refcount patch
Avg Std Devn Avg Std Devn
Real: 100.59 0.38 100.63 0.43
User: 1209.60 0.37 1209.91 0.31
System: 81.52 0.42 81.64 0.34
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Acked-by: Andi Kleen <ak@suse.de>
Cc: Christoph Lameter <clameter@sgi.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-09-19 12:46:47 +07:00
|
|
|
else if (!shared_pol && pol != current->mempolicy)
|
|
|
|
mpol_get(pol); /* vma or other task's policy */
|
2005-04-17 05:20:36 +07:00
|
|
|
return pol;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Return a zonelist representing a mempolicy */
|
2005-10-07 13:46:04 +07:00
|
|
|
static struct zonelist *zonelist_policy(gfp_t gfp, struct mempolicy *policy)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
int nd;
|
|
|
|
|
|
|
|
switch (policy->policy) {
|
|
|
|
case MPOL_PREFERRED:
|
|
|
|
nd = policy->v.preferred_node;
|
|
|
|
if (nd < 0)
|
|
|
|
nd = numa_node_id();
|
|
|
|
break;
|
|
|
|
case MPOL_BIND:
|
|
|
|
/* Lower zones don't get a policy applied */
|
|
|
|
/* Careful: current->mems_allowed might have moved */
|
2006-09-26 13:31:19 +07:00
|
|
|
if (gfp_zone(gfp) >= policy_zone)
|
2005-04-17 05:20:36 +07:00
|
|
|
if (cpuset_zonelist_valid_mems_allowed(policy->v.zonelist))
|
|
|
|
return policy->v.zonelist;
|
|
|
|
/*FALL THROUGH*/
|
|
|
|
case MPOL_INTERLEAVE: /* should not happen */
|
|
|
|
case MPOL_DEFAULT:
|
|
|
|
nd = numa_node_id();
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
nd = 0;
|
|
|
|
BUG();
|
|
|
|
}
|
2005-10-21 13:55:38 +07:00
|
|
|
return NODE_DATA(nd)->node_zonelists + gfp_zone(gfp);
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Do dynamic interleaving for a process */
|
|
|
|
static unsigned interleave_nodes(struct mempolicy *policy)
|
|
|
|
{
|
|
|
|
unsigned nid, next;
|
|
|
|
struct task_struct *me = current;
|
|
|
|
|
|
|
|
nid = me->il_next;
|
2005-10-30 08:15:48 +07:00
|
|
|
next = next_node(nid, policy->v.nodes);
|
2005-04-17 05:20:36 +07:00
|
|
|
if (next >= MAX_NUMNODES)
|
2005-10-30 08:15:48 +07:00
|
|
|
next = first_node(policy->v.nodes);
|
2005-04-17 05:20:36 +07:00
|
|
|
me->il_next = next;
|
|
|
|
return nid;
|
|
|
|
}
|
|
|
|
|
2006-01-19 08:42:36 +07:00
|
|
|
/*
|
|
|
|
* Depending on the memory policy provide a node from which to allocate the
|
|
|
|
* next slab entry.
|
|
|
|
*/
|
|
|
|
unsigned slab_node(struct mempolicy *policy)
|
|
|
|
{
|
2006-09-27 15:50:08 +07:00
|
|
|
int pol = policy ? policy->policy : MPOL_DEFAULT;
|
|
|
|
|
|
|
|
switch (pol) {
|
2006-01-19 08:42:36 +07:00
|
|
|
case MPOL_INTERLEAVE:
|
|
|
|
return interleave_nodes(policy);
|
|
|
|
|
|
|
|
case MPOL_BIND:
|
|
|
|
/*
|
|
|
|
* Follow bind policy behavior and start allocation at the
|
|
|
|
* first node.
|
|
|
|
*/
|
2006-09-26 13:31:55 +07:00
|
|
|
return zone_to_nid(policy->v.zonelist->zones[0]);
|
2006-01-19 08:42:36 +07:00
|
|
|
|
|
|
|
case MPOL_PREFERRED:
|
|
|
|
if (policy->v.preferred_node >= 0)
|
|
|
|
return policy->v.preferred_node;
|
|
|
|
/* Fall through */
|
|
|
|
|
|
|
|
default:
|
|
|
|
return numa_node_id();
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/* Do static interleaving for a VMA with known offset. */
|
|
|
|
static unsigned offset_il_node(struct mempolicy *pol,
|
|
|
|
struct vm_area_struct *vma, unsigned long off)
|
|
|
|
{
|
2005-10-30 08:15:48 +07:00
|
|
|
unsigned nnodes = nodes_weight(pol->v.nodes);
|
2005-04-17 05:20:36 +07:00
|
|
|
unsigned target = (unsigned)off % nnodes;
|
|
|
|
int c;
|
|
|
|
int nid = -1;
|
|
|
|
|
|
|
|
c = 0;
|
|
|
|
do {
|
2005-10-30 08:15:48 +07:00
|
|
|
nid = next_node(nid, pol->v.nodes);
|
2005-04-17 05:20:36 +07:00
|
|
|
c++;
|
|
|
|
} while (c <= target);
|
|
|
|
return nid;
|
|
|
|
}
|
|
|
|
|
2006-01-06 15:10:46 +07:00
|
|
|
/* Determine a node number for interleave */
|
|
|
|
static inline unsigned interleave_nid(struct mempolicy *pol,
|
|
|
|
struct vm_area_struct *vma, unsigned long addr, int shift)
|
|
|
|
{
|
|
|
|
if (vma) {
|
|
|
|
unsigned long off;
|
|
|
|
|
2006-09-01 11:27:53 +07:00
|
|
|
/*
|
|
|
|
* for small pages, there is no difference between
|
|
|
|
* shift and PAGE_SHIFT, so the bit-shift is safe.
|
|
|
|
* for huge pages, since vm_pgoff is in units of small
|
|
|
|
* pages, we need to shift off the always 0 bits to get
|
|
|
|
* a useful offset.
|
|
|
|
*/
|
|
|
|
BUG_ON(shift < PAGE_SHIFT);
|
|
|
|
off = vma->vm_pgoff >> (shift - PAGE_SHIFT);
|
2006-01-06 15:10:46 +07:00
|
|
|
off += (addr - vma->vm_start) >> shift;
|
|
|
|
return offset_il_node(pol, vma, off);
|
|
|
|
} else
|
|
|
|
return interleave_nodes(pol);
|
|
|
|
}
|
|
|
|
|
2006-02-04 03:51:14 +07:00
|
|
|
#ifdef CONFIG_HUGETLBFS
|
Fix NUMA Memory Policy Reference Counting
This patch proposes fixes to the reference counting of memory policy in the
page allocation paths and in show_numa_map(). Extracted from my "Memory
Policy Cleanups and Enhancements" series as stand-alone.
Shared policy lookup [shmem] has always added a reference to the policy,
but this was never unrefed after page allocation or after formatting the
numa map data.
Default system policy should not require additional ref counting, nor
should the current task's task policy. However, show_numa_map() calls
get_vma_policy() to examine what may be [likely is] another task's policy.
The latter case needs protection against freeing of the policy.
This patch adds a reference count to a mempolicy returned by
get_vma_policy() when the policy is a vma policy or another task's
mempolicy. Again, shared policy is already reference counted on lookup. A
matching "unref" [__mpol_free()] is performed in alloc_page_vma() for
shared and vma policies, and in show_numa_map() for shared and another
task's mempolicy. We can call __mpol_free() directly, saving an admittedly
inexpensive inline NULL test, because we know we have a non-NULL policy.
Handling policy ref counts for hugepages is a bit trickier.
huge_zonelist() returns a zone list that might come from a shared or vma
'BIND policy. In this case, we should hold the reference until after the
huge page allocation in dequeue_hugepage(). The patch modifies
huge_zonelist() to return a pointer to the mempolicy if it needs to be
unref'd after allocation.
Kernel Build [16cpu, 32GB, ia64] - average of 10 runs:
w/o patch w/ refcount patch
Avg Std Devn Avg Std Devn
Real: 100.59 0.38 100.63 0.43
User: 1209.60 0.37 1209.91 0.31
System: 81.52 0.42 81.64 0.34
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Acked-by: Andi Kleen <ak@suse.de>
Cc: Christoph Lameter <clameter@sgi.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-09-19 12:46:47 +07:00
|
|
|
/*
|
|
|
|
* huge_zonelist(@vma, @addr, @gfp_flags, @mpol)
|
|
|
|
* @vma = virtual memory area whose policy is sought
|
|
|
|
* @addr = address in @vma for shared policy lookup and interleave policy
|
|
|
|
* @gfp_flags = for requested zone
|
|
|
|
* @mpol = pointer to mempolicy pointer for reference counted 'BIND policy
|
|
|
|
*
|
|
|
|
* Returns a zonelist suitable for a huge page allocation.
|
|
|
|
* If the effective policy is 'BIND, returns pointer to policy's zonelist.
|
|
|
|
* If it is also a policy for which get_vma_policy() returns an extra
|
|
|
|
* reference, we must hold that reference until after allocation.
|
|
|
|
* In that case, return policy via @mpol so hugetlb allocation can drop
|
|
|
|
* the reference. For non-'BIND referenced policies, we can/do drop the
|
|
|
|
* reference here, so the caller doesn't need to know about the special case
|
|
|
|
* for default and current task policy.
|
|
|
|
*/
|
2007-07-17 18:03:13 +07:00
|
|
|
struct zonelist *huge_zonelist(struct vm_area_struct *vma, unsigned long addr,
|
Fix NUMA Memory Policy Reference Counting
This patch proposes fixes to the reference counting of memory policy in the
page allocation paths and in show_numa_map(). Extracted from my "Memory
Policy Cleanups and Enhancements" series as stand-alone.
Shared policy lookup [shmem] has always added a reference to the policy,
but this was never unrefed after page allocation or after formatting the
numa map data.
Default system policy should not require additional ref counting, nor
should the current task's task policy. However, show_numa_map() calls
get_vma_policy() to examine what may be [likely is] another task's policy.
The latter case needs protection against freeing of the policy.
This patch adds a reference count to a mempolicy returned by
get_vma_policy() when the policy is a vma policy or another task's
mempolicy. Again, shared policy is already reference counted on lookup. A
matching "unref" [__mpol_free()] is performed in alloc_page_vma() for
shared and vma policies, and in show_numa_map() for shared and another
task's mempolicy. We can call __mpol_free() directly, saving an admittedly
inexpensive inline NULL test, because we know we have a non-NULL policy.
Handling policy ref counts for hugepages is a bit trickier.
huge_zonelist() returns a zone list that might come from a shared or vma
'BIND policy. In this case, we should hold the reference until after the
huge page allocation in dequeue_hugepage(). The patch modifies
huge_zonelist() to return a pointer to the mempolicy if it needs to be
unref'd after allocation.
Kernel Build [16cpu, 32GB, ia64] - average of 10 runs:
w/o patch w/ refcount patch
Avg Std Devn Avg Std Devn
Real: 100.59 0.38 100.63 0.43
User: 1209.60 0.37 1209.91 0.31
System: 81.52 0.42 81.64 0.34
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Acked-by: Andi Kleen <ak@suse.de>
Cc: Christoph Lameter <clameter@sgi.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-09-19 12:46:47 +07:00
|
|
|
gfp_t gfp_flags, struct mempolicy **mpol)
|
2006-01-06 15:10:46 +07:00
|
|
|
{
|
|
|
|
struct mempolicy *pol = get_vma_policy(current, vma, addr);
|
Fix NUMA Memory Policy Reference Counting
This patch proposes fixes to the reference counting of memory policy in the
page allocation paths and in show_numa_map(). Extracted from my "Memory
Policy Cleanups and Enhancements" series as stand-alone.
Shared policy lookup [shmem] has always added a reference to the policy,
but this was never unrefed after page allocation or after formatting the
numa map data.
Default system policy should not require additional ref counting, nor
should the current task's task policy. However, show_numa_map() calls
get_vma_policy() to examine what may be [likely is] another task's policy.
The latter case needs protection against freeing of the policy.
This patch adds a reference count to a mempolicy returned by
get_vma_policy() when the policy is a vma policy or another task's
mempolicy. Again, shared policy is already reference counted on lookup. A
matching "unref" [__mpol_free()] is performed in alloc_page_vma() for
shared and vma policies, and in show_numa_map() for shared and another
task's mempolicy. We can call __mpol_free() directly, saving an admittedly
inexpensive inline NULL test, because we know we have a non-NULL policy.
Handling policy ref counts for hugepages is a bit trickier.
huge_zonelist() returns a zone list that might come from a shared or vma
'BIND policy. In this case, we should hold the reference until after the
huge page allocation in dequeue_hugepage(). The patch modifies
huge_zonelist() to return a pointer to the mempolicy if it needs to be
unref'd after allocation.
Kernel Build [16cpu, 32GB, ia64] - average of 10 runs:
w/o patch w/ refcount patch
Avg Std Devn Avg Std Devn
Real: 100.59 0.38 100.63 0.43
User: 1209.60 0.37 1209.91 0.31
System: 81.52 0.42 81.64 0.34
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Acked-by: Andi Kleen <ak@suse.de>
Cc: Christoph Lameter <clameter@sgi.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-09-19 12:46:47 +07:00
|
|
|
struct zonelist *zl;
|
2006-01-06 15:10:46 +07:00
|
|
|
|
Fix NUMA Memory Policy Reference Counting
This patch proposes fixes to the reference counting of memory policy in the
page allocation paths and in show_numa_map(). Extracted from my "Memory
Policy Cleanups and Enhancements" series as stand-alone.
Shared policy lookup [shmem] has always added a reference to the policy,
but this was never unrefed after page allocation or after formatting the
numa map data.
Default system policy should not require additional ref counting, nor
should the current task's task policy. However, show_numa_map() calls
get_vma_policy() to examine what may be [likely is] another task's policy.
The latter case needs protection against freeing of the policy.
This patch adds a reference count to a mempolicy returned by
get_vma_policy() when the policy is a vma policy or another task's
mempolicy. Again, shared policy is already reference counted on lookup. A
matching "unref" [__mpol_free()] is performed in alloc_page_vma() for
shared and vma policies, and in show_numa_map() for shared and another
task's mempolicy. We can call __mpol_free() directly, saving an admittedly
inexpensive inline NULL test, because we know we have a non-NULL policy.
Handling policy ref counts for hugepages is a bit trickier.
huge_zonelist() returns a zone list that might come from a shared or vma
'BIND policy. In this case, we should hold the reference until after the
huge page allocation in dequeue_hugepage(). The patch modifies
huge_zonelist() to return a pointer to the mempolicy if it needs to be
unref'd after allocation.
Kernel Build [16cpu, 32GB, ia64] - average of 10 runs:
w/o patch w/ refcount patch
Avg Std Devn Avg Std Devn
Real: 100.59 0.38 100.63 0.43
User: 1209.60 0.37 1209.91 0.31
System: 81.52 0.42 81.64 0.34
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Acked-by: Andi Kleen <ak@suse.de>
Cc: Christoph Lameter <clameter@sgi.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-09-19 12:46:47 +07:00
|
|
|
*mpol = NULL; /* probably no unref needed */
|
2006-01-06 15:10:46 +07:00
|
|
|
if (pol->policy == MPOL_INTERLEAVE) {
|
|
|
|
unsigned nid;
|
|
|
|
|
|
|
|
nid = interleave_nid(pol, vma, addr, HPAGE_SHIFT);
|
Fix NUMA Memory Policy Reference Counting
This patch proposes fixes to the reference counting of memory policy in the
page allocation paths and in show_numa_map(). Extracted from my "Memory
Policy Cleanups and Enhancements" series as stand-alone.
Shared policy lookup [shmem] has always added a reference to the policy,
but this was never unrefed after page allocation or after formatting the
numa map data.
Default system policy should not require additional ref counting, nor
should the current task's task policy. However, show_numa_map() calls
get_vma_policy() to examine what may be [likely is] another task's policy.
The latter case needs protection against freeing of the policy.
This patch adds a reference count to a mempolicy returned by
get_vma_policy() when the policy is a vma policy or another task's
mempolicy. Again, shared policy is already reference counted on lookup. A
matching "unref" [__mpol_free()] is performed in alloc_page_vma() for
shared and vma policies, and in show_numa_map() for shared and another
task's mempolicy. We can call __mpol_free() directly, saving an admittedly
inexpensive inline NULL test, because we know we have a non-NULL policy.
Handling policy ref counts for hugepages is a bit trickier.
huge_zonelist() returns a zone list that might come from a shared or vma
'BIND policy. In this case, we should hold the reference until after the
huge page allocation in dequeue_hugepage(). The patch modifies
huge_zonelist() to return a pointer to the mempolicy if it needs to be
unref'd after allocation.
Kernel Build [16cpu, 32GB, ia64] - average of 10 runs:
w/o patch w/ refcount patch
Avg Std Devn Avg Std Devn
Real: 100.59 0.38 100.63 0.43
User: 1209.60 0.37 1209.91 0.31
System: 81.52 0.42 81.64 0.34
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Acked-by: Andi Kleen <ak@suse.de>
Cc: Christoph Lameter <clameter@sgi.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-09-19 12:46:47 +07:00
|
|
|
__mpol_free(pol); /* finished with pol */
|
2007-07-17 18:03:13 +07:00
|
|
|
return NODE_DATA(nid)->node_zonelists + gfp_zone(gfp_flags);
|
2006-01-06 15:10:46 +07:00
|
|
|
}
|
Fix NUMA Memory Policy Reference Counting
This patch proposes fixes to the reference counting of memory policy in the
page allocation paths and in show_numa_map(). Extracted from my "Memory
Policy Cleanups and Enhancements" series as stand-alone.
Shared policy lookup [shmem] has always added a reference to the policy,
but this was never unrefed after page allocation or after formatting the
numa map data.
Default system policy should not require additional ref counting, nor
should the current task's task policy. However, show_numa_map() calls
get_vma_policy() to examine what may be [likely is] another task's policy.
The latter case needs protection against freeing of the policy.
This patch adds a reference count to a mempolicy returned by
get_vma_policy() when the policy is a vma policy or another task's
mempolicy. Again, shared policy is already reference counted on lookup. A
matching "unref" [__mpol_free()] is performed in alloc_page_vma() for
shared and vma policies, and in show_numa_map() for shared and another
task's mempolicy. We can call __mpol_free() directly, saving an admittedly
inexpensive inline NULL test, because we know we have a non-NULL policy.
Handling policy ref counts for hugepages is a bit trickier.
huge_zonelist() returns a zone list that might come from a shared or vma
'BIND policy. In this case, we should hold the reference until after the
huge page allocation in dequeue_hugepage(). The patch modifies
huge_zonelist() to return a pointer to the mempolicy if it needs to be
unref'd after allocation.
Kernel Build [16cpu, 32GB, ia64] - average of 10 runs:
w/o patch w/ refcount patch
Avg Std Devn Avg Std Devn
Real: 100.59 0.38 100.63 0.43
User: 1209.60 0.37 1209.91 0.31
System: 81.52 0.42 81.64 0.34
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Acked-by: Andi Kleen <ak@suse.de>
Cc: Christoph Lameter <clameter@sgi.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-09-19 12:46:47 +07:00
|
|
|
|
|
|
|
zl = zonelist_policy(GFP_HIGHUSER, pol);
|
|
|
|
if (unlikely(pol != &default_policy && pol != current->mempolicy)) {
|
|
|
|
if (pol->policy != MPOL_BIND)
|
|
|
|
__mpol_free(pol); /* finished with pol */
|
|
|
|
else
|
|
|
|
*mpol = pol; /* unref needed after allocation */
|
|
|
|
}
|
|
|
|
return zl;
|
2006-01-06 15:10:46 +07:00
|
|
|
}
|
2006-02-04 03:51:14 +07:00
|
|
|
#endif
|
2006-01-06 15:10:46 +07:00
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/* Allocate a page in interleaved policy.
|
|
|
|
Own path because it needs to do special accounting. */
|
2005-10-30 08:15:49 +07:00
|
|
|
static struct page *alloc_page_interleave(gfp_t gfp, unsigned order,
|
|
|
|
unsigned nid)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
struct zonelist *zl;
|
|
|
|
struct page *page;
|
|
|
|
|
2005-10-21 13:55:38 +07:00
|
|
|
zl = NODE_DATA(nid)->node_zonelists + gfp_zone(gfp);
|
2005-04-17 05:20:36 +07:00
|
|
|
page = __alloc_pages(gfp, order, zl);
|
2006-06-30 15:55:44 +07:00
|
|
|
if (page && page_zone(page) == zl->zones[0])
|
|
|
|
inc_zone_page_state(page, NUMA_INTERLEAVE_HIT);
|
2005-04-17 05:20:36 +07:00
|
|
|
return page;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* alloc_page_vma - Allocate a page for a VMA.
|
|
|
|
*
|
|
|
|
* @gfp:
|
|
|
|
* %GFP_USER user allocation.
|
|
|
|
* %GFP_KERNEL kernel allocations,
|
|
|
|
* %GFP_HIGHMEM highmem/user allocations,
|
|
|
|
* %GFP_FS allocation should not call back into a file system.
|
|
|
|
* %GFP_ATOMIC don't sleep.
|
|
|
|
*
|
|
|
|
* @vma: Pointer to VMA or NULL if not available.
|
|
|
|
* @addr: Virtual Address of the allocation. Must be inside the VMA.
|
|
|
|
*
|
|
|
|
* This function allocates a page from the kernel page pool and applies
|
|
|
|
* a NUMA policy associated with the VMA or the current process.
|
|
|
|
* When VMA is not NULL caller must hold down_read on the mmap_sem of the
|
|
|
|
* mm_struct of the VMA to prevent it from going away. Should be used for
|
|
|
|
* all allocations for pages that will be mapped into
|
|
|
|
* user space. Returns NULL when no page can be allocated.
|
|
|
|
*
|
|
|
|
* Should be called with the mm_sem of the vma hold.
|
|
|
|
*/
|
|
|
|
struct page *
|
2005-10-07 13:46:04 +07:00
|
|
|
alloc_page_vma(gfp_t gfp, struct vm_area_struct *vma, unsigned long addr)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
2005-09-04 05:54:45 +07:00
|
|
|
struct mempolicy *pol = get_vma_policy(current, vma, addr);
|
Fix NUMA Memory Policy Reference Counting
This patch proposes fixes to the reference counting of memory policy in the
page allocation paths and in show_numa_map(). Extracted from my "Memory
Policy Cleanups and Enhancements" series as stand-alone.
Shared policy lookup [shmem] has always added a reference to the policy,
but this was never unrefed after page allocation or after formatting the
numa map data.
Default system policy should not require additional ref counting, nor
should the current task's task policy. However, show_numa_map() calls
get_vma_policy() to examine what may be [likely is] another task's policy.
The latter case needs protection against freeing of the policy.
This patch adds a reference count to a mempolicy returned by
get_vma_policy() when the policy is a vma policy or another task's
mempolicy. Again, shared policy is already reference counted on lookup. A
matching "unref" [__mpol_free()] is performed in alloc_page_vma() for
shared and vma policies, and in show_numa_map() for shared and another
task's mempolicy. We can call __mpol_free() directly, saving an admittedly
inexpensive inline NULL test, because we know we have a non-NULL policy.
Handling policy ref counts for hugepages is a bit trickier.
huge_zonelist() returns a zone list that might come from a shared or vma
'BIND policy. In this case, we should hold the reference until after the
huge page allocation in dequeue_hugepage(). The patch modifies
huge_zonelist() to return a pointer to the mempolicy if it needs to be
unref'd after allocation.
Kernel Build [16cpu, 32GB, ia64] - average of 10 runs:
w/o patch w/ refcount patch
Avg Std Devn Avg Std Devn
Real: 100.59 0.38 100.63 0.43
User: 1209.60 0.37 1209.91 0.31
System: 81.52 0.42 81.64 0.34
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Acked-by: Andi Kleen <ak@suse.de>
Cc: Christoph Lameter <clameter@sgi.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-09-19 12:46:47 +07:00
|
|
|
struct zonelist *zl;
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2006-01-08 16:01:54 +07:00
|
|
|
cpuset_update_task_memory_state();
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
if (unlikely(pol->policy == MPOL_INTERLEAVE)) {
|
|
|
|
unsigned nid;
|
2006-01-06 15:10:46 +07:00
|
|
|
|
|
|
|
nid = interleave_nid(pol, vma, addr, PAGE_SHIFT);
|
2005-04-17 05:20:36 +07:00
|
|
|
return alloc_page_interleave(gfp, 0, nid);
|
|
|
|
}
|
Fix NUMA Memory Policy Reference Counting
This patch proposes fixes to the reference counting of memory policy in the
page allocation paths and in show_numa_map(). Extracted from my "Memory
Policy Cleanups and Enhancements" series as stand-alone.
Shared policy lookup [shmem] has always added a reference to the policy,
but this was never unrefed after page allocation or after formatting the
numa map data.
Default system policy should not require additional ref counting, nor
should the current task's task policy. However, show_numa_map() calls
get_vma_policy() to examine what may be [likely is] another task's policy.
The latter case needs protection against freeing of the policy.
This patch adds a reference count to a mempolicy returned by
get_vma_policy() when the policy is a vma policy or another task's
mempolicy. Again, shared policy is already reference counted on lookup. A
matching "unref" [__mpol_free()] is performed in alloc_page_vma() for
shared and vma policies, and in show_numa_map() for shared and another
task's mempolicy. We can call __mpol_free() directly, saving an admittedly
inexpensive inline NULL test, because we know we have a non-NULL policy.
Handling policy ref counts for hugepages is a bit trickier.
huge_zonelist() returns a zone list that might come from a shared or vma
'BIND policy. In this case, we should hold the reference until after the
huge page allocation in dequeue_hugepage(). The patch modifies
huge_zonelist() to return a pointer to the mempolicy if it needs to be
unref'd after allocation.
Kernel Build [16cpu, 32GB, ia64] - average of 10 runs:
w/o patch w/ refcount patch
Avg Std Devn Avg Std Devn
Real: 100.59 0.38 100.63 0.43
User: 1209.60 0.37 1209.91 0.31
System: 81.52 0.42 81.64 0.34
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Acked-by: Andi Kleen <ak@suse.de>
Cc: Christoph Lameter <clameter@sgi.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-09-19 12:46:47 +07:00
|
|
|
zl = zonelist_policy(gfp, pol);
|
|
|
|
if (pol != &default_policy && pol != current->mempolicy) {
|
|
|
|
/*
|
|
|
|
* slow path: ref counted policy -- shared or vma
|
|
|
|
*/
|
|
|
|
struct page *page = __alloc_pages(gfp, 0, zl);
|
|
|
|
__mpol_free(pol);
|
|
|
|
return page;
|
|
|
|
}
|
|
|
|
/*
|
|
|
|
* fast path: default or task policy
|
|
|
|
*/
|
|
|
|
return __alloc_pages(gfp, 0, zl);
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* alloc_pages_current - Allocate pages.
|
|
|
|
*
|
|
|
|
* @gfp:
|
|
|
|
* %GFP_USER user allocation,
|
|
|
|
* %GFP_KERNEL kernel allocation,
|
|
|
|
* %GFP_HIGHMEM highmem allocation,
|
|
|
|
* %GFP_FS don't call back into a file system.
|
|
|
|
* %GFP_ATOMIC don't sleep.
|
|
|
|
* @order: Power of two of allocation size in pages. 0 is a single page.
|
|
|
|
*
|
|
|
|
* Allocate a page from the kernel page pool. When not in
|
|
|
|
* interrupt context and apply the current process NUMA policy.
|
|
|
|
* Returns NULL when no page can be allocated.
|
|
|
|
*
|
2006-01-08 16:01:54 +07:00
|
|
|
* Don't call cpuset_update_task_memory_state() unless
|
2005-04-17 05:20:36 +07:00
|
|
|
* 1) it's ok to take cpuset_sem (can WAIT), and
|
|
|
|
* 2) allocating for current task (not interrupt).
|
|
|
|
*/
|
2005-10-07 13:46:04 +07:00
|
|
|
struct page *alloc_pages_current(gfp_t gfp, unsigned order)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
struct mempolicy *pol = current->mempolicy;
|
|
|
|
|
|
|
|
if ((gfp & __GFP_WAIT) && !in_interrupt())
|
2006-01-08 16:01:54 +07:00
|
|
|
cpuset_update_task_memory_state();
|
2006-09-26 13:31:40 +07:00
|
|
|
if (!pol || in_interrupt() || (gfp & __GFP_THISNODE))
|
2005-04-17 05:20:36 +07:00
|
|
|
pol = &default_policy;
|
|
|
|
if (pol->policy == MPOL_INTERLEAVE)
|
|
|
|
return alloc_page_interleave(gfp, order, interleave_nodes(pol));
|
|
|
|
return __alloc_pages(gfp, order, zonelist_policy(gfp, pol));
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(alloc_pages_current);
|
|
|
|
|
[PATCH] cpuset: rebind vma mempolicies fix
Fix more of longstanding bug in cpuset/mempolicy interaction.
NUMA mempolicies (mm/mempolicy.c) are constrained by the current tasks cpuset
to just the Memory Nodes allowed by that cpuset. The kernel maintains
internal state for each mempolicy, tracking what nodes are used for the
MPOL_INTERLEAVE, MPOL_BIND or MPOL_PREFERRED policies.
When a tasks cpuset memory placement changes, whether because the cpuset
changed, or because the task was attached to a different cpuset, then the
tasks mempolicies have to be rebound to the new cpuset placement, so as to
preserve the cpuset-relative numbering of the nodes in that policy.
An earlier fix handled such mempolicy rebinding for mempolicies attached to a
task.
This fix rebinds mempolicies attached to vma's (address ranges in a tasks
address space.) Due to the need to hold the task->mm->mmap_sem semaphore while
updating vma's, the rebinding of vma mempolicies has to be done when the
cpuset memory placement is changed, at which time mmap_sem can be safely
acquired. The tasks mempolicy is rebound later, when the task next attempts
to allocate memory and notices that its task->cpuset_mems_generation is
out-of-date with its cpusets mems_generation.
Because walking the tasklist to find all tasks attached to a changing cpuset
requires holding tasklist_lock, a spinlock, one cannot update the vma's of the
affected tasks while doing the tasklist scan. In general, one cannot acquire
a semaphore (which can sleep) while already holding a spinlock (such as
tasklist_lock). So a list of mm references has to be built up during the
tasklist scan, then the tasklist lock dropped, then for each mm, its mmap_sem
acquired, and the vma's in that mm rebound.
Once the tasklist lock is dropped, affected tasks may fork new tasks, before
their mm's are rebound. A kernel global 'cpuset_being_rebound' is set to
point to the cpuset being rebound (there can only be one; cpuset modifications
are done under a global 'manage_sem' semaphore), and the mpol_copy code that
is used to copy a tasks mempolicies during fork catches such forking tasks,
and ensures their children are also rebound.
When a task is moved to a different cpuset, it is easier, as there is only one
task involved. It's mm->vma's are scanned, using the same
mpol_rebind_policy() as used above.
It may happen that both the mpol_copy hook and the update done via the
tasklist scan update the same mm twice. This is ok, as the mempolicies of
each vma in an mm keep track of what mems_allowed they are relative to, and
safely no-op a second request to rebind to the same nodes.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-08 16:01:59 +07:00
|
|
|
/*
|
|
|
|
* If mpol_copy() sees current->cpuset == cpuset_being_rebound, then it
|
|
|
|
* rebinds the mempolicy its copying by calling mpol_rebind_policy()
|
|
|
|
* with the mems_allowed returned by cpuset_mems_allowed(). This
|
|
|
|
* keeps mempolicies cpuset relative after its cpuset moves. See
|
|
|
|
* further kernel/cpuset.c update_nodemask().
|
|
|
|
*/
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/* Slow path of a mempolicy copy */
|
|
|
|
struct mempolicy *__mpol_copy(struct mempolicy *old)
|
|
|
|
{
|
|
|
|
struct mempolicy *new = kmem_cache_alloc(policy_cache, GFP_KERNEL);
|
|
|
|
|
|
|
|
if (!new)
|
|
|
|
return ERR_PTR(-ENOMEM);
|
[PATCH] cpuset: rebind vma mempolicies fix
Fix more of longstanding bug in cpuset/mempolicy interaction.
NUMA mempolicies (mm/mempolicy.c) are constrained by the current tasks cpuset
to just the Memory Nodes allowed by that cpuset. The kernel maintains
internal state for each mempolicy, tracking what nodes are used for the
MPOL_INTERLEAVE, MPOL_BIND or MPOL_PREFERRED policies.
When a tasks cpuset memory placement changes, whether because the cpuset
changed, or because the task was attached to a different cpuset, then the
tasks mempolicies have to be rebound to the new cpuset placement, so as to
preserve the cpuset-relative numbering of the nodes in that policy.
An earlier fix handled such mempolicy rebinding for mempolicies attached to a
task.
This fix rebinds mempolicies attached to vma's (address ranges in a tasks
address space.) Due to the need to hold the task->mm->mmap_sem semaphore while
updating vma's, the rebinding of vma mempolicies has to be done when the
cpuset memory placement is changed, at which time mmap_sem can be safely
acquired. The tasks mempolicy is rebound later, when the task next attempts
to allocate memory and notices that its task->cpuset_mems_generation is
out-of-date with its cpusets mems_generation.
Because walking the tasklist to find all tasks attached to a changing cpuset
requires holding tasklist_lock, a spinlock, one cannot update the vma's of the
affected tasks while doing the tasklist scan. In general, one cannot acquire
a semaphore (which can sleep) while already holding a spinlock (such as
tasklist_lock). So a list of mm references has to be built up during the
tasklist scan, then the tasklist lock dropped, then for each mm, its mmap_sem
acquired, and the vma's in that mm rebound.
Once the tasklist lock is dropped, affected tasks may fork new tasks, before
their mm's are rebound. A kernel global 'cpuset_being_rebound' is set to
point to the cpuset being rebound (there can only be one; cpuset modifications
are done under a global 'manage_sem' semaphore), and the mpol_copy code that
is used to copy a tasks mempolicies during fork catches such forking tasks,
and ensures their children are also rebound.
When a task is moved to a different cpuset, it is easier, as there is only one
task involved. It's mm->vma's are scanned, using the same
mpol_rebind_policy() as used above.
It may happen that both the mpol_copy hook and the update done via the
tasklist scan update the same mm twice. This is ok, as the mempolicies of
each vma in an mm keep track of what mems_allowed they are relative to, and
safely no-op a second request to rebind to the same nodes.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-08 16:01:59 +07:00
|
|
|
if (current_cpuset_is_being_rebound()) {
|
|
|
|
nodemask_t mems = cpuset_mems_allowed(current);
|
|
|
|
mpol_rebind_policy(old, &mems);
|
|
|
|
}
|
2005-04-17 05:20:36 +07:00
|
|
|
*new = *old;
|
|
|
|
atomic_set(&new->refcnt, 1);
|
|
|
|
if (new->policy == MPOL_BIND) {
|
|
|
|
int sz = ksize(old->v.zonelist);
|
2006-12-07 11:33:17 +07:00
|
|
|
new->v.zonelist = kmemdup(old->v.zonelist, sz, GFP_KERNEL);
|
2005-04-17 05:20:36 +07:00
|
|
|
if (!new->v.zonelist) {
|
|
|
|
kmem_cache_free(policy_cache, new);
|
|
|
|
return ERR_PTR(-ENOMEM);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return new;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Slow path of a mempolicy comparison */
|
|
|
|
int __mpol_equal(struct mempolicy *a, struct mempolicy *b)
|
|
|
|
{
|
|
|
|
if (!a || !b)
|
|
|
|
return 0;
|
|
|
|
if (a->policy != b->policy)
|
|
|
|
return 0;
|
|
|
|
switch (a->policy) {
|
|
|
|
case MPOL_DEFAULT:
|
|
|
|
return 1;
|
|
|
|
case MPOL_INTERLEAVE:
|
2005-10-30 08:15:48 +07:00
|
|
|
return nodes_equal(a->v.nodes, b->v.nodes);
|
2005-04-17 05:20:36 +07:00
|
|
|
case MPOL_PREFERRED:
|
|
|
|
return a->v.preferred_node == b->v.preferred_node;
|
|
|
|
case MPOL_BIND: {
|
|
|
|
int i;
|
|
|
|
for (i = 0; a->v.zonelist->zones[i]; i++)
|
|
|
|
if (a->v.zonelist->zones[i] != b->v.zonelist->zones[i])
|
|
|
|
return 0;
|
|
|
|
return b->v.zonelist->zones[i] == NULL;
|
|
|
|
}
|
|
|
|
default:
|
|
|
|
BUG();
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Slow path of a mpol destructor. */
|
|
|
|
void __mpol_free(struct mempolicy *p)
|
|
|
|
{
|
|
|
|
if (!atomic_dec_and_test(&p->refcnt))
|
|
|
|
return;
|
|
|
|
if (p->policy == MPOL_BIND)
|
|
|
|
kfree(p->v.zonelist);
|
|
|
|
p->policy = MPOL_DEFAULT;
|
|
|
|
kmem_cache_free(policy_cache, p);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Shared memory backing store policy support.
|
|
|
|
*
|
|
|
|
* Remember policies even when nobody has shared memory mapped.
|
|
|
|
* The policies are kept in Red-Black tree linked from the inode.
|
|
|
|
* They are protected by the sp->lock spinlock, which should be held
|
|
|
|
* for any accesses to the tree.
|
|
|
|
*/
|
|
|
|
|
|
|
|
/* lookup first element intersecting start-end */
|
|
|
|
/* Caller holds sp->lock */
|
|
|
|
static struct sp_node *
|
|
|
|
sp_lookup(struct shared_policy *sp, unsigned long start, unsigned long end)
|
|
|
|
{
|
|
|
|
struct rb_node *n = sp->root.rb_node;
|
|
|
|
|
|
|
|
while (n) {
|
|
|
|
struct sp_node *p = rb_entry(n, struct sp_node, nd);
|
|
|
|
|
|
|
|
if (start >= p->end)
|
|
|
|
n = n->rb_right;
|
|
|
|
else if (end <= p->start)
|
|
|
|
n = n->rb_left;
|
|
|
|
else
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (!n)
|
|
|
|
return NULL;
|
|
|
|
for (;;) {
|
|
|
|
struct sp_node *w = NULL;
|
|
|
|
struct rb_node *prev = rb_prev(n);
|
|
|
|
if (!prev)
|
|
|
|
break;
|
|
|
|
w = rb_entry(prev, struct sp_node, nd);
|
|
|
|
if (w->end <= start)
|
|
|
|
break;
|
|
|
|
n = prev;
|
|
|
|
}
|
|
|
|
return rb_entry(n, struct sp_node, nd);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Insert a new shared policy into the list. */
|
|
|
|
/* Caller holds sp->lock */
|
|
|
|
static void sp_insert(struct shared_policy *sp, struct sp_node *new)
|
|
|
|
{
|
|
|
|
struct rb_node **p = &sp->root.rb_node;
|
|
|
|
struct rb_node *parent = NULL;
|
|
|
|
struct sp_node *nd;
|
|
|
|
|
|
|
|
while (*p) {
|
|
|
|
parent = *p;
|
|
|
|
nd = rb_entry(parent, struct sp_node, nd);
|
|
|
|
if (new->start < nd->start)
|
|
|
|
p = &(*p)->rb_left;
|
|
|
|
else if (new->end > nd->end)
|
|
|
|
p = &(*p)->rb_right;
|
|
|
|
else
|
|
|
|
BUG();
|
|
|
|
}
|
|
|
|
rb_link_node(&new->nd, parent, p);
|
|
|
|
rb_insert_color(&new->nd, &sp->root);
|
2007-07-16 13:38:16 +07:00
|
|
|
pr_debug("inserting %lx-%lx: %d\n", new->start, new->end,
|
2005-04-17 05:20:36 +07:00
|
|
|
new->policy ? new->policy->policy : 0);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Find shared policy intersecting idx */
|
|
|
|
struct mempolicy *
|
|
|
|
mpol_shared_policy_lookup(struct shared_policy *sp, unsigned long idx)
|
|
|
|
{
|
|
|
|
struct mempolicy *pol = NULL;
|
|
|
|
struct sp_node *sn;
|
|
|
|
|
|
|
|
if (!sp->root.rb_node)
|
|
|
|
return NULL;
|
|
|
|
spin_lock(&sp->lock);
|
|
|
|
sn = sp_lookup(sp, idx, idx+1);
|
|
|
|
if (sn) {
|
|
|
|
mpol_get(sn->policy);
|
|
|
|
pol = sn->policy;
|
|
|
|
}
|
|
|
|
spin_unlock(&sp->lock);
|
|
|
|
return pol;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void sp_delete(struct shared_policy *sp, struct sp_node *n)
|
|
|
|
{
|
2007-07-16 13:38:16 +07:00
|
|
|
pr_debug("deleting %lx-l%lx\n", n->start, n->end);
|
2005-04-17 05:20:36 +07:00
|
|
|
rb_erase(&n->nd, &sp->root);
|
|
|
|
mpol_free(n->policy);
|
|
|
|
kmem_cache_free(sn_cache, n);
|
|
|
|
}
|
|
|
|
|
2007-10-16 15:26:26 +07:00
|
|
|
static struct sp_node *sp_alloc(unsigned long start, unsigned long end,
|
|
|
|
struct mempolicy *pol)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
struct sp_node *n = kmem_cache_alloc(sn_cache, GFP_KERNEL);
|
|
|
|
|
|
|
|
if (!n)
|
|
|
|
return NULL;
|
|
|
|
n->start = start;
|
|
|
|
n->end = end;
|
|
|
|
mpol_get(pol);
|
|
|
|
n->policy = pol;
|
|
|
|
return n;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Replace a policy range. */
|
|
|
|
static int shared_policy_replace(struct shared_policy *sp, unsigned long start,
|
|
|
|
unsigned long end, struct sp_node *new)
|
|
|
|
{
|
|
|
|
struct sp_node *n, *new2 = NULL;
|
|
|
|
|
|
|
|
restart:
|
|
|
|
spin_lock(&sp->lock);
|
|
|
|
n = sp_lookup(sp, start, end);
|
|
|
|
/* Take care of old policies in the same range. */
|
|
|
|
while (n && n->start < end) {
|
|
|
|
struct rb_node *next = rb_next(&n->nd);
|
|
|
|
if (n->start >= start) {
|
|
|
|
if (n->end <= end)
|
|
|
|
sp_delete(sp, n);
|
|
|
|
else
|
|
|
|
n->start = end;
|
|
|
|
} else {
|
|
|
|
/* Old policy spanning whole new range. */
|
|
|
|
if (n->end > end) {
|
|
|
|
if (!new2) {
|
|
|
|
spin_unlock(&sp->lock);
|
|
|
|
new2 = sp_alloc(end, n->end, n->policy);
|
|
|
|
if (!new2)
|
|
|
|
return -ENOMEM;
|
|
|
|
goto restart;
|
|
|
|
}
|
|
|
|
n->end = start;
|
|
|
|
sp_insert(sp, new2);
|
|
|
|
new2 = NULL;
|
|
|
|
break;
|
|
|
|
} else
|
|
|
|
n->end = start;
|
|
|
|
}
|
|
|
|
if (!next)
|
|
|
|
break;
|
|
|
|
n = rb_entry(next, struct sp_node, nd);
|
|
|
|
}
|
|
|
|
if (new)
|
|
|
|
sp_insert(sp, new);
|
|
|
|
spin_unlock(&sp->lock);
|
|
|
|
if (new2) {
|
|
|
|
mpol_free(new2->policy);
|
|
|
|
kmem_cache_free(sn_cache, new2);
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2006-01-15 04:20:48 +07:00
|
|
|
void mpol_shared_policy_init(struct shared_policy *info, int policy,
|
|
|
|
nodemask_t *policy_nodes)
|
|
|
|
{
|
|
|
|
info->root = RB_ROOT;
|
|
|
|
spin_lock_init(&info->lock);
|
|
|
|
|
|
|
|
if (policy != MPOL_DEFAULT) {
|
|
|
|
struct mempolicy *newpol;
|
|
|
|
|
|
|
|
/* Falls back to MPOL_DEFAULT on any error */
|
|
|
|
newpol = mpol_new(policy, policy_nodes);
|
|
|
|
if (!IS_ERR(newpol)) {
|
|
|
|
/* Create pseudo-vma that contains just the policy */
|
|
|
|
struct vm_area_struct pvma;
|
|
|
|
|
|
|
|
memset(&pvma, 0, sizeof(struct vm_area_struct));
|
|
|
|
/* Policy covers entire file */
|
|
|
|
pvma.vm_end = TASK_SIZE;
|
|
|
|
mpol_set_shared_policy(info, &pvma, newpol);
|
|
|
|
mpol_free(newpol);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
int mpol_set_shared_policy(struct shared_policy *info,
|
|
|
|
struct vm_area_struct *vma, struct mempolicy *npol)
|
|
|
|
{
|
|
|
|
int err;
|
|
|
|
struct sp_node *new = NULL;
|
|
|
|
unsigned long sz = vma_pages(vma);
|
|
|
|
|
2007-07-16 13:38:16 +07:00
|
|
|
pr_debug("set_shared_policy %lx sz %lu %d %lx\n",
|
2005-04-17 05:20:36 +07:00
|
|
|
vma->vm_pgoff,
|
|
|
|
sz, npol? npol->policy : -1,
|
2007-07-16 13:38:16 +07:00
|
|
|
npol ? nodes_addr(npol->v.nodes)[0] : -1);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
if (npol) {
|
|
|
|
new = sp_alloc(vma->vm_pgoff, vma->vm_pgoff + sz, npol);
|
|
|
|
if (!new)
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
err = shared_policy_replace(info, vma->vm_pgoff, vma->vm_pgoff+sz, new);
|
|
|
|
if (err && new)
|
|
|
|
kmem_cache_free(sn_cache, new);
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Free a backing policy store on inode delete. */
|
|
|
|
void mpol_free_shared_policy(struct shared_policy *p)
|
|
|
|
{
|
|
|
|
struct sp_node *n;
|
|
|
|
struct rb_node *next;
|
|
|
|
|
|
|
|
if (!p->root.rb_node)
|
|
|
|
return;
|
|
|
|
spin_lock(&p->lock);
|
|
|
|
next = rb_first(&p->root);
|
|
|
|
while (next) {
|
|
|
|
n = rb_entry(next, struct sp_node, nd);
|
|
|
|
next = rb_next(&n->nd);
|
2005-07-28 01:43:50 +07:00
|
|
|
rb_erase(&n->nd, &p->root);
|
2005-04-17 05:20:36 +07:00
|
|
|
mpol_free(n->policy);
|
|
|
|
kmem_cache_free(sn_cache, n);
|
|
|
|
}
|
|
|
|
spin_unlock(&p->lock);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* assumes fs == KERNEL_DS */
|
|
|
|
void __init numa_policy_init(void)
|
|
|
|
{
|
2007-07-16 13:38:15 +07:00
|
|
|
nodemask_t interleave_nodes;
|
|
|
|
unsigned long largest = 0;
|
|
|
|
int nid, prefer = 0;
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
policy_cache = kmem_cache_create("numa_policy",
|
|
|
|
sizeof(struct mempolicy),
|
2007-07-20 08:11:58 +07:00
|
|
|
0, SLAB_PANIC, NULL);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
sn_cache = kmem_cache_create("shared_policy_node",
|
|
|
|
sizeof(struct sp_node),
|
2007-07-20 08:11:58 +07:00
|
|
|
0, SLAB_PANIC, NULL);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2007-07-16 13:38:15 +07:00
|
|
|
/*
|
|
|
|
* Set interleaving policy for system init. Interleaving is only
|
|
|
|
* enabled across suitably sized nodes (default is >= 16MB), or
|
|
|
|
* fall back to the largest node if they're all smaller.
|
|
|
|
*/
|
|
|
|
nodes_clear(interleave_nodes);
|
2007-10-16 15:25:35 +07:00
|
|
|
for_each_node_state(nid, N_HIGH_MEMORY) {
|
2007-07-16 13:38:15 +07:00
|
|
|
unsigned long total_pages = node_present_pages(nid);
|
|
|
|
|
|
|
|
/* Preserve the largest node */
|
|
|
|
if (largest < total_pages) {
|
|
|
|
largest = total_pages;
|
|
|
|
prefer = nid;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Interleave this node? */
|
|
|
|
if ((total_pages << PAGE_SHIFT) >= (16 << 20))
|
|
|
|
node_set(nid, interleave_nodes);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* All too small, use the largest */
|
|
|
|
if (unlikely(nodes_empty(interleave_nodes)))
|
|
|
|
node_set(prefer, interleave_nodes);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2007-07-16 13:38:15 +07:00
|
|
|
if (do_set_mempolicy(MPOL_INTERLEAVE, &interleave_nodes))
|
2005-04-17 05:20:36 +07:00
|
|
|
printk("numa_policy_init: interleaving failed\n");
|
|
|
|
}
|
|
|
|
|
2005-10-30 08:16:59 +07:00
|
|
|
/* Reset policy of current process to default */
|
2005-04-17 05:20:36 +07:00
|
|
|
void numa_default_policy(void)
|
|
|
|
{
|
2005-10-30 08:16:59 +07:00
|
|
|
do_set_mempolicy(MPOL_DEFAULT, NULL);
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
[PATCH] cpusets: automatic numa mempolicy rebinding
This patch automatically updates a tasks NUMA mempolicy when its cpuset
memory placement changes. It does so within the context of the task,
without any need to support low level external mempolicy manipulation.
If a system is not using cpusets, or if running on a system with just the
root (all-encompassing) cpuset, then this remap is a no-op. Only when a
task is moved between cpusets, or a cpusets memory placement is changed
does the following apply. Otherwise, the main routine below,
rebind_policy() is not even called.
When mixing cpusets, scheduler affinity, and NUMA mempolicies, the
essential role of cpusets is to place jobs (several related tasks) on a set
of CPUs and Memory Nodes, the essential role of sched_setaffinity is to
manage a jobs processor placement within its allowed cpuset, and the
essential role of NUMA mempolicy (mbind, set_mempolicy) is to manage a jobs
memory placement within its allowed cpuset.
However, CPU affinity and NUMA memory placement are managed within the
kernel using absolute system wide numbering, not cpuset relative numbering.
This is ok until a job is migrated to a different cpuset, or what's the
same, a jobs cpuset is moved to different CPUs and Memory Nodes.
Then the CPU affinity and NUMA memory placement of the tasks in the job
need to be updated, to preserve their cpuset-relative position. This can
be done for CPU affinity using sched_setaffinity() from user code, as one
task can modify anothers CPU affinity. This cannot be done from an
external task for NUMA memory placement, as that can only be modified in
the context of the task using it.
However, it easy enough to remap a tasks NUMA mempolicy automatically when
a task is migrated, using the existing cpuset mechanism to trigger a
refresh of a tasks memory placement after its cpuset has changed. All that
is needed is the old and new nodemask, and notice to the task that it needs
to rebind its mempolicy. The tasks mems_allowed has the old mask, the
tasks cpuset has the new mask, and the existing
cpuset_update_current_mems_allowed() mechanism provides the notice. The
bitmap/cpumask/nodemask remap operators provide the cpuset relative
calculations.
This patch leaves open a couple of issues:
1) Updating vma and shmfs/tmpfs/hugetlbfs memory policies:
These mempolicies may reference nodes outside of those allowed to
the current task by its cpuset. Tasks are migrated as part of jobs,
which reside on what might be several cpusets in a subtree. When such
a job is migrated, all NUMA memory policy references to nodes within
that cpuset subtree should be translated, and references to any nodes
outside that subtree should be left untouched. A future patch will
provide the cpuset mechanism needed to mark such subtrees. With that
patch, we will be able to correctly migrate these other memory policies
across a job migration.
2) Updating cpuset, affinity and memory policies in user space:
This is harder. Any placement state stored in user space using
system-wide numbering will be invalidated across a migration. More
work will be required to provide user code with a migration-safe means
to manage its cpuset relative placement, while preserving the current
API's that pass system wide numbers, not cpuset relative numbers across
the kernel-user boundary.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-31 06:02:36 +07:00
|
|
|
|
|
|
|
/* Migrate a policy to a different set of nodes */
|
2007-10-16 15:26:26 +07:00
|
|
|
static void mpol_rebind_policy(struct mempolicy *pol,
|
|
|
|
const nodemask_t *newmask)
|
[PATCH] cpusets: automatic numa mempolicy rebinding
This patch automatically updates a tasks NUMA mempolicy when its cpuset
memory placement changes. It does so within the context of the task,
without any need to support low level external mempolicy manipulation.
If a system is not using cpusets, or if running on a system with just the
root (all-encompassing) cpuset, then this remap is a no-op. Only when a
task is moved between cpusets, or a cpusets memory placement is changed
does the following apply. Otherwise, the main routine below,
rebind_policy() is not even called.
When mixing cpusets, scheduler affinity, and NUMA mempolicies, the
essential role of cpusets is to place jobs (several related tasks) on a set
of CPUs and Memory Nodes, the essential role of sched_setaffinity is to
manage a jobs processor placement within its allowed cpuset, and the
essential role of NUMA mempolicy (mbind, set_mempolicy) is to manage a jobs
memory placement within its allowed cpuset.
However, CPU affinity and NUMA memory placement are managed within the
kernel using absolute system wide numbering, not cpuset relative numbering.
This is ok until a job is migrated to a different cpuset, or what's the
same, a jobs cpuset is moved to different CPUs and Memory Nodes.
Then the CPU affinity and NUMA memory placement of the tasks in the job
need to be updated, to preserve their cpuset-relative position. This can
be done for CPU affinity using sched_setaffinity() from user code, as one
task can modify anothers CPU affinity. This cannot be done from an
external task for NUMA memory placement, as that can only be modified in
the context of the task using it.
However, it easy enough to remap a tasks NUMA mempolicy automatically when
a task is migrated, using the existing cpuset mechanism to trigger a
refresh of a tasks memory placement after its cpuset has changed. All that
is needed is the old and new nodemask, and notice to the task that it needs
to rebind its mempolicy. The tasks mems_allowed has the old mask, the
tasks cpuset has the new mask, and the existing
cpuset_update_current_mems_allowed() mechanism provides the notice. The
bitmap/cpumask/nodemask remap operators provide the cpuset relative
calculations.
This patch leaves open a couple of issues:
1) Updating vma and shmfs/tmpfs/hugetlbfs memory policies:
These mempolicies may reference nodes outside of those allowed to
the current task by its cpuset. Tasks are migrated as part of jobs,
which reside on what might be several cpusets in a subtree. When such
a job is migrated, all NUMA memory policy references to nodes within
that cpuset subtree should be translated, and references to any nodes
outside that subtree should be left untouched. A future patch will
provide the cpuset mechanism needed to mark such subtrees. With that
patch, we will be able to correctly migrate these other memory policies
across a job migration.
2) Updating cpuset, affinity and memory policies in user space:
This is harder. Any placement state stored in user space using
system-wide numbering will be invalidated across a migration. More
work will be required to provide user code with a migration-safe means
to manage its cpuset relative placement, while preserving the current
API's that pass system wide numbers, not cpuset relative numbers across
the kernel-user boundary.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-31 06:02:36 +07:00
|
|
|
{
|
[PATCH] cpuset: numa_policy_rebind cleanup
Cleanup, reorganize and make more robust the mempolicy.c code to rebind
mempolicies relative to the containing cpuset after a tasks memory placement
changes.
The real motivator for this cleanup patch is to lay more groundwork for the
upcoming patch to correctly rebind NUMA mempolicies that are attached to vma's
after the containing cpuset memory placement changes.
NUMA mempolicies are constrained by the cpuset their task is a member of.
When either (1) a task is moved to a different cpuset, or (2) the 'mems'
mems_allowed of a cpuset is changed, then the NUMA mempolicies have embedded
node numbers (for MPOL_BIND, MPOL_INTERLEAVE and MPOL_PREFERRED) that need to
be recalculated, relative to their new cpuset placement.
The old code used an unreliable method of determining what was the old
mems_allowed constraining the mempolicy. It just looked at the tasks
mems_allowed value. This sort of worked with the present code, that just
rebinds the -task- mempolicy, and leaves any -vma- mempolicies broken,
referring to the old nodes. But in an upcoming patch, the vma mempolicies
will be rebound as well. Then the order in which the various task and vma
mempolicies are updated will no longer be deterministic, and one can no longer
count on the task->mems_allowed holding the old value for as long as needed.
It's not even clear if the current code was guaranteed to work reliably for
task mempolicies.
So I added a mems_allowed field to each mempolicy, stating exactly what
mems_allowed the policy is relative to, and updated synchronously and reliably
anytime that the mempolicy is rebound.
Also removed a useless wrapper routine, numa_policy_rebind(), and had its
caller, cpuset_update_task_memory_state(), call directly to the rewritten
policy_rebind() routine, and made that rebind routine extern instead of
static, and added a "mpol_" prefix to its name, making it
mpol_rebind_policy().
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-08 16:01:56 +07:00
|
|
|
nodemask_t *mpolmask;
|
[PATCH] cpusets: automatic numa mempolicy rebinding
This patch automatically updates a tasks NUMA mempolicy when its cpuset
memory placement changes. It does so within the context of the task,
without any need to support low level external mempolicy manipulation.
If a system is not using cpusets, or if running on a system with just the
root (all-encompassing) cpuset, then this remap is a no-op. Only when a
task is moved between cpusets, or a cpusets memory placement is changed
does the following apply. Otherwise, the main routine below,
rebind_policy() is not even called.
When mixing cpusets, scheduler affinity, and NUMA mempolicies, the
essential role of cpusets is to place jobs (several related tasks) on a set
of CPUs and Memory Nodes, the essential role of sched_setaffinity is to
manage a jobs processor placement within its allowed cpuset, and the
essential role of NUMA mempolicy (mbind, set_mempolicy) is to manage a jobs
memory placement within its allowed cpuset.
However, CPU affinity and NUMA memory placement are managed within the
kernel using absolute system wide numbering, not cpuset relative numbering.
This is ok until a job is migrated to a different cpuset, or what's the
same, a jobs cpuset is moved to different CPUs and Memory Nodes.
Then the CPU affinity and NUMA memory placement of the tasks in the job
need to be updated, to preserve their cpuset-relative position. This can
be done for CPU affinity using sched_setaffinity() from user code, as one
task can modify anothers CPU affinity. This cannot be done from an
external task for NUMA memory placement, as that can only be modified in
the context of the task using it.
However, it easy enough to remap a tasks NUMA mempolicy automatically when
a task is migrated, using the existing cpuset mechanism to trigger a
refresh of a tasks memory placement after its cpuset has changed. All that
is needed is the old and new nodemask, and notice to the task that it needs
to rebind its mempolicy. The tasks mems_allowed has the old mask, the
tasks cpuset has the new mask, and the existing
cpuset_update_current_mems_allowed() mechanism provides the notice. The
bitmap/cpumask/nodemask remap operators provide the cpuset relative
calculations.
This patch leaves open a couple of issues:
1) Updating vma and shmfs/tmpfs/hugetlbfs memory policies:
These mempolicies may reference nodes outside of those allowed to
the current task by its cpuset. Tasks are migrated as part of jobs,
which reside on what might be several cpusets in a subtree. When such
a job is migrated, all NUMA memory policy references to nodes within
that cpuset subtree should be translated, and references to any nodes
outside that subtree should be left untouched. A future patch will
provide the cpuset mechanism needed to mark such subtrees. With that
patch, we will be able to correctly migrate these other memory policies
across a job migration.
2) Updating cpuset, affinity and memory policies in user space:
This is harder. Any placement state stored in user space using
system-wide numbering will be invalidated across a migration. More
work will be required to provide user code with a migration-safe means
to manage its cpuset relative placement, while preserving the current
API's that pass system wide numbers, not cpuset relative numbers across
the kernel-user boundary.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-31 06:02:36 +07:00
|
|
|
nodemask_t tmp;
|
|
|
|
|
|
|
|
if (!pol)
|
|
|
|
return;
|
[PATCH] cpuset: numa_policy_rebind cleanup
Cleanup, reorganize and make more robust the mempolicy.c code to rebind
mempolicies relative to the containing cpuset after a tasks memory placement
changes.
The real motivator for this cleanup patch is to lay more groundwork for the
upcoming patch to correctly rebind NUMA mempolicies that are attached to vma's
after the containing cpuset memory placement changes.
NUMA mempolicies are constrained by the cpuset their task is a member of.
When either (1) a task is moved to a different cpuset, or (2) the 'mems'
mems_allowed of a cpuset is changed, then the NUMA mempolicies have embedded
node numbers (for MPOL_BIND, MPOL_INTERLEAVE and MPOL_PREFERRED) that need to
be recalculated, relative to their new cpuset placement.
The old code used an unreliable method of determining what was the old
mems_allowed constraining the mempolicy. It just looked at the tasks
mems_allowed value. This sort of worked with the present code, that just
rebinds the -task- mempolicy, and leaves any -vma- mempolicies broken,
referring to the old nodes. But in an upcoming patch, the vma mempolicies
will be rebound as well. Then the order in which the various task and vma
mempolicies are updated will no longer be deterministic, and one can no longer
count on the task->mems_allowed holding the old value for as long as needed.
It's not even clear if the current code was guaranteed to work reliably for
task mempolicies.
So I added a mems_allowed field to each mempolicy, stating exactly what
mems_allowed the policy is relative to, and updated synchronously and reliably
anytime that the mempolicy is rebound.
Also removed a useless wrapper routine, numa_policy_rebind(), and had its
caller, cpuset_update_task_memory_state(), call directly to the rewritten
policy_rebind() routine, and made that rebind routine extern instead of
static, and added a "mpol_" prefix to its name, making it
mpol_rebind_policy().
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-08 16:01:56 +07:00
|
|
|
mpolmask = &pol->cpuset_mems_allowed;
|
|
|
|
if (nodes_equal(*mpolmask, *newmask))
|
|
|
|
return;
|
[PATCH] cpusets: automatic numa mempolicy rebinding
This patch automatically updates a tasks NUMA mempolicy when its cpuset
memory placement changes. It does so within the context of the task,
without any need to support low level external mempolicy manipulation.
If a system is not using cpusets, or if running on a system with just the
root (all-encompassing) cpuset, then this remap is a no-op. Only when a
task is moved between cpusets, or a cpusets memory placement is changed
does the following apply. Otherwise, the main routine below,
rebind_policy() is not even called.
When mixing cpusets, scheduler affinity, and NUMA mempolicies, the
essential role of cpusets is to place jobs (several related tasks) on a set
of CPUs and Memory Nodes, the essential role of sched_setaffinity is to
manage a jobs processor placement within its allowed cpuset, and the
essential role of NUMA mempolicy (mbind, set_mempolicy) is to manage a jobs
memory placement within its allowed cpuset.
However, CPU affinity and NUMA memory placement are managed within the
kernel using absolute system wide numbering, not cpuset relative numbering.
This is ok until a job is migrated to a different cpuset, or what's the
same, a jobs cpuset is moved to different CPUs and Memory Nodes.
Then the CPU affinity and NUMA memory placement of the tasks in the job
need to be updated, to preserve their cpuset-relative position. This can
be done for CPU affinity using sched_setaffinity() from user code, as one
task can modify anothers CPU affinity. This cannot be done from an
external task for NUMA memory placement, as that can only be modified in
the context of the task using it.
However, it easy enough to remap a tasks NUMA mempolicy automatically when
a task is migrated, using the existing cpuset mechanism to trigger a
refresh of a tasks memory placement after its cpuset has changed. All that
is needed is the old and new nodemask, and notice to the task that it needs
to rebind its mempolicy. The tasks mems_allowed has the old mask, the
tasks cpuset has the new mask, and the existing
cpuset_update_current_mems_allowed() mechanism provides the notice. The
bitmap/cpumask/nodemask remap operators provide the cpuset relative
calculations.
This patch leaves open a couple of issues:
1) Updating vma and shmfs/tmpfs/hugetlbfs memory policies:
These mempolicies may reference nodes outside of those allowed to
the current task by its cpuset. Tasks are migrated as part of jobs,
which reside on what might be several cpusets in a subtree. When such
a job is migrated, all NUMA memory policy references to nodes within
that cpuset subtree should be translated, and references to any nodes
outside that subtree should be left untouched. A future patch will
provide the cpuset mechanism needed to mark such subtrees. With that
patch, we will be able to correctly migrate these other memory policies
across a job migration.
2) Updating cpuset, affinity and memory policies in user space:
This is harder. Any placement state stored in user space using
system-wide numbering will be invalidated across a migration. More
work will be required to provide user code with a migration-safe means
to manage its cpuset relative placement, while preserving the current
API's that pass system wide numbers, not cpuset relative numbers across
the kernel-user boundary.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-31 06:02:36 +07:00
|
|
|
|
|
|
|
switch (pol->policy) {
|
|
|
|
case MPOL_DEFAULT:
|
|
|
|
break;
|
|
|
|
case MPOL_INTERLEAVE:
|
[PATCH] cpuset: numa_policy_rebind cleanup
Cleanup, reorganize and make more robust the mempolicy.c code to rebind
mempolicies relative to the containing cpuset after a tasks memory placement
changes.
The real motivator for this cleanup patch is to lay more groundwork for the
upcoming patch to correctly rebind NUMA mempolicies that are attached to vma's
after the containing cpuset memory placement changes.
NUMA mempolicies are constrained by the cpuset their task is a member of.
When either (1) a task is moved to a different cpuset, or (2) the 'mems'
mems_allowed of a cpuset is changed, then the NUMA mempolicies have embedded
node numbers (for MPOL_BIND, MPOL_INTERLEAVE and MPOL_PREFERRED) that need to
be recalculated, relative to their new cpuset placement.
The old code used an unreliable method of determining what was the old
mems_allowed constraining the mempolicy. It just looked at the tasks
mems_allowed value. This sort of worked with the present code, that just
rebinds the -task- mempolicy, and leaves any -vma- mempolicies broken,
referring to the old nodes. But in an upcoming patch, the vma mempolicies
will be rebound as well. Then the order in which the various task and vma
mempolicies are updated will no longer be deterministic, and one can no longer
count on the task->mems_allowed holding the old value for as long as needed.
It's not even clear if the current code was guaranteed to work reliably for
task mempolicies.
So I added a mems_allowed field to each mempolicy, stating exactly what
mems_allowed the policy is relative to, and updated synchronously and reliably
anytime that the mempolicy is rebound.
Also removed a useless wrapper routine, numa_policy_rebind(), and had its
caller, cpuset_update_task_memory_state(), call directly to the rewritten
policy_rebind() routine, and made that rebind routine extern instead of
static, and added a "mpol_" prefix to its name, making it
mpol_rebind_policy().
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-08 16:01:56 +07:00
|
|
|
nodes_remap(tmp, pol->v.nodes, *mpolmask, *newmask);
|
[PATCH] cpusets: automatic numa mempolicy rebinding
This patch automatically updates a tasks NUMA mempolicy when its cpuset
memory placement changes. It does so within the context of the task,
without any need to support low level external mempolicy manipulation.
If a system is not using cpusets, or if running on a system with just the
root (all-encompassing) cpuset, then this remap is a no-op. Only when a
task is moved between cpusets, or a cpusets memory placement is changed
does the following apply. Otherwise, the main routine below,
rebind_policy() is not even called.
When mixing cpusets, scheduler affinity, and NUMA mempolicies, the
essential role of cpusets is to place jobs (several related tasks) on a set
of CPUs and Memory Nodes, the essential role of sched_setaffinity is to
manage a jobs processor placement within its allowed cpuset, and the
essential role of NUMA mempolicy (mbind, set_mempolicy) is to manage a jobs
memory placement within its allowed cpuset.
However, CPU affinity and NUMA memory placement are managed within the
kernel using absolute system wide numbering, not cpuset relative numbering.
This is ok until a job is migrated to a different cpuset, or what's the
same, a jobs cpuset is moved to different CPUs and Memory Nodes.
Then the CPU affinity and NUMA memory placement of the tasks in the job
need to be updated, to preserve their cpuset-relative position. This can
be done for CPU affinity using sched_setaffinity() from user code, as one
task can modify anothers CPU affinity. This cannot be done from an
external task for NUMA memory placement, as that can only be modified in
the context of the task using it.
However, it easy enough to remap a tasks NUMA mempolicy automatically when
a task is migrated, using the existing cpuset mechanism to trigger a
refresh of a tasks memory placement after its cpuset has changed. All that
is needed is the old and new nodemask, and notice to the task that it needs
to rebind its mempolicy. The tasks mems_allowed has the old mask, the
tasks cpuset has the new mask, and the existing
cpuset_update_current_mems_allowed() mechanism provides the notice. The
bitmap/cpumask/nodemask remap operators provide the cpuset relative
calculations.
This patch leaves open a couple of issues:
1) Updating vma and shmfs/tmpfs/hugetlbfs memory policies:
These mempolicies may reference nodes outside of those allowed to
the current task by its cpuset. Tasks are migrated as part of jobs,
which reside on what might be several cpusets in a subtree. When such
a job is migrated, all NUMA memory policy references to nodes within
that cpuset subtree should be translated, and references to any nodes
outside that subtree should be left untouched. A future patch will
provide the cpuset mechanism needed to mark such subtrees. With that
patch, we will be able to correctly migrate these other memory policies
across a job migration.
2) Updating cpuset, affinity and memory policies in user space:
This is harder. Any placement state stored in user space using
system-wide numbering will be invalidated across a migration. More
work will be required to provide user code with a migration-safe means
to manage its cpuset relative placement, while preserving the current
API's that pass system wide numbers, not cpuset relative numbers across
the kernel-user boundary.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-31 06:02:36 +07:00
|
|
|
pol->v.nodes = tmp;
|
[PATCH] cpuset: numa_policy_rebind cleanup
Cleanup, reorganize and make more robust the mempolicy.c code to rebind
mempolicies relative to the containing cpuset after a tasks memory placement
changes.
The real motivator for this cleanup patch is to lay more groundwork for the
upcoming patch to correctly rebind NUMA mempolicies that are attached to vma's
after the containing cpuset memory placement changes.
NUMA mempolicies are constrained by the cpuset their task is a member of.
When either (1) a task is moved to a different cpuset, or (2) the 'mems'
mems_allowed of a cpuset is changed, then the NUMA mempolicies have embedded
node numbers (for MPOL_BIND, MPOL_INTERLEAVE and MPOL_PREFERRED) that need to
be recalculated, relative to their new cpuset placement.
The old code used an unreliable method of determining what was the old
mems_allowed constraining the mempolicy. It just looked at the tasks
mems_allowed value. This sort of worked with the present code, that just
rebinds the -task- mempolicy, and leaves any -vma- mempolicies broken,
referring to the old nodes. But in an upcoming patch, the vma mempolicies
will be rebound as well. Then the order in which the various task and vma
mempolicies are updated will no longer be deterministic, and one can no longer
count on the task->mems_allowed holding the old value for as long as needed.
It's not even clear if the current code was guaranteed to work reliably for
task mempolicies.
So I added a mems_allowed field to each mempolicy, stating exactly what
mems_allowed the policy is relative to, and updated synchronously and reliably
anytime that the mempolicy is rebound.
Also removed a useless wrapper routine, numa_policy_rebind(), and had its
caller, cpuset_update_task_memory_state(), call directly to the rewritten
policy_rebind() routine, and made that rebind routine extern instead of
static, and added a "mpol_" prefix to its name, making it
mpol_rebind_policy().
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-08 16:01:56 +07:00
|
|
|
*mpolmask = *newmask;
|
|
|
|
current->il_next = node_remap(current->il_next,
|
|
|
|
*mpolmask, *newmask);
|
[PATCH] cpusets: automatic numa mempolicy rebinding
This patch automatically updates a tasks NUMA mempolicy when its cpuset
memory placement changes. It does so within the context of the task,
without any need to support low level external mempolicy manipulation.
If a system is not using cpusets, or if running on a system with just the
root (all-encompassing) cpuset, then this remap is a no-op. Only when a
task is moved between cpusets, or a cpusets memory placement is changed
does the following apply. Otherwise, the main routine below,
rebind_policy() is not even called.
When mixing cpusets, scheduler affinity, and NUMA mempolicies, the
essential role of cpusets is to place jobs (several related tasks) on a set
of CPUs and Memory Nodes, the essential role of sched_setaffinity is to
manage a jobs processor placement within its allowed cpuset, and the
essential role of NUMA mempolicy (mbind, set_mempolicy) is to manage a jobs
memory placement within its allowed cpuset.
However, CPU affinity and NUMA memory placement are managed within the
kernel using absolute system wide numbering, not cpuset relative numbering.
This is ok until a job is migrated to a different cpuset, or what's the
same, a jobs cpuset is moved to different CPUs and Memory Nodes.
Then the CPU affinity and NUMA memory placement of the tasks in the job
need to be updated, to preserve their cpuset-relative position. This can
be done for CPU affinity using sched_setaffinity() from user code, as one
task can modify anothers CPU affinity. This cannot be done from an
external task for NUMA memory placement, as that can only be modified in
the context of the task using it.
However, it easy enough to remap a tasks NUMA mempolicy automatically when
a task is migrated, using the existing cpuset mechanism to trigger a
refresh of a tasks memory placement after its cpuset has changed. All that
is needed is the old and new nodemask, and notice to the task that it needs
to rebind its mempolicy. The tasks mems_allowed has the old mask, the
tasks cpuset has the new mask, and the existing
cpuset_update_current_mems_allowed() mechanism provides the notice. The
bitmap/cpumask/nodemask remap operators provide the cpuset relative
calculations.
This patch leaves open a couple of issues:
1) Updating vma and shmfs/tmpfs/hugetlbfs memory policies:
These mempolicies may reference nodes outside of those allowed to
the current task by its cpuset. Tasks are migrated as part of jobs,
which reside on what might be several cpusets in a subtree. When such
a job is migrated, all NUMA memory policy references to nodes within
that cpuset subtree should be translated, and references to any nodes
outside that subtree should be left untouched. A future patch will
provide the cpuset mechanism needed to mark such subtrees. With that
patch, we will be able to correctly migrate these other memory policies
across a job migration.
2) Updating cpuset, affinity and memory policies in user space:
This is harder. Any placement state stored in user space using
system-wide numbering will be invalidated across a migration. More
work will be required to provide user code with a migration-safe means
to manage its cpuset relative placement, while preserving the current
API's that pass system wide numbers, not cpuset relative numbers across
the kernel-user boundary.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-31 06:02:36 +07:00
|
|
|
break;
|
|
|
|
case MPOL_PREFERRED:
|
|
|
|
pol->v.preferred_node = node_remap(pol->v.preferred_node,
|
[PATCH] cpuset: numa_policy_rebind cleanup
Cleanup, reorganize and make more robust the mempolicy.c code to rebind
mempolicies relative to the containing cpuset after a tasks memory placement
changes.
The real motivator for this cleanup patch is to lay more groundwork for the
upcoming patch to correctly rebind NUMA mempolicies that are attached to vma's
after the containing cpuset memory placement changes.
NUMA mempolicies are constrained by the cpuset their task is a member of.
When either (1) a task is moved to a different cpuset, or (2) the 'mems'
mems_allowed of a cpuset is changed, then the NUMA mempolicies have embedded
node numbers (for MPOL_BIND, MPOL_INTERLEAVE and MPOL_PREFERRED) that need to
be recalculated, relative to their new cpuset placement.
The old code used an unreliable method of determining what was the old
mems_allowed constraining the mempolicy. It just looked at the tasks
mems_allowed value. This sort of worked with the present code, that just
rebinds the -task- mempolicy, and leaves any -vma- mempolicies broken,
referring to the old nodes. But in an upcoming patch, the vma mempolicies
will be rebound as well. Then the order in which the various task and vma
mempolicies are updated will no longer be deterministic, and one can no longer
count on the task->mems_allowed holding the old value for as long as needed.
It's not even clear if the current code was guaranteed to work reliably for
task mempolicies.
So I added a mems_allowed field to each mempolicy, stating exactly what
mems_allowed the policy is relative to, and updated synchronously and reliably
anytime that the mempolicy is rebound.
Also removed a useless wrapper routine, numa_policy_rebind(), and had its
caller, cpuset_update_task_memory_state(), call directly to the rewritten
policy_rebind() routine, and made that rebind routine extern instead of
static, and added a "mpol_" prefix to its name, making it
mpol_rebind_policy().
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-08 16:01:56 +07:00
|
|
|
*mpolmask, *newmask);
|
|
|
|
*mpolmask = *newmask;
|
[PATCH] cpusets: automatic numa mempolicy rebinding
This patch automatically updates a tasks NUMA mempolicy when its cpuset
memory placement changes. It does so within the context of the task,
without any need to support low level external mempolicy manipulation.
If a system is not using cpusets, or if running on a system with just the
root (all-encompassing) cpuset, then this remap is a no-op. Only when a
task is moved between cpusets, or a cpusets memory placement is changed
does the following apply. Otherwise, the main routine below,
rebind_policy() is not even called.
When mixing cpusets, scheduler affinity, and NUMA mempolicies, the
essential role of cpusets is to place jobs (several related tasks) on a set
of CPUs and Memory Nodes, the essential role of sched_setaffinity is to
manage a jobs processor placement within its allowed cpuset, and the
essential role of NUMA mempolicy (mbind, set_mempolicy) is to manage a jobs
memory placement within its allowed cpuset.
However, CPU affinity and NUMA memory placement are managed within the
kernel using absolute system wide numbering, not cpuset relative numbering.
This is ok until a job is migrated to a different cpuset, or what's the
same, a jobs cpuset is moved to different CPUs and Memory Nodes.
Then the CPU affinity and NUMA memory placement of the tasks in the job
need to be updated, to preserve their cpuset-relative position. This can
be done for CPU affinity using sched_setaffinity() from user code, as one
task can modify anothers CPU affinity. This cannot be done from an
external task for NUMA memory placement, as that can only be modified in
the context of the task using it.
However, it easy enough to remap a tasks NUMA mempolicy automatically when
a task is migrated, using the existing cpuset mechanism to trigger a
refresh of a tasks memory placement after its cpuset has changed. All that
is needed is the old and new nodemask, and notice to the task that it needs
to rebind its mempolicy. The tasks mems_allowed has the old mask, the
tasks cpuset has the new mask, and the existing
cpuset_update_current_mems_allowed() mechanism provides the notice. The
bitmap/cpumask/nodemask remap operators provide the cpuset relative
calculations.
This patch leaves open a couple of issues:
1) Updating vma and shmfs/tmpfs/hugetlbfs memory policies:
These mempolicies may reference nodes outside of those allowed to
the current task by its cpuset. Tasks are migrated as part of jobs,
which reside on what might be several cpusets in a subtree. When such
a job is migrated, all NUMA memory policy references to nodes within
that cpuset subtree should be translated, and references to any nodes
outside that subtree should be left untouched. A future patch will
provide the cpuset mechanism needed to mark such subtrees. With that
patch, we will be able to correctly migrate these other memory policies
across a job migration.
2) Updating cpuset, affinity and memory policies in user space:
This is harder. Any placement state stored in user space using
system-wide numbering will be invalidated across a migration. More
work will be required to provide user code with a migration-safe means
to manage its cpuset relative placement, while preserving the current
API's that pass system wide numbers, not cpuset relative numbers across
the kernel-user boundary.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-31 06:02:36 +07:00
|
|
|
break;
|
|
|
|
case MPOL_BIND: {
|
|
|
|
nodemask_t nodes;
|
|
|
|
struct zone **z;
|
|
|
|
struct zonelist *zonelist;
|
|
|
|
|
|
|
|
nodes_clear(nodes);
|
|
|
|
for (z = pol->v.zonelist->zones; *z; z++)
|
2006-09-26 13:31:55 +07:00
|
|
|
node_set(zone_to_nid(*z), nodes);
|
[PATCH] cpuset: numa_policy_rebind cleanup
Cleanup, reorganize and make more robust the mempolicy.c code to rebind
mempolicies relative to the containing cpuset after a tasks memory placement
changes.
The real motivator for this cleanup patch is to lay more groundwork for the
upcoming patch to correctly rebind NUMA mempolicies that are attached to vma's
after the containing cpuset memory placement changes.
NUMA mempolicies are constrained by the cpuset their task is a member of.
When either (1) a task is moved to a different cpuset, or (2) the 'mems'
mems_allowed of a cpuset is changed, then the NUMA mempolicies have embedded
node numbers (for MPOL_BIND, MPOL_INTERLEAVE and MPOL_PREFERRED) that need to
be recalculated, relative to their new cpuset placement.
The old code used an unreliable method of determining what was the old
mems_allowed constraining the mempolicy. It just looked at the tasks
mems_allowed value. This sort of worked with the present code, that just
rebinds the -task- mempolicy, and leaves any -vma- mempolicies broken,
referring to the old nodes. But in an upcoming patch, the vma mempolicies
will be rebound as well. Then the order in which the various task and vma
mempolicies are updated will no longer be deterministic, and one can no longer
count on the task->mems_allowed holding the old value for as long as needed.
It's not even clear if the current code was guaranteed to work reliably for
task mempolicies.
So I added a mems_allowed field to each mempolicy, stating exactly what
mems_allowed the policy is relative to, and updated synchronously and reliably
anytime that the mempolicy is rebound.
Also removed a useless wrapper routine, numa_policy_rebind(), and had its
caller, cpuset_update_task_memory_state(), call directly to the rewritten
policy_rebind() routine, and made that rebind routine extern instead of
static, and added a "mpol_" prefix to its name, making it
mpol_rebind_policy().
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-08 16:01:56 +07:00
|
|
|
nodes_remap(tmp, nodes, *mpolmask, *newmask);
|
[PATCH] cpusets: automatic numa mempolicy rebinding
This patch automatically updates a tasks NUMA mempolicy when its cpuset
memory placement changes. It does so within the context of the task,
without any need to support low level external mempolicy manipulation.
If a system is not using cpusets, or if running on a system with just the
root (all-encompassing) cpuset, then this remap is a no-op. Only when a
task is moved between cpusets, or a cpusets memory placement is changed
does the following apply. Otherwise, the main routine below,
rebind_policy() is not even called.
When mixing cpusets, scheduler affinity, and NUMA mempolicies, the
essential role of cpusets is to place jobs (several related tasks) on a set
of CPUs and Memory Nodes, the essential role of sched_setaffinity is to
manage a jobs processor placement within its allowed cpuset, and the
essential role of NUMA mempolicy (mbind, set_mempolicy) is to manage a jobs
memory placement within its allowed cpuset.
However, CPU affinity and NUMA memory placement are managed within the
kernel using absolute system wide numbering, not cpuset relative numbering.
This is ok until a job is migrated to a different cpuset, or what's the
same, a jobs cpuset is moved to different CPUs and Memory Nodes.
Then the CPU affinity and NUMA memory placement of the tasks in the job
need to be updated, to preserve their cpuset-relative position. This can
be done for CPU affinity using sched_setaffinity() from user code, as one
task can modify anothers CPU affinity. This cannot be done from an
external task for NUMA memory placement, as that can only be modified in
the context of the task using it.
However, it easy enough to remap a tasks NUMA mempolicy automatically when
a task is migrated, using the existing cpuset mechanism to trigger a
refresh of a tasks memory placement after its cpuset has changed. All that
is needed is the old and new nodemask, and notice to the task that it needs
to rebind its mempolicy. The tasks mems_allowed has the old mask, the
tasks cpuset has the new mask, and the existing
cpuset_update_current_mems_allowed() mechanism provides the notice. The
bitmap/cpumask/nodemask remap operators provide the cpuset relative
calculations.
This patch leaves open a couple of issues:
1) Updating vma and shmfs/tmpfs/hugetlbfs memory policies:
These mempolicies may reference nodes outside of those allowed to
the current task by its cpuset. Tasks are migrated as part of jobs,
which reside on what might be several cpusets in a subtree. When such
a job is migrated, all NUMA memory policy references to nodes within
that cpuset subtree should be translated, and references to any nodes
outside that subtree should be left untouched. A future patch will
provide the cpuset mechanism needed to mark such subtrees. With that
patch, we will be able to correctly migrate these other memory policies
across a job migration.
2) Updating cpuset, affinity and memory policies in user space:
This is harder. Any placement state stored in user space using
system-wide numbering will be invalidated across a migration. More
work will be required to provide user code with a migration-safe means
to manage its cpuset relative placement, while preserving the current
API's that pass system wide numbers, not cpuset relative numbers across
the kernel-user boundary.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-31 06:02:36 +07:00
|
|
|
nodes = tmp;
|
|
|
|
|
|
|
|
zonelist = bind_zonelist(&nodes);
|
|
|
|
|
|
|
|
/* If no mem, then zonelist is NULL and we keep old zonelist.
|
|
|
|
* If that old zonelist has no remaining mems_allowed nodes,
|
|
|
|
* then zonelist_policy() will "FALL THROUGH" to MPOL_DEFAULT.
|
|
|
|
*/
|
|
|
|
|
2007-02-21 04:57:49 +07:00
|
|
|
if (!IS_ERR(zonelist)) {
|
[PATCH] cpusets: automatic numa mempolicy rebinding
This patch automatically updates a tasks NUMA mempolicy when its cpuset
memory placement changes. It does so within the context of the task,
without any need to support low level external mempolicy manipulation.
If a system is not using cpusets, or if running on a system with just the
root (all-encompassing) cpuset, then this remap is a no-op. Only when a
task is moved between cpusets, or a cpusets memory placement is changed
does the following apply. Otherwise, the main routine below,
rebind_policy() is not even called.
When mixing cpusets, scheduler affinity, and NUMA mempolicies, the
essential role of cpusets is to place jobs (several related tasks) on a set
of CPUs and Memory Nodes, the essential role of sched_setaffinity is to
manage a jobs processor placement within its allowed cpuset, and the
essential role of NUMA mempolicy (mbind, set_mempolicy) is to manage a jobs
memory placement within its allowed cpuset.
However, CPU affinity and NUMA memory placement are managed within the
kernel using absolute system wide numbering, not cpuset relative numbering.
This is ok until a job is migrated to a different cpuset, or what's the
same, a jobs cpuset is moved to different CPUs and Memory Nodes.
Then the CPU affinity and NUMA memory placement of the tasks in the job
need to be updated, to preserve their cpuset-relative position. This can
be done for CPU affinity using sched_setaffinity() from user code, as one
task can modify anothers CPU affinity. This cannot be done from an
external task for NUMA memory placement, as that can only be modified in
the context of the task using it.
However, it easy enough to remap a tasks NUMA mempolicy automatically when
a task is migrated, using the existing cpuset mechanism to trigger a
refresh of a tasks memory placement after its cpuset has changed. All that
is needed is the old and new nodemask, and notice to the task that it needs
to rebind its mempolicy. The tasks mems_allowed has the old mask, the
tasks cpuset has the new mask, and the existing
cpuset_update_current_mems_allowed() mechanism provides the notice. The
bitmap/cpumask/nodemask remap operators provide the cpuset relative
calculations.
This patch leaves open a couple of issues:
1) Updating vma and shmfs/tmpfs/hugetlbfs memory policies:
These mempolicies may reference nodes outside of those allowed to
the current task by its cpuset. Tasks are migrated as part of jobs,
which reside on what might be several cpusets in a subtree. When such
a job is migrated, all NUMA memory policy references to nodes within
that cpuset subtree should be translated, and references to any nodes
outside that subtree should be left untouched. A future patch will
provide the cpuset mechanism needed to mark such subtrees. With that
patch, we will be able to correctly migrate these other memory policies
across a job migration.
2) Updating cpuset, affinity and memory policies in user space:
This is harder. Any placement state stored in user space using
system-wide numbering will be invalidated across a migration. More
work will be required to provide user code with a migration-safe means
to manage its cpuset relative placement, while preserving the current
API's that pass system wide numbers, not cpuset relative numbers across
the kernel-user boundary.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-31 06:02:36 +07:00
|
|
|
/* Good - got mem - substitute new zonelist */
|
|
|
|
kfree(pol->v.zonelist);
|
|
|
|
pol->v.zonelist = zonelist;
|
|
|
|
}
|
[PATCH] cpuset: numa_policy_rebind cleanup
Cleanup, reorganize and make more robust the mempolicy.c code to rebind
mempolicies relative to the containing cpuset after a tasks memory placement
changes.
The real motivator for this cleanup patch is to lay more groundwork for the
upcoming patch to correctly rebind NUMA mempolicies that are attached to vma's
after the containing cpuset memory placement changes.
NUMA mempolicies are constrained by the cpuset their task is a member of.
When either (1) a task is moved to a different cpuset, or (2) the 'mems'
mems_allowed of a cpuset is changed, then the NUMA mempolicies have embedded
node numbers (for MPOL_BIND, MPOL_INTERLEAVE and MPOL_PREFERRED) that need to
be recalculated, relative to their new cpuset placement.
The old code used an unreliable method of determining what was the old
mems_allowed constraining the mempolicy. It just looked at the tasks
mems_allowed value. This sort of worked with the present code, that just
rebinds the -task- mempolicy, and leaves any -vma- mempolicies broken,
referring to the old nodes. But in an upcoming patch, the vma mempolicies
will be rebound as well. Then the order in which the various task and vma
mempolicies are updated will no longer be deterministic, and one can no longer
count on the task->mems_allowed holding the old value for as long as needed.
It's not even clear if the current code was guaranteed to work reliably for
task mempolicies.
So I added a mems_allowed field to each mempolicy, stating exactly what
mems_allowed the policy is relative to, and updated synchronously and reliably
anytime that the mempolicy is rebound.
Also removed a useless wrapper routine, numa_policy_rebind(), and had its
caller, cpuset_update_task_memory_state(), call directly to the rewritten
policy_rebind() routine, and made that rebind routine extern instead of
static, and added a "mpol_" prefix to its name, making it
mpol_rebind_policy().
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-08 16:01:56 +07:00
|
|
|
*mpolmask = *newmask;
|
[PATCH] cpusets: automatic numa mempolicy rebinding
This patch automatically updates a tasks NUMA mempolicy when its cpuset
memory placement changes. It does so within the context of the task,
without any need to support low level external mempolicy manipulation.
If a system is not using cpusets, or if running on a system with just the
root (all-encompassing) cpuset, then this remap is a no-op. Only when a
task is moved between cpusets, or a cpusets memory placement is changed
does the following apply. Otherwise, the main routine below,
rebind_policy() is not even called.
When mixing cpusets, scheduler affinity, and NUMA mempolicies, the
essential role of cpusets is to place jobs (several related tasks) on a set
of CPUs and Memory Nodes, the essential role of sched_setaffinity is to
manage a jobs processor placement within its allowed cpuset, and the
essential role of NUMA mempolicy (mbind, set_mempolicy) is to manage a jobs
memory placement within its allowed cpuset.
However, CPU affinity and NUMA memory placement are managed within the
kernel using absolute system wide numbering, not cpuset relative numbering.
This is ok until a job is migrated to a different cpuset, or what's the
same, a jobs cpuset is moved to different CPUs and Memory Nodes.
Then the CPU affinity and NUMA memory placement of the tasks in the job
need to be updated, to preserve their cpuset-relative position. This can
be done for CPU affinity using sched_setaffinity() from user code, as one
task can modify anothers CPU affinity. This cannot be done from an
external task for NUMA memory placement, as that can only be modified in
the context of the task using it.
However, it easy enough to remap a tasks NUMA mempolicy automatically when
a task is migrated, using the existing cpuset mechanism to trigger a
refresh of a tasks memory placement after its cpuset has changed. All that
is needed is the old and new nodemask, and notice to the task that it needs
to rebind its mempolicy. The tasks mems_allowed has the old mask, the
tasks cpuset has the new mask, and the existing
cpuset_update_current_mems_allowed() mechanism provides the notice. The
bitmap/cpumask/nodemask remap operators provide the cpuset relative
calculations.
This patch leaves open a couple of issues:
1) Updating vma and shmfs/tmpfs/hugetlbfs memory policies:
These mempolicies may reference nodes outside of those allowed to
the current task by its cpuset. Tasks are migrated as part of jobs,
which reside on what might be several cpusets in a subtree. When such
a job is migrated, all NUMA memory policy references to nodes within
that cpuset subtree should be translated, and references to any nodes
outside that subtree should be left untouched. A future patch will
provide the cpuset mechanism needed to mark such subtrees. With that
patch, we will be able to correctly migrate these other memory policies
across a job migration.
2) Updating cpuset, affinity and memory policies in user space:
This is harder. Any placement state stored in user space using
system-wide numbering will be invalidated across a migration. More
work will be required to provide user code with a migration-safe means
to manage its cpuset relative placement, while preserving the current
API's that pass system wide numbers, not cpuset relative numbers across
the kernel-user boundary.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-31 06:02:36 +07:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
default:
|
|
|
|
BUG();
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
[PATCH] cpuset: numa_policy_rebind cleanup
Cleanup, reorganize and make more robust the mempolicy.c code to rebind
mempolicies relative to the containing cpuset after a tasks memory placement
changes.
The real motivator for this cleanup patch is to lay more groundwork for the
upcoming patch to correctly rebind NUMA mempolicies that are attached to vma's
after the containing cpuset memory placement changes.
NUMA mempolicies are constrained by the cpuset their task is a member of.
When either (1) a task is moved to a different cpuset, or (2) the 'mems'
mems_allowed of a cpuset is changed, then the NUMA mempolicies have embedded
node numbers (for MPOL_BIND, MPOL_INTERLEAVE and MPOL_PREFERRED) that need to
be recalculated, relative to their new cpuset placement.
The old code used an unreliable method of determining what was the old
mems_allowed constraining the mempolicy. It just looked at the tasks
mems_allowed value. This sort of worked with the present code, that just
rebinds the -task- mempolicy, and leaves any -vma- mempolicies broken,
referring to the old nodes. But in an upcoming patch, the vma mempolicies
will be rebound as well. Then the order in which the various task and vma
mempolicies are updated will no longer be deterministic, and one can no longer
count on the task->mems_allowed holding the old value for as long as needed.
It's not even clear if the current code was guaranteed to work reliably for
task mempolicies.
So I added a mems_allowed field to each mempolicy, stating exactly what
mems_allowed the policy is relative to, and updated synchronously and reliably
anytime that the mempolicy is rebound.
Also removed a useless wrapper routine, numa_policy_rebind(), and had its
caller, cpuset_update_task_memory_state(), call directly to the rewritten
policy_rebind() routine, and made that rebind routine extern instead of
static, and added a "mpol_" prefix to its name, making it
mpol_rebind_policy().
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-08 16:01:56 +07:00
|
|
|
* Wrapper for mpol_rebind_policy() that just requires task
|
|
|
|
* pointer, and updates task mempolicy.
|
[PATCH] cpusets: automatic numa mempolicy rebinding
This patch automatically updates a tasks NUMA mempolicy when its cpuset
memory placement changes. It does so within the context of the task,
without any need to support low level external mempolicy manipulation.
If a system is not using cpusets, or if running on a system with just the
root (all-encompassing) cpuset, then this remap is a no-op. Only when a
task is moved between cpusets, or a cpusets memory placement is changed
does the following apply. Otherwise, the main routine below,
rebind_policy() is not even called.
When mixing cpusets, scheduler affinity, and NUMA mempolicies, the
essential role of cpusets is to place jobs (several related tasks) on a set
of CPUs and Memory Nodes, the essential role of sched_setaffinity is to
manage a jobs processor placement within its allowed cpuset, and the
essential role of NUMA mempolicy (mbind, set_mempolicy) is to manage a jobs
memory placement within its allowed cpuset.
However, CPU affinity and NUMA memory placement are managed within the
kernel using absolute system wide numbering, not cpuset relative numbering.
This is ok until a job is migrated to a different cpuset, or what's the
same, a jobs cpuset is moved to different CPUs and Memory Nodes.
Then the CPU affinity and NUMA memory placement of the tasks in the job
need to be updated, to preserve their cpuset-relative position. This can
be done for CPU affinity using sched_setaffinity() from user code, as one
task can modify anothers CPU affinity. This cannot be done from an
external task for NUMA memory placement, as that can only be modified in
the context of the task using it.
However, it easy enough to remap a tasks NUMA mempolicy automatically when
a task is migrated, using the existing cpuset mechanism to trigger a
refresh of a tasks memory placement after its cpuset has changed. All that
is needed is the old and new nodemask, and notice to the task that it needs
to rebind its mempolicy. The tasks mems_allowed has the old mask, the
tasks cpuset has the new mask, and the existing
cpuset_update_current_mems_allowed() mechanism provides the notice. The
bitmap/cpumask/nodemask remap operators provide the cpuset relative
calculations.
This patch leaves open a couple of issues:
1) Updating vma and shmfs/tmpfs/hugetlbfs memory policies:
These mempolicies may reference nodes outside of those allowed to
the current task by its cpuset. Tasks are migrated as part of jobs,
which reside on what might be several cpusets in a subtree. When such
a job is migrated, all NUMA memory policy references to nodes within
that cpuset subtree should be translated, and references to any nodes
outside that subtree should be left untouched. A future patch will
provide the cpuset mechanism needed to mark such subtrees. With that
patch, we will be able to correctly migrate these other memory policies
across a job migration.
2) Updating cpuset, affinity and memory policies in user space:
This is harder. Any placement state stored in user space using
system-wide numbering will be invalidated across a migration. More
work will be required to provide user code with a migration-safe means
to manage its cpuset relative placement, while preserving the current
API's that pass system wide numbers, not cpuset relative numbers across
the kernel-user boundary.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-31 06:02:36 +07:00
|
|
|
*/
|
[PATCH] cpuset: numa_policy_rebind cleanup
Cleanup, reorganize and make more robust the mempolicy.c code to rebind
mempolicies relative to the containing cpuset after a tasks memory placement
changes.
The real motivator for this cleanup patch is to lay more groundwork for the
upcoming patch to correctly rebind NUMA mempolicies that are attached to vma's
after the containing cpuset memory placement changes.
NUMA mempolicies are constrained by the cpuset their task is a member of.
When either (1) a task is moved to a different cpuset, or (2) the 'mems'
mems_allowed of a cpuset is changed, then the NUMA mempolicies have embedded
node numbers (for MPOL_BIND, MPOL_INTERLEAVE and MPOL_PREFERRED) that need to
be recalculated, relative to their new cpuset placement.
The old code used an unreliable method of determining what was the old
mems_allowed constraining the mempolicy. It just looked at the tasks
mems_allowed value. This sort of worked with the present code, that just
rebinds the -task- mempolicy, and leaves any -vma- mempolicies broken,
referring to the old nodes. But in an upcoming patch, the vma mempolicies
will be rebound as well. Then the order in which the various task and vma
mempolicies are updated will no longer be deterministic, and one can no longer
count on the task->mems_allowed holding the old value for as long as needed.
It's not even clear if the current code was guaranteed to work reliably for
task mempolicies.
So I added a mems_allowed field to each mempolicy, stating exactly what
mems_allowed the policy is relative to, and updated synchronously and reliably
anytime that the mempolicy is rebound.
Also removed a useless wrapper routine, numa_policy_rebind(), and had its
caller, cpuset_update_task_memory_state(), call directly to the rewritten
policy_rebind() routine, and made that rebind routine extern instead of
static, and added a "mpol_" prefix to its name, making it
mpol_rebind_policy().
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-08 16:01:56 +07:00
|
|
|
|
|
|
|
void mpol_rebind_task(struct task_struct *tsk, const nodemask_t *new)
|
[PATCH] cpusets: automatic numa mempolicy rebinding
This patch automatically updates a tasks NUMA mempolicy when its cpuset
memory placement changes. It does so within the context of the task,
without any need to support low level external mempolicy manipulation.
If a system is not using cpusets, or if running on a system with just the
root (all-encompassing) cpuset, then this remap is a no-op. Only when a
task is moved between cpusets, or a cpusets memory placement is changed
does the following apply. Otherwise, the main routine below,
rebind_policy() is not even called.
When mixing cpusets, scheduler affinity, and NUMA mempolicies, the
essential role of cpusets is to place jobs (several related tasks) on a set
of CPUs and Memory Nodes, the essential role of sched_setaffinity is to
manage a jobs processor placement within its allowed cpuset, and the
essential role of NUMA mempolicy (mbind, set_mempolicy) is to manage a jobs
memory placement within its allowed cpuset.
However, CPU affinity and NUMA memory placement are managed within the
kernel using absolute system wide numbering, not cpuset relative numbering.
This is ok until a job is migrated to a different cpuset, or what's the
same, a jobs cpuset is moved to different CPUs and Memory Nodes.
Then the CPU affinity and NUMA memory placement of the tasks in the job
need to be updated, to preserve their cpuset-relative position. This can
be done for CPU affinity using sched_setaffinity() from user code, as one
task can modify anothers CPU affinity. This cannot be done from an
external task for NUMA memory placement, as that can only be modified in
the context of the task using it.
However, it easy enough to remap a tasks NUMA mempolicy automatically when
a task is migrated, using the existing cpuset mechanism to trigger a
refresh of a tasks memory placement after its cpuset has changed. All that
is needed is the old and new nodemask, and notice to the task that it needs
to rebind its mempolicy. The tasks mems_allowed has the old mask, the
tasks cpuset has the new mask, and the existing
cpuset_update_current_mems_allowed() mechanism provides the notice. The
bitmap/cpumask/nodemask remap operators provide the cpuset relative
calculations.
This patch leaves open a couple of issues:
1) Updating vma and shmfs/tmpfs/hugetlbfs memory policies:
These mempolicies may reference nodes outside of those allowed to
the current task by its cpuset. Tasks are migrated as part of jobs,
which reside on what might be several cpusets in a subtree. When such
a job is migrated, all NUMA memory policy references to nodes within
that cpuset subtree should be translated, and references to any nodes
outside that subtree should be left untouched. A future patch will
provide the cpuset mechanism needed to mark such subtrees. With that
patch, we will be able to correctly migrate these other memory policies
across a job migration.
2) Updating cpuset, affinity and memory policies in user space:
This is harder. Any placement state stored in user space using
system-wide numbering will be invalidated across a migration. More
work will be required to provide user code with a migration-safe means
to manage its cpuset relative placement, while preserving the current
API's that pass system wide numbers, not cpuset relative numbers across
the kernel-user boundary.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-31 06:02:36 +07:00
|
|
|
{
|
[PATCH] cpuset: numa_policy_rebind cleanup
Cleanup, reorganize and make more robust the mempolicy.c code to rebind
mempolicies relative to the containing cpuset after a tasks memory placement
changes.
The real motivator for this cleanup patch is to lay more groundwork for the
upcoming patch to correctly rebind NUMA mempolicies that are attached to vma's
after the containing cpuset memory placement changes.
NUMA mempolicies are constrained by the cpuset their task is a member of.
When either (1) a task is moved to a different cpuset, or (2) the 'mems'
mems_allowed of a cpuset is changed, then the NUMA mempolicies have embedded
node numbers (for MPOL_BIND, MPOL_INTERLEAVE and MPOL_PREFERRED) that need to
be recalculated, relative to their new cpuset placement.
The old code used an unreliable method of determining what was the old
mems_allowed constraining the mempolicy. It just looked at the tasks
mems_allowed value. This sort of worked with the present code, that just
rebinds the -task- mempolicy, and leaves any -vma- mempolicies broken,
referring to the old nodes. But in an upcoming patch, the vma mempolicies
will be rebound as well. Then the order in which the various task and vma
mempolicies are updated will no longer be deterministic, and one can no longer
count on the task->mems_allowed holding the old value for as long as needed.
It's not even clear if the current code was guaranteed to work reliably for
task mempolicies.
So I added a mems_allowed field to each mempolicy, stating exactly what
mems_allowed the policy is relative to, and updated synchronously and reliably
anytime that the mempolicy is rebound.
Also removed a useless wrapper routine, numa_policy_rebind(), and had its
caller, cpuset_update_task_memory_state(), call directly to the rewritten
policy_rebind() routine, and made that rebind routine extern instead of
static, and added a "mpol_" prefix to its name, making it
mpol_rebind_policy().
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-08 16:01:56 +07:00
|
|
|
mpol_rebind_policy(tsk->mempolicy, new);
|
[PATCH] cpusets: automatic numa mempolicy rebinding
This patch automatically updates a tasks NUMA mempolicy when its cpuset
memory placement changes. It does so within the context of the task,
without any need to support low level external mempolicy manipulation.
If a system is not using cpusets, or if running on a system with just the
root (all-encompassing) cpuset, then this remap is a no-op. Only when a
task is moved between cpusets, or a cpusets memory placement is changed
does the following apply. Otherwise, the main routine below,
rebind_policy() is not even called.
When mixing cpusets, scheduler affinity, and NUMA mempolicies, the
essential role of cpusets is to place jobs (several related tasks) on a set
of CPUs and Memory Nodes, the essential role of sched_setaffinity is to
manage a jobs processor placement within its allowed cpuset, and the
essential role of NUMA mempolicy (mbind, set_mempolicy) is to manage a jobs
memory placement within its allowed cpuset.
However, CPU affinity and NUMA memory placement are managed within the
kernel using absolute system wide numbering, not cpuset relative numbering.
This is ok until a job is migrated to a different cpuset, or what's the
same, a jobs cpuset is moved to different CPUs and Memory Nodes.
Then the CPU affinity and NUMA memory placement of the tasks in the job
need to be updated, to preserve their cpuset-relative position. This can
be done for CPU affinity using sched_setaffinity() from user code, as one
task can modify anothers CPU affinity. This cannot be done from an
external task for NUMA memory placement, as that can only be modified in
the context of the task using it.
However, it easy enough to remap a tasks NUMA mempolicy automatically when
a task is migrated, using the existing cpuset mechanism to trigger a
refresh of a tasks memory placement after its cpuset has changed. All that
is needed is the old and new nodemask, and notice to the task that it needs
to rebind its mempolicy. The tasks mems_allowed has the old mask, the
tasks cpuset has the new mask, and the existing
cpuset_update_current_mems_allowed() mechanism provides the notice. The
bitmap/cpumask/nodemask remap operators provide the cpuset relative
calculations.
This patch leaves open a couple of issues:
1) Updating vma and shmfs/tmpfs/hugetlbfs memory policies:
These mempolicies may reference nodes outside of those allowed to
the current task by its cpuset. Tasks are migrated as part of jobs,
which reside on what might be several cpusets in a subtree. When such
a job is migrated, all NUMA memory policy references to nodes within
that cpuset subtree should be translated, and references to any nodes
outside that subtree should be left untouched. A future patch will
provide the cpuset mechanism needed to mark such subtrees. With that
patch, we will be able to correctly migrate these other memory policies
across a job migration.
2) Updating cpuset, affinity and memory policies in user space:
This is harder. Any placement state stored in user space using
system-wide numbering will be invalidated across a migration. More
work will be required to provide user code with a migration-safe means
to manage its cpuset relative placement, while preserving the current
API's that pass system wide numbers, not cpuset relative numbers across
the kernel-user boundary.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-31 06:02:36 +07:00
|
|
|
}
|
2006-01-08 16:01:02 +07:00
|
|
|
|
[PATCH] cpuset: rebind vma mempolicies fix
Fix more of longstanding bug in cpuset/mempolicy interaction.
NUMA mempolicies (mm/mempolicy.c) are constrained by the current tasks cpuset
to just the Memory Nodes allowed by that cpuset. The kernel maintains
internal state for each mempolicy, tracking what nodes are used for the
MPOL_INTERLEAVE, MPOL_BIND or MPOL_PREFERRED policies.
When a tasks cpuset memory placement changes, whether because the cpuset
changed, or because the task was attached to a different cpuset, then the
tasks mempolicies have to be rebound to the new cpuset placement, so as to
preserve the cpuset-relative numbering of the nodes in that policy.
An earlier fix handled such mempolicy rebinding for mempolicies attached to a
task.
This fix rebinds mempolicies attached to vma's (address ranges in a tasks
address space.) Due to the need to hold the task->mm->mmap_sem semaphore while
updating vma's, the rebinding of vma mempolicies has to be done when the
cpuset memory placement is changed, at which time mmap_sem can be safely
acquired. The tasks mempolicy is rebound later, when the task next attempts
to allocate memory and notices that its task->cpuset_mems_generation is
out-of-date with its cpusets mems_generation.
Because walking the tasklist to find all tasks attached to a changing cpuset
requires holding tasklist_lock, a spinlock, one cannot update the vma's of the
affected tasks while doing the tasklist scan. In general, one cannot acquire
a semaphore (which can sleep) while already holding a spinlock (such as
tasklist_lock). So a list of mm references has to be built up during the
tasklist scan, then the tasklist lock dropped, then for each mm, its mmap_sem
acquired, and the vma's in that mm rebound.
Once the tasklist lock is dropped, affected tasks may fork new tasks, before
their mm's are rebound. A kernel global 'cpuset_being_rebound' is set to
point to the cpuset being rebound (there can only be one; cpuset modifications
are done under a global 'manage_sem' semaphore), and the mpol_copy code that
is used to copy a tasks mempolicies during fork catches such forking tasks,
and ensures their children are also rebound.
When a task is moved to a different cpuset, it is easier, as there is only one
task involved. It's mm->vma's are scanned, using the same
mpol_rebind_policy() as used above.
It may happen that both the mpol_copy hook and the update done via the
tasklist scan update the same mm twice. This is ok, as the mempolicies of
each vma in an mm keep track of what mems_allowed they are relative to, and
safely no-op a second request to rebind to the same nodes.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-08 16:01:59 +07:00
|
|
|
/*
|
|
|
|
* Rebind each vma in mm to new nodemask.
|
|
|
|
*
|
|
|
|
* Call holding a reference to mm. Takes mm->mmap_sem during call.
|
|
|
|
*/
|
|
|
|
|
|
|
|
void mpol_rebind_mm(struct mm_struct *mm, nodemask_t *new)
|
|
|
|
{
|
|
|
|
struct vm_area_struct *vma;
|
|
|
|
|
|
|
|
down_write(&mm->mmap_sem);
|
|
|
|
for (vma = mm->mmap; vma; vma = vma->vm_next)
|
|
|
|
mpol_rebind_policy(vma->vm_policy, new);
|
|
|
|
up_write(&mm->mmap_sem);
|
|
|
|
}
|
|
|
|
|
2006-01-08 16:01:02 +07:00
|
|
|
/*
|
|
|
|
* Display pages allocated per node and memory policy via /proc.
|
|
|
|
*/
|
|
|
|
|
2006-12-07 11:40:36 +07:00
|
|
|
static const char * const policy_types[] =
|
|
|
|
{ "default", "prefer", "bind", "interleave" };
|
2006-01-08 16:01:02 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Convert a mempolicy into a string.
|
|
|
|
* Returns the number of characters in buffer (if positive)
|
|
|
|
* or an error (negative)
|
|
|
|
*/
|
|
|
|
static inline int mpol_to_str(char *buffer, int maxlen, struct mempolicy *pol)
|
|
|
|
{
|
|
|
|
char *p = buffer;
|
|
|
|
int l;
|
|
|
|
nodemask_t nodes;
|
|
|
|
int mode = pol ? pol->policy : MPOL_DEFAULT;
|
|
|
|
|
|
|
|
switch (mode) {
|
|
|
|
case MPOL_DEFAULT:
|
|
|
|
nodes_clear(nodes);
|
|
|
|
break;
|
|
|
|
|
|
|
|
case MPOL_PREFERRED:
|
|
|
|
nodes_clear(nodes);
|
|
|
|
node_set(pol->v.preferred_node, nodes);
|
|
|
|
break;
|
|
|
|
|
|
|
|
case MPOL_BIND:
|
|
|
|
get_zonemask(pol, &nodes);
|
|
|
|
break;
|
|
|
|
|
|
|
|
case MPOL_INTERLEAVE:
|
|
|
|
nodes = pol->v.nodes;
|
|
|
|
break;
|
|
|
|
|
|
|
|
default:
|
|
|
|
BUG();
|
|
|
|
return -EFAULT;
|
|
|
|
}
|
|
|
|
|
|
|
|
l = strlen(policy_types[mode]);
|
|
|
|
if (buffer + maxlen < p + l + 1)
|
|
|
|
return -ENOSPC;
|
|
|
|
|
|
|
|
strcpy(p, policy_types[mode]);
|
|
|
|
p += l;
|
|
|
|
|
|
|
|
if (!nodes_empty(nodes)) {
|
|
|
|
if (buffer + maxlen < p + 2)
|
|
|
|
return -ENOSPC;
|
|
|
|
*p++ = '=';
|
|
|
|
p += nodelist_scnprintf(p, buffer + maxlen - p, nodes);
|
|
|
|
}
|
|
|
|
return p - buffer;
|
|
|
|
}
|
|
|
|
|
|
|
|
struct numa_maps {
|
|
|
|
unsigned long pages;
|
|
|
|
unsigned long anon;
|
2006-03-07 06:42:53 +07:00
|
|
|
unsigned long active;
|
|
|
|
unsigned long writeback;
|
2006-01-08 16:01:02 +07:00
|
|
|
unsigned long mapcount_max;
|
2006-03-07 06:42:53 +07:00
|
|
|
unsigned long dirty;
|
|
|
|
unsigned long swapcache;
|
2006-01-08 16:01:02 +07:00
|
|
|
unsigned long node[MAX_NUMNODES];
|
|
|
|
};
|
|
|
|
|
2006-03-07 06:42:53 +07:00
|
|
|
static void gather_stats(struct page *page, void *private, int pte_dirty)
|
2006-01-08 16:01:02 +07:00
|
|
|
{
|
|
|
|
struct numa_maps *md = private;
|
|
|
|
int count = page_mapcount(page);
|
|
|
|
|
2006-03-07 06:42:53 +07:00
|
|
|
md->pages++;
|
|
|
|
if (pte_dirty || PageDirty(page))
|
|
|
|
md->dirty++;
|
2006-01-08 16:01:02 +07:00
|
|
|
|
2006-03-07 06:42:53 +07:00
|
|
|
if (PageSwapCache(page))
|
|
|
|
md->swapcache++;
|
2006-01-08 16:01:02 +07:00
|
|
|
|
2006-03-07 06:42:53 +07:00
|
|
|
if (PageActive(page))
|
|
|
|
md->active++;
|
|
|
|
|
|
|
|
if (PageWriteback(page))
|
|
|
|
md->writeback++;
|
2006-01-08 16:01:02 +07:00
|
|
|
|
|
|
|
if (PageAnon(page))
|
|
|
|
md->anon++;
|
|
|
|
|
2006-03-07 06:42:53 +07:00
|
|
|
if (count > md->mapcount_max)
|
|
|
|
md->mapcount_max = count;
|
|
|
|
|
2006-01-08 16:01:02 +07:00
|
|
|
md->node[page_to_nid(page)]++;
|
|
|
|
}
|
|
|
|
|
2006-03-08 12:55:22 +07:00
|
|
|
#ifdef CONFIG_HUGETLB_PAGE
|
2006-03-07 06:42:53 +07:00
|
|
|
static void check_huge_range(struct vm_area_struct *vma,
|
|
|
|
unsigned long start, unsigned long end,
|
|
|
|
struct numa_maps *md)
|
|
|
|
{
|
|
|
|
unsigned long addr;
|
|
|
|
struct page *page;
|
|
|
|
|
|
|
|
for (addr = start; addr < end; addr += HPAGE_SIZE) {
|
|
|
|
pte_t *ptep = huge_pte_offset(vma->vm_mm, addr & HPAGE_MASK);
|
|
|
|
pte_t pte;
|
|
|
|
|
|
|
|
if (!ptep)
|
|
|
|
continue;
|
|
|
|
|
|
|
|
pte = *ptep;
|
|
|
|
if (pte_none(pte))
|
|
|
|
continue;
|
|
|
|
|
|
|
|
page = pte_page(pte);
|
|
|
|
if (!page)
|
|
|
|
continue;
|
|
|
|
|
|
|
|
gather_stats(page, md, pte_dirty(*ptep));
|
|
|
|
}
|
|
|
|
}
|
2006-03-08 12:55:22 +07:00
|
|
|
#else
|
|
|
|
static inline void check_huge_range(struct vm_area_struct *vma,
|
|
|
|
unsigned long start, unsigned long end,
|
|
|
|
struct numa_maps *md)
|
|
|
|
{
|
|
|
|
}
|
|
|
|
#endif
|
2006-03-07 06:42:53 +07:00
|
|
|
|
2006-01-08 16:01:02 +07:00
|
|
|
int show_numa_map(struct seq_file *m, void *v)
|
|
|
|
{
|
2006-06-26 14:25:55 +07:00
|
|
|
struct proc_maps_private *priv = m->private;
|
2006-01-08 16:01:02 +07:00
|
|
|
struct vm_area_struct *vma = v;
|
|
|
|
struct numa_maps *md;
|
2006-03-07 06:42:53 +07:00
|
|
|
struct file *file = vma->vm_file;
|
|
|
|
struct mm_struct *mm = vma->vm_mm;
|
Fix NUMA Memory Policy Reference Counting
This patch proposes fixes to the reference counting of memory policy in the
page allocation paths and in show_numa_map(). Extracted from my "Memory
Policy Cleanups and Enhancements" series as stand-alone.
Shared policy lookup [shmem] has always added a reference to the policy,
but this was never unrefed after page allocation or after formatting the
numa map data.
Default system policy should not require additional ref counting, nor
should the current task's task policy. However, show_numa_map() calls
get_vma_policy() to examine what may be [likely is] another task's policy.
The latter case needs protection against freeing of the policy.
This patch adds a reference count to a mempolicy returned by
get_vma_policy() when the policy is a vma policy or another task's
mempolicy. Again, shared policy is already reference counted on lookup. A
matching "unref" [__mpol_free()] is performed in alloc_page_vma() for
shared and vma policies, and in show_numa_map() for shared and another
task's mempolicy. We can call __mpol_free() directly, saving an admittedly
inexpensive inline NULL test, because we know we have a non-NULL policy.
Handling policy ref counts for hugepages is a bit trickier.
huge_zonelist() returns a zone list that might come from a shared or vma
'BIND policy. In this case, we should hold the reference until after the
huge page allocation in dequeue_hugepage(). The patch modifies
huge_zonelist() to return a pointer to the mempolicy if it needs to be
unref'd after allocation.
Kernel Build [16cpu, 32GB, ia64] - average of 10 runs:
w/o patch w/ refcount patch
Avg Std Devn Avg Std Devn
Real: 100.59 0.38 100.63 0.43
User: 1209.60 0.37 1209.91 0.31
System: 81.52 0.42 81.64 0.34
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Acked-by: Andi Kleen <ak@suse.de>
Cc: Christoph Lameter <clameter@sgi.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-09-19 12:46:47 +07:00
|
|
|
struct mempolicy *pol;
|
2006-01-08 16:01:02 +07:00
|
|
|
int n;
|
|
|
|
char buffer[50];
|
|
|
|
|
2006-03-07 06:42:53 +07:00
|
|
|
if (!mm)
|
2006-01-08 16:01:02 +07:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
md = kzalloc(sizeof(struct numa_maps), GFP_KERNEL);
|
|
|
|
if (!md)
|
|
|
|
return 0;
|
|
|
|
|
Fix NUMA Memory Policy Reference Counting
This patch proposes fixes to the reference counting of memory policy in the
page allocation paths and in show_numa_map(). Extracted from my "Memory
Policy Cleanups and Enhancements" series as stand-alone.
Shared policy lookup [shmem] has always added a reference to the policy,
but this was never unrefed after page allocation or after formatting the
numa map data.
Default system policy should not require additional ref counting, nor
should the current task's task policy. However, show_numa_map() calls
get_vma_policy() to examine what may be [likely is] another task's policy.
The latter case needs protection against freeing of the policy.
This patch adds a reference count to a mempolicy returned by
get_vma_policy() when the policy is a vma policy or another task's
mempolicy. Again, shared policy is already reference counted on lookup. A
matching "unref" [__mpol_free()] is performed in alloc_page_vma() for
shared and vma policies, and in show_numa_map() for shared and another
task's mempolicy. We can call __mpol_free() directly, saving an admittedly
inexpensive inline NULL test, because we know we have a non-NULL policy.
Handling policy ref counts for hugepages is a bit trickier.
huge_zonelist() returns a zone list that might come from a shared or vma
'BIND policy. In this case, we should hold the reference until after the
huge page allocation in dequeue_hugepage(). The patch modifies
huge_zonelist() to return a pointer to the mempolicy if it needs to be
unref'd after allocation.
Kernel Build [16cpu, 32GB, ia64] - average of 10 runs:
w/o patch w/ refcount patch
Avg Std Devn Avg Std Devn
Real: 100.59 0.38 100.63 0.43
User: 1209.60 0.37 1209.91 0.31
System: 81.52 0.42 81.64 0.34
Signed-off-by: Lee Schermerhorn <lee.schermerhorn@hp.com>
Acked-by: Andi Kleen <ak@suse.de>
Cc: Christoph Lameter <clameter@sgi.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-09-19 12:46:47 +07:00
|
|
|
pol = get_vma_policy(priv->task, vma, vma->vm_start);
|
|
|
|
mpol_to_str(buffer, sizeof(buffer), pol);
|
|
|
|
/*
|
|
|
|
* unref shared or other task's mempolicy
|
|
|
|
*/
|
|
|
|
if (pol != &default_policy && pol != current->mempolicy)
|
|
|
|
__mpol_free(pol);
|
2006-03-07 06:42:53 +07:00
|
|
|
|
|
|
|
seq_printf(m, "%08lx %s", vma->vm_start, buffer);
|
|
|
|
|
|
|
|
if (file) {
|
|
|
|
seq_printf(m, " file=");
|
2006-12-08 17:37:21 +07:00
|
|
|
seq_path(m, file->f_path.mnt, file->f_path.dentry, "\n\t= ");
|
2006-03-07 06:42:53 +07:00
|
|
|
} else if (vma->vm_start <= mm->brk && vma->vm_end >= mm->start_brk) {
|
|
|
|
seq_printf(m, " heap");
|
|
|
|
} else if (vma->vm_start <= mm->start_stack &&
|
|
|
|
vma->vm_end >= mm->start_stack) {
|
|
|
|
seq_printf(m, " stack");
|
|
|
|
}
|
|
|
|
|
|
|
|
if (is_vm_hugetlb_page(vma)) {
|
|
|
|
check_huge_range(vma, vma->vm_start, vma->vm_end, md);
|
|
|
|
seq_printf(m, " huge");
|
|
|
|
} else {
|
2006-03-02 17:54:37 +07:00
|
|
|
check_pgd_range(vma, vma->vm_start, vma->vm_end,
|
2007-10-16 15:25:35 +07:00
|
|
|
&node_states[N_HIGH_MEMORY], MPOL_MF_STATS, md);
|
2006-03-07 06:42:53 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
if (!md->pages)
|
|
|
|
goto out;
|
2006-01-08 16:01:02 +07:00
|
|
|
|
2006-03-07 06:42:53 +07:00
|
|
|
if (md->anon)
|
|
|
|
seq_printf(m," anon=%lu",md->anon);
|
2006-01-08 16:01:02 +07:00
|
|
|
|
2006-03-07 06:42:53 +07:00
|
|
|
if (md->dirty)
|
|
|
|
seq_printf(m," dirty=%lu",md->dirty);
|
2006-01-08 16:01:02 +07:00
|
|
|
|
2006-03-07 06:42:53 +07:00
|
|
|
if (md->pages != md->anon && md->pages != md->dirty)
|
|
|
|
seq_printf(m, " mapped=%lu", md->pages);
|
2006-01-08 16:01:02 +07:00
|
|
|
|
2006-03-07 06:42:53 +07:00
|
|
|
if (md->mapcount_max > 1)
|
|
|
|
seq_printf(m, " mapmax=%lu", md->mapcount_max);
|
2006-01-08 16:01:02 +07:00
|
|
|
|
2006-03-07 06:42:53 +07:00
|
|
|
if (md->swapcache)
|
|
|
|
seq_printf(m," swapcache=%lu", md->swapcache);
|
|
|
|
|
|
|
|
if (md->active < md->pages && !is_vm_hugetlb_page(vma))
|
|
|
|
seq_printf(m," active=%lu", md->active);
|
|
|
|
|
|
|
|
if (md->writeback)
|
|
|
|
seq_printf(m," writeback=%lu", md->writeback);
|
|
|
|
|
2007-10-16 15:25:35 +07:00
|
|
|
for_each_node_state(n, N_HIGH_MEMORY)
|
2006-03-07 06:42:53 +07:00
|
|
|
if (md->node[n])
|
|
|
|
seq_printf(m, " N%d=%lu", n, md->node[n]);
|
|
|
|
out:
|
|
|
|
seq_putc(m, '\n');
|
2006-01-08 16:01:02 +07:00
|
|
|
kfree(md);
|
|
|
|
|
|
|
|
if (m->count < m->size)
|
2006-06-26 14:25:55 +07:00
|
|
|
m->version = (vma != priv->tail_vma) ? vma->vm_start : 0;
|
2006-01-08 16:01:02 +07:00
|
|
|
return 0;
|
|
|
|
}
|