Florian Fainelli says:
====================
net: systemport: RX path and suspend fixes
These two patches fix a race condition where we have our RX interrupts
enabled, but not NAPI for the RX path, and the second patch fixes an
issue for packets stuck in RX fifo during a suspend/resume cycle.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
bcm_sysport_resume() was missing an UniMAC reset which can lead to
various receive FIFO corruptions coming out of a suspend cycle. If the
RX FIFO is stuck, it will deliver corrupted/duplicate packets towards
the host CPU interface.
This could be reproduced on crowded network and when Wake-on-LAN is
enabled for this particular interface because the switch still forwards
packets towards the host CPU interface (SYSTEMPORT), and we had to leave
the UniMAC RX enable bit on to allow matching MagicPackets.
Once we re-enter the resume function, there is a small window during
which the UniMAC receive is still enabled, and we start queueing
packets, but the RDMA and RBUF engines are not ready, which leads to
having packets stuck in the UniMAC RX FIFO, ultimately delivered towards
the host CPU as corrupted.
Fixes: 40755a0fce ("net: systemport: add suspend and resume support")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There is currently a small window during which the SYSTEMPORT adapter
enables its RX interrupts without having enabled its NAPI handler, which
can result in packets to be discarded during interface bringup.
A similar but more serious window exists in bcm_sysport_resume() during
which we can have the RDMA engine not fully prepared to receive packets
and yet having RX interrupts enabled.
Fix this my moving the RX interrupt enable down to
bcm_sysport_netif_start() after napi_enable() for the RX path is called,
which fixes both call sites: bcm_sysport_open() and
bcm_sysport_resume().
Fixes: b02e6d9ba7 ("net: systemport: add bcm_sysport_netif_{enable,stop}")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix kernel-doc warning in <linux/skbuff.h> by making both headers_start
and headers_end private fields.
Warning(..//include/linux/skbuff.h:654): No description found for parameter 'headers_end[0]'
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Marvell phy 88E1145 configuration & initialization was missing a case
for initializing SGMII mode. This patch adds that case.
Signed-off-by: Vince Bridgers <vbridger@opensource.altera.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We have a race that can lead us to miss skinny extent items in the function
btrfs_lookup_extent_info() when the skinny metadata feature is enabled.
So basically the sequence of steps is:
1) We search in the extent tree for the skinny extent, which returns > 0
(not found);
2) We check the previous item in the returned leaf for a non-skinny extent,
and we don't find it;
3) Because we didn't find the non-skinny extent in step 2), we release our
path to search the extent tree again, but this time for a non-skinny
extent key;
4) Right after we released our path in step 3), a skinny extent was inserted
in the extent tree (delayed refs were run) - our second extent tree search
will miss it, because it's not looking for a skinny extent;
5) After the second search returned (with ret > 0), we look for any delayed
ref for our extent's bytenr (and we do it while holding a read lock on the
leaf), but we won't find any, as such delayed ref had just run and completed
after we released out path in step 3) before doing the second search.
Fix this by removing completely the path release and re-search logic. This is
safe, because if we seach for a metadata item and we don't find it, we have the
guarantee that the returned leaf is the one where the item would be inserted,
and so path->slots[0] > 0 and path->slots[0] - 1 must be the slot where the
non-skinny extent item is if it exists. The only case where path->slots[0] is
zero is when there are no smaller keys in the tree (i.e. no left siblings for
our leaf), in which case the re-search logic isn't needed as well.
This race has been present since the introduction of skinny metadata (change
3173a18f70).
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: Miao Xie <miaox@cn.fujitsu.com>
Signed-off-by: Chris Mason <clm@fb.com>
Pull two nfsd fixes from Bruce Fields:
"One regression from the 3.16 xdr rewrite, one an older bug exposed by
a separate bug in the client's new SEEK code"
* 'for-3.18' of git://linux-nfs.org/~bfields/linux:
nfsd4: fix crash on unknown operation number
nfsd4: fix response size estimation for OP_SEQUENCE
the accounting of the ftrace_ops trampoline logic. One was that the
old hash was not updated before calling the modify code for an ftrace_ops.
The second bug was what let the first bug go unnoticed, as the update would
check the current hash for all ftrace_ops (where it should only check the
old hash for modified ones). This let things work when only one ftrace_ops
was registered to a function, but could break if more than one was
registered depending on the order of the look ups.
The worse thing that can happen if this bug triggers is that the ftrace
self checks would find an anomaly and shut itself down.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQEcBAABAgAGBQJUToYWAAoJEEjnJuOKh9ldfS8H/36CL5E+4itux9tIhf13Untj
FSi3EzvEdrTYu7IhdyRB6N7cp07g79jU3v40ZDLxDHzG2i4VLft/Z3uzIC0Z6mhL
kJZCCWpUTAKJO/UPFcenEZ7eiL+B+5QVOc1Oxcet0odG5HWkEZG62va/MrhB9k/7
uUNRqXNjg7w2rG0TK2qjcTHiPGJ9h7/wG9RgYktAIs27BUmip5sRS1IMyFL51Gpo
UNtIKGtG6/4hizdlHhWBuAa6ErM37GPskx3iP/45xiAu3J8SIbOk1FBe+4Xk+DZQ
hZK479hzlk6OU/M2vDJefG1d6zeQ7y00LMkUIAPiUEgayXAXpYX7UjV13CLQeGU=
=HrhJ
-----END PGP SIGNATURE-----
Merge tag 'trace-fixes-v3.18-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace
Pull ftrace trampoline accounting fixes from Steven Rostedt:
"Adding the new code for 3.19, I discovered a couple of minor bugs with
the accounting of the ftrace_ops trampoline logic.
One was that the old hash was not updated before calling the modify
code for an ftrace_ops. The second bug was what let the first bug go
unnoticed, as the update would check the current hash for all
ftrace_ops (where it should only check the old hash for modified
ones). This let things work when only one ftrace_ops was registered
to a function, but could break if more than one was registered
depending on the order of the look ups.
The worse thing that can happen if this bug triggers is that the
ftrace self checks would find an anomaly and shut itself down"
* tag 'trace-fixes-v3.18-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace:
ftrace: Fix checking of trampoline ftrace_ops in finding trampoline
ftrace: Set ops->old_hash on modifying what an ops hooks to
Commit 35ce7f29a4 (rcu: Create rcuo kthreads only for onlined CPUs)
avoids creating rcuo kthreads for CPUs that never come online. This
fixes a bug in many instances of firmware: Instead of lying about their
age, these systems instead lie about the number of CPUs that they have.
Before commit 35ce7f29a4, this could result in huge numbers of useless
rcuo kthreads being created.
It appears that experience indicates that I should have told the
people suffering from this problem to fix their broken firmware, but
I instead produced what turned out to be a partial fix. The missing
piece supplied by this commit makes sure that rcu_barrier() knows not to
post callbacks for no-CBs CPUs that have not yet come online, because
otherwise rcu_barrier() will hang on systems having firmware that lies
about the number of CPUs.
It is tempting to simply have rcu_barrier() refuse to post a callback on
any no-CBs CPU that does not have an rcuo kthread. This unfortunately
does not work because rcu_barrier() is required to wait for all pending
callbacks. It is therefore required to wait even for those callbacks
that cannot possibly be invoked. Even if doing so hangs the system.
Given that posting a callback to a no-CBs CPU that does not yet have an
rcuo kthread can hang rcu_barrier(), It is tempting to report an error
in this case. Unfortunately, this will result in false positives at
boot time, when it is perfectly legal to post callbacks to the boot CPU
before the scheduler has started, in other words, before it is legal
to invoke rcu_barrier().
So this commit instead has rcu_barrier() avoid posting callbacks to
CPUs having neither rcuo kthread nor pending callbacks, and has it
complain bitterly if it finds CPUs having no rcuo kthread but some
pending callbacks. And when rcu_barrier() does find CPUs having no rcuo
kthread but pending callbacks, as noted earlier, it has no choice but
to hang indefinitely.
Reported-by: Yanko Kaneti <yaneti@declera.com>
Reported-by: Jay Vosburgh <jay.vosburgh@canonical.com>
Reported-by: Meelis Roos <mroos@linux.ee>
Reported-by: Eric B Munson <emunson@akamai.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Eric B Munson <emunson@akamai.com>
Tested-by: Jay Vosburgh <jay.vosburgh@canonical.com>
Tested-by: Yanko Kaneti <yaneti@declera.com>
Tested-by: Kevin Fenzi <kevin@scrye.com>
Tested-by: Meelis Roos <mroos@linux.ee>
Pull ARM fixes from Russell King:
"A couple of ARM fixes.
We fix some printk formats for ptrdiff_t quantities which cause GCC
4.9 to complain, and we also blacklist known buggy GCC 4.8.x compilers
as their miscompilation is serious enough to cause filesystem
corruption, even through many distros have fixed their versions"
* 'fixes' of git://ftp.arm.linux.org.uk/~rmk/linux-arm:
ARM: fix some printk formats
ARM: Blacklist GCC 4.8.0 to GCC 4.8.2 - PR58854
When unmapping a range of pages in zap_pte_range, the page being
unmapped is added to an mmu_gather_batch structure for asynchronous
freeing. If we run out of space in the batch structure before the range
has been completely unmapped, then we break out of the loop, force a
TLB flush and free the pages that we have batched so far. If there are
further pages to unmap, then we resume the loop where we left off.
Unfortunately, we forget to update addr when we break out of the loop,
which causes us to truncate the range being invalidated as the end
address is exclusive. When we re-enter the loop at the same address, the
page has already been freed and the pte_present test will fail, meaning
that we do not reconsider the address for invalidation.
This patch fixes the problem by incrementing addr by the PAGE_SIZE
before breaking out of the loop on batch failure.
Signed-off-by: Will Deacon <will.deacon@arm.com>
Cc: stable@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
when slave 0 has no phy and slave 1 connected to phy, driver probe will
fail as there is no phy id present for slave 0 device tree, so continuing
even though no phy-id found, also moving mac-id read later to ensure
mac-id is read from device tree even when phy-id entry in not found.
Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
John W. Linville says:
====================
pull request: wireless 2014-10-28
Please pull this batch of fixes intended for the 3.18 stream!
For the mac80211 bits, Johannes says:
"Here are a few fixes for the wireless stack: one fixes the
RTS rate, one for a debugfs file, one to return the correct
channel to userspace, a sanity check for a userspace value
and the remaining two are just documentation fixes."
For the iwlwifi bits, Emmanuel says:
"I revert here a patch that caused interoperability issues.
dvm gets a fix for a bug that was reported by many users.
Two minor fixes for BT Coex and platform power fix that helps
reducing latency when the PCIe link goes to low power states."
In addition...
Felix Fietkau adds a couple of ath code fixes related to regulatory
rule enforcement.
Hauke Mehrtens fixes a build break with bcma when CONFIG_OF_ADDRESS
is not set.
Karsten Wiese provides a trio of minor fixes for rtl8192cu.
Kees Cook prevents a potential information leak in rtlwifi.
Larry Finger also brings a trio of minor fixes for rtlwifi.
Rafał Miłecki adds a device ID to the bcma bus driver.
Rickard Strandqvist offers some strn* -> strl* changes in brcmfmac
to eliminate non-terminated string issues.
Sujith Manoharan avoids some ath9k stalls by enabling HW queue control
only for MCC.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Andrew Lunn says:
====================
DSA tagging mismatches
The second patch is a fix, which should be applied to -rc. It is
possible to get a DSA configuration which does not work. The patch
stops this happening.
The first patch detects this situation, and errors out the probe of
DSA, making it more obvious something is wrong. It is not required to
apply it -rc.
v2 fixes the use case pointed out by Florian, that a switch driver
may use DSA_TAG_PROTO_NONE which the patch did not correctly handle.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
The mv88e6171 can support two different tagging protocols, DSA and
EDSA. The switch driver structure only allows one protocol to be
enumerated, and DSA was chosen. However the Kconfig entry ensures the
EDSA tagging code is built. With a minimal configuration, we then end
up with a mismatch. The probe is successful, EDSA tagging is used, but
the switch is configured for DSA, resulting in mangled packets.
Change the switch driver structure to enumerate EDSA, fixing the
mismatch.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Fixes: 42f2725394 ("net: DSA: Marvell mv88e6171 switch driver")
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If there is a mismatch between enabled tagging protocols and the
protocol the switch supports, error out, rather than continue with a
situation which is unlikely to work.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
cc: alexander.h.duyck@intel.com
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
psmouse_reconnect() will not be called if psmouse driver is not bound to
the serio port, so there is no point in checking that. Also, as coded, it
introduces potential NULL dereference in psmouse_dbg() in case psmouse is
indeed NULL. Let's just remove it.
Detected by Coverity: CID 146528
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
I believe the intent of the code was to drop oldest bytes from the queue,
not the latest if we drop one byte and both latest and some oldest of we
are dropping more than one.
Acked-by: Jan-Benedict Glaw <jbglaw@lug-owl.de>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Coverity pointed out that at return point error is always 0 so the
conditional is not needed.
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Fix a memory leak with scsi-mq triggered by commands with large data
transfer length.
Fixes: c53c6d6a68 ("scatterlist: allow chaining to preallocated chunks")
Cc: <stable@vger.kernel.org> # 3.17.x
Signed-off-by: Tony Battersby <tonyb@cybernetics.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
The linux-next wiki at http://linux.f-seidel.de/linux-next/pmwiki has
been gone for several months now.
Signed-off-by: Jim Davis <jim.epost@gmail.com>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
The prctl test code in Documentation/ tries to show how to
use a call that only makes sense on x86. Restrict it there
so that other platforms don't try to call asm("rdtsc").
Signed-off-by: Alexander Graf <agraf@suse.de>
Acked-by: Peter Foley <pefoley2@pefoley.com>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
In compat mode, we copy each field of snd_pcm_status struct but don't
touch the reserved fields, and this leaves uninitialized values
there. Meanwhile the native ioctl does zero-clear the whole
structure, so we should follow the same rule in compat mode, too.
Reported-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Fix duplicate XT-PIC seen in /proc/interrupts on x86 systems
that make use of 8259A Programmable Interrupt Controllers.
Specifically convert output like this:
CPU0
0: 76573 XT-PIC-XT-PIC timer
1: 11 XT-PIC-XT-PIC i8042
2: 0 XT-PIC-XT-PIC cascade
4: 8 XT-PIC-XT-PIC serial
6: 3 XT-PIC-XT-PIC floppy
7: 0 XT-PIC-XT-PIC parport0
8: 1 XT-PIC-XT-PIC rtc0
10: 448 XT-PIC-XT-PIC fddi0
12: 23 XT-PIC-XT-PIC eth0
14: 2464 XT-PIC-XT-PIC ide0
NMI: 0 Non-maskable interrupts
ERR: 0
to one like this:
CPU0
0: 122033 XT-PIC timer
1: 11 XT-PIC i8042
2: 0 XT-PIC cascade
4: 8 XT-PIC serial
6: 3 XT-PIC floppy
7: 0 XT-PIC parport0
8: 1 XT-PIC rtc0
10: 145 XT-PIC fddi0
12: 31 XT-PIC eth0
14: 2245 XT-PIC ide0
NMI: 0 Non-maskable interrupts
ERR: 0
that is one like we used to have from ~2.2 till it was changed
sometime.
The rationale is there is no value in this duplicate
information, it merely clutters output and looks ugly. We only
have one handler for 8259A interrupts so there is no need to
give it a name separate from the name already given to
irq_chip.
We could define meaningful names for handlers based on bits in
the ELCR register on systems that have it or the value of the
LTIM bit we use in ICW1 otherwise (hardcoded to 0 though with
MCA support gone), to tell edge-triggered and level-triggered
inputs apart. While that information does not affect 8259A
interrupt handlers it could help people determine which lines
are shareable and which are not. That is material for a
separate change though.
Any tools that parse /proc/interrupts are supposed not to be
affected since it was many years we used the format this change
converts back to.
Signed-off-by: Maciej W. Rozycki <macro@linux-mips.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/alpine.LFD.2.11.1410260147190.21390@eddie.linux-mips.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The bug referenced by the comment in this commit was not
completely fixed in GCC 4.8.2, as I mentioned in a thread back
in February:
https://lkml.org/lkml/2014/2/12/797
The conclusion at that time was to make the quirk unconditional
until the bug could be found and fixed in GCC. Unfortunately,
when I submitted the patch (commit a9f18034) I left a comment
in that claimed the bug was fixed in GCC 4.8.2+.
This comment is inaccurate, and should be removed.
Signed-off-by: Steven Noonan <steven@uplinklabs.net>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Jakub Jelinek <jakub@redhat.com>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Link: http://lkml.kernel.org/r/1414274982-14040-1-git-send-email-steven@uplinklabs.net
Cc: Ingo Molnar <mingo@kernel.org>
struct perf_event_mmap_page has members called "index" and
"cap_user_rdpmc". Spell them correctly in the examples.
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: linux-api@vger.kernel.org
Link: http://lkml.kernel.org/r/320ba26391a8123cc16e5f02d24d34bd404332fd.1412313343.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Andy reported that the current state of event_idx is rather confused.
So remove all but the x86_pmu implementation and change the default to
return 0 (the safe option).
Reported-by: Andy Lutomirski <luto@amacapital.net>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Christoph Lameter <cl@linux.com>
Cc: Cody P Schafer <cody@linux.vnet.ibm.com>
Cc: Cody P Schafer <dev@codyps.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Cc: Himangi Saraogi <himangi774@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Paul Gortmaker <paul.gortmaker@windriver.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: sukadev@linux.vnet.ibm.com <sukadev@linux.vnet.ibm.com>
Cc: Thomas Huth <thuth@linux.vnet.ibm.com>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Cc: linux390@de.ibm.com
Cc: linuxppc-dev@lists.ozlabs.org
Cc: linux-s390@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Andy spotted the fail in what was intended as a conditional printk level.
Reported-by: Andy Lutomirski <luto@amacapital.net>
Fixes: cc6cd47e73 ("perf/x86: Tone down kernel messages when the PMU check fails in a virtual environment")
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/20141007124757.GH19379@twins.programming.kicks-ass.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
1) switched_to_dl() check is wrong. We reschedule only
if rq->curr is deadline task, and we do not reschedule
if it's a lower priority task. But we must always
preempt a task of other classes.
2) dl_task_timer():
Policy does not change in case of priority inheritance.
rt_mutex_setprio() changes prio, while policy remains old.
So we lose some balancing logic in dl_task_timer() and
switched_to_dl() when we check policy instead of priority. Boosted
task may be rq->curr.
(I didn't change switched_from_dl() because no check is necessary
there at all).
I've looked at this place(switched_to_dl) several times and even fixed
this function, but found just now... I suppose some performance tests
may work better after this.
Signed-off-by: Kirill Tkhai <ktkhai@parallels.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Juri Lelli <juri.lelli@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/1413909356.19914.128.camel@tkhai
Signed-off-by: Ingo Molnar <mingo@kernel.org>
preempt_schedule_context() does preempt_enable_notrace() at the end
and this can call the same function again; exception_exit() is heavy
and it is quite possible that need-resched is true again.
1. Change this code to dec preempt_count() and check need_resched()
by hand.
2. As Linus suggested, we can use the PREEMPT_ACTIVE bit and avoid
the enable/disable dance around __schedule(). But in this case
we need to move into sched/core.c.
3. Cosmetic, but x86 forgets to declare this function. This doesn't
really matter because it is only called by asm helpers, still it
make sense to add the declaration into asm/preempt.h to match
preempt_schedule().
Reported-by: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Alexander Graf <agraf@suse.de>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Christoph Lameter <cl@linux.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Peter Anvin <hpa@zytor.com>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Cc: Chuck Ebbert <cebbert.lkml@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Link: http://lkml.kernel.org/r/20141005202322.GB27962@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
While offling node by hot removing memory, the following divide error
occurs:
divide error: 0000 [#1] SMP
[...]
Call Trace:
[...] handle_mm_fault
[...] ? try_to_wake_up
[...] ? wake_up_state
[...] __do_page_fault
[...] ? do_futex
[...] ? put_prev_entity
[...] ? __switch_to
[...] do_page_fault
[...] page_fault
[...]
RIP [<ffffffff810a7081>] task_numa_fault
RSP <ffff88084eb2bcb0>
The issue occurs as follows:
1. When page fault occurs and page is allocated from node 1,
task_struct->numa_faults_buffer_memory[] of node 1 is
incremented and p->numa_faults_locality[] is also incremented
as follows:
o numa_faults_buffer_memory[] o numa_faults_locality[]
NR_NUMA_HINT_FAULT_TYPES
| 0 | 1 |
---------------------------------- ----------------------
node 0 | 0 | 0 | remote | 0 |
node 1 | 0 | 1 | locale | 1 |
---------------------------------- ----------------------
2. node 1 is offlined by hot removing memory.
3. When page fault occurs, fault_types[] is calculated by using
p->numa_faults_buffer_memory[] of all online nodes in
task_numa_placement(). But node 1 was offline by step 2. So
the fault_types[] is calculated by using only
p->numa_faults_buffer_memory[] of node 0. So both of fault_types[]
are set to 0.
4. The values(0) of fault_types[] pass to update_task_scan_period().
5. numa_faults_locality[1] is set to 1. So the following division is
calculated.
static void update_task_scan_period(struct task_struct *p,
unsigned long shared, unsigned long private){
...
ratio = DIV_ROUND_UP(private * NUMA_PERIOD_SLOTS, (private + shared));
}
6. But both of private and shared are set to 0. So divide error
occurs here.
The divide error is rare case because the trigger is node offline.
This patch always increments denominator for avoiding divide error.
Signed-off-by: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/54475703.8000505@jp.fujitsu.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Unlocked access to dst_rq->curr in task_numa_compare() is racy.
If curr task is exiting this may be a reason of use-after-free:
task_numa_compare() do_exit()
... current->flags |= PF_EXITING;
... release_task()
... ~~delayed_put_task_struct()~~
... schedule()
rcu_read_lock() ...
cur = ACCESS_ONCE(dst_rq->curr) ...
... rq->curr = next;
... context_switch()
... finish_task_switch()
... put_task_struct()
... __put_task_struct()
... free_task_struct()
task_numa_assign() ...
get_task_struct() ...
As noted by Oleg:
<<The lockless get_task_struct(tsk) is only safe if tsk == current
and didn't pass exit_notify(), or if this tsk was found on a rcu
protected list (say, for_each_process() or find_task_by_vpid()).
IOW, it is only safe if release_task() was not called before we
take rcu_read_lock(), in this case we can rely on the fact that
delayed_put_pid() can not drop the (potentially) last reference
until rcu_read_unlock().
And as Kirill pointed out task_numa_compare()->task_numa_assign()
path does get_task_struct(dst_rq->curr) and this is not safe. The
task_struct itself can't go away, but rcu_read_lock() can't save
us from the final put_task_struct() in finish_task_switch(); this
reference goes away without rcu gp>>
The patch provides simple check of PF_EXITING flag. If it's not set,
this guarantees that call_rcu() of delayed_put_task_struct() callback
hasn't happened yet, so we can safely do get_task_struct() in
task_numa_assign().
Locked dst_rq->lock protects from concurrency with the last schedule().
Reusing or unmapping of cur's memory may happen without it.
Suggested-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Kirill Tkhai <ktkhai@parallels.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/1413962231.19914.130.camel@tkhai
Signed-off-by: Ingo Molnar <mingo@kernel.org>
dl_task_timer() is racy against several paths. Daniel noticed that
the replenishment timer may experience a race condition against an
enqueue_dl_entity() called from rt_mutex_setprio(). With his own
words:
rt_mutex_setprio() resets p->dl.dl_throttled. So the pattern is:
start_dl_timer() throttled = 1, rt_mutex_setprio() throlled = 0,
sched_switch() -> enqueue_task(), dl_task_timer-> enqueue_task()
throttled is 0
=> BUG_ON(on_dl_rq(dl_se)) fires as the scheduling entity is already
enqueued on the -deadline runqueue.
As we do for the other races, we just bail out in the replenishment
timer code.
Reported-by: Daniel Wagner <daniel.wagner@bmw-carit.de>
Tested-by: Daniel Wagner <daniel.wagner@bmw-carit.de>
Signed-off-by: Juri Lelli <juri.lelli@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: vincent@legout.info
Cc: Dario Faggioli <raistlin@linux.it>
Cc: Michael Trimarchi <michael@amarulasolutions.com>
Cc: Fabio Checconi <fchecconi@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/1414142198-18552-5-git-send-email-juri.lelli@arm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
In the deboost path, right after the dl_boosted flag has been
reset, we can currently end up replenishing using -deadline
parameters of a !SCHED_DEADLINE entity. This of course causes
a bug, as those parameters are empty.
In the case depicted above it is safe to simply bail out, as
the deboosted task is going to be back to its original scheduling
class anyway.
Reported-by: Daniel Wagner <daniel.wagner@bmw-carit.de>
Tested-by: Daniel Wagner <daniel.wagner@bmw-carit.de>
Signed-off-by: Juri Lelli <juri.lelli@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: vincent@legout.info
Cc: Dario Faggioli <raistlin@linux.it>
Cc: Michael Trimarchi <michael@amarulasolutions.com>
Cc: Fabio Checconi <fchecconi@gmail.com>
Link: http://lkml.kernel.org/r/1414142198-18552-4-git-send-email-juri.lelli@arm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The race may happen when somebody is changing task_group of a forking task.
Child's cgroup is the same as parent's after dup_task_struct() (there just
memory copying). Also, cfs_rq and rt_rq are the same as parent's.
But if parent changes its task_group before it's called cgroup_post_fork(),
we do not reflect this situation on child. Child's cfs_rq and rt_rq remain
the same, while child's task_group changes in cgroup_post_fork().
To fix this we introduce fork() method, which calls sched_move_task() directly.
This function changes sched_task_group on appropriate (also its logic has
no problem with freshly created tasks, so we shouldn't introduce something
special; we are able just to use it).
Possibly, this decides the Burke Libbey's problem: https://lkml.org/lkml/2014/10/24/456
Signed-off-by: Kirill Tkhai <ktkhai@parallels.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/1414405105.19914.169.camel@tkhai
Signed-off-by: Ingo Molnar <mingo@kernel.org>
In certain circumstances the PSL (Power Service Layer, which provides
translation services for CXL hardware) can send an interrupt for a
segment miss that the kernel has already handled. This can happen if
multiple translations for the same segment are queued in the PSL before
the kernel has restarted the first translation.
The CXL driver does not expect this situation and does not check if a
segment had already been handled. This could cause a duplicate segment
table entry which in turn caused a PSL error taking down the card.
This patch fixes the issue by checking for existing entries in the
segment table that match the segment we are trying to insert, so as to
avoid inserting duplicate entries.
Signed-off-by: Ian Munsie <imunsie@au1.ibm.com>
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This patch makes copro_calculate_slb() mask the ESID by the correct mask
for 1T vs 256M segments.
This has no effect by itself as the extra bits were ignored, but it
makes debugging the segment table entries easier and means that we can
directly compare the ESID values for duplicates without needing to worry
about masking in the comparison.
This will be used to simplify a comparison in the following patch.
Signed-off-by: Ian Munsie <imunsie@au1.ibm.com>
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This moves the segment table hash calculation from cxl_load_segment()
into find_free_sste() since that is the only place it is actually used.
Signed-off-by: Ian Munsie <imunsie@au1.ibm.com>
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This patch simplifies the process of finding a free segment table entry
by disabling the secondary hash. This reduces the number of possible
entries in the segment table for a given address from 16 to 8.
Due to the large segment sizes we use it is extremely unlikely that the
secondary hash would ever have been used in practice, so this should not
have any negative impacts and may even improve performance due to the
reduced number of comparisons that software & hardware need to perform.
This patch clears the SC bit in the hardware's state register
(CXL_PSL_SR_An) to disable the secondary hash in the hardware since we
can no longer fill out entries using it.
Signed-off-by: Ian Munsie <imunsie@au1.ibm.com>
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Fengguang Wu reported a boot crash on the x86 platform
via the 0-day Linux Kernel Performance Test:
cma: dma_contiguous_reserve: reserving 31 MiB for global area
BUG: Int 6: CR2 (null)
[<41850786>] dump_stack+0x16/0x18
[<41d2b1db>] early_idt_handler+0x6b/0x6b
[<41072227>] ? __phys_addr+0x2e/0xca
[<41d4ee4d>] cma_declare_contiguous+0x3c/0x2d7
[<41d6d359>] dma_contiguous_reserve_area+0x27/0x47
[<41d6d4d1>] dma_contiguous_reserve+0x158/0x163
[<41d33e0f>] setup_arch+0x79b/0xc68
[<41d2b7cf>] start_kernel+0x9c/0x456
[<41d2b2ca>] i386_start_kernel+0x79/0x7d
(See details at: https://lkml.org/lkml/2014/10/8/708)
It is because dma_contiguous_reserve() is called before
initmem_init() in x86, the variable high_memory is not
initialized but accessed by __pa(high_memory) in
dma_contiguous_reserve().
This patch moves dma_contiguous_reserve() after initmem_init()
so that high_memory is initialized before accessed.
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Weijie Yang <weijie.yang@samsung.com>
Acked-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Marek Szyprowski <m.szyprowski@samsung.com>
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Cc: iamjoonsoo.kim@lge.com
Cc: 'Linux-MM' <linux-mm@kvack.org>
Cc: 'Weijie Yang' <weijie.yang.kh@gmail.com>
Link: http://lkml.kernel.org/r/000101cfef69%2431e528a0%2495af79e0%24%25yang@samsung.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This reverts commit bf7588a085.
Ben says although the code is not correct "[this] fix was completely
wrong and does more damages than it fixes things."
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Use daddr instead of reaching into dest.
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Alex Gartrell <agartrell@fb.com>
Signed-off-by: Simon Horman <horms@verge.net.au>
introduce two configs:
- hidden CONFIG_BPF to select eBPF interpreter that classic socket filters
depend on
- visible CONFIG_BPF_SYSCALL (default off) that tracing and sockets can use
that solves several problems:
- tracing and others that wish to use eBPF don't need to depend on NET.
They can use BPF_SYSCALL to allow loading from userspace or select BPF
to use it directly from kernel in NET-less configs.
- in 3.18 programs cannot be attached to events yet, so don't force it on
- when the rest of eBPF infra is there in 3.19+, it's still useful to
switch it off to minimize kernel size
bloat-o-meter on x64 shows:
add/remove: 0/60 grow/shrink: 0/2 up/down: 0/-15601 (-15601)
tested with many different config combinations. Hopefully didn't miss anything.
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Acked-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
