The kernel's implementation of notifier chains is unsafe. There is no
protection against entries being added to or removed from a chain while the
chain is in use. The issues were discussed in this thread:
http://marc.theaimsgroup.com/?l=linux-kernel&m=113018709002036&w=2
We noticed that notifier chains in the kernel fall into two basic usage
classes:
"Blocking" chains are always called from a process context
and the callout routines are allowed to sleep;
"Atomic" chains can be called from an atomic context and
the callout routines are not allowed to sleep.
We decided to codify this distinction and make it part of the API. Therefore
this set of patches introduces three new, parallel APIs: one for blocking
notifiers, one for atomic notifiers, and one for "raw" notifiers (which is
really just the old API under a new name). New kinds of data structures are
used for the heads of the chains, and new routines are defined for
registration, unregistration, and calling a chain. The three APIs are
explained in include/linux/notifier.h and their implementation is in
kernel/sys.c.
With atomic and blocking chains, the implementation guarantees that the chain
links will not be corrupted and that chain callers will not get messed up by
entries being added or removed. For raw chains the implementation provides no
guarantees at all; users of this API must provide their own protections. (The
idea was that situations may come up where the assumptions of the atomic and
blocking APIs are not appropriate, so it should be possible for users to
handle these things in their own way.)
There are some limitations, which should not be too hard to live with. For
atomic/blocking chains, registration and unregistration must always be done in
a process context since the chain is protected by a mutex/rwsem. Also, a
callout routine for a non-raw chain must not try to register or unregister
entries on its own chain. (This did happen in a couple of places and the code
had to be changed to avoid it.)
Since atomic chains may be called from within an NMI handler, they cannot use
spinlocks for synchronization. Instead we use RCU. The overhead falls almost
entirely in the unregister routine, which is okay since unregistration is much
less frequent that calling a chain.
Here is the list of chains that we adjusted and their classifications. None
of them use the raw API, so for the moment it is only a placeholder.
ATOMIC CHAINS
-------------
arch/i386/kernel/traps.c: i386die_chain
arch/ia64/kernel/traps.c: ia64die_chain
arch/powerpc/kernel/traps.c: powerpc_die_chain
arch/sparc64/kernel/traps.c: sparc64die_chain
arch/x86_64/kernel/traps.c: die_chain
drivers/char/ipmi/ipmi_si_intf.c: xaction_notifier_list
kernel/panic.c: panic_notifier_list
kernel/profile.c: task_free_notifier
net/bluetooth/hci_core.c: hci_notifier
net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_chain
net/ipv4/netfilter/ip_conntrack_core.c: ip_conntrack_expect_chain
net/ipv6/addrconf.c: inet6addr_chain
net/netfilter/nf_conntrack_core.c: nf_conntrack_chain
net/netfilter/nf_conntrack_core.c: nf_conntrack_expect_chain
net/netlink/af_netlink.c: netlink_chain
BLOCKING CHAINS
---------------
arch/powerpc/platforms/pseries/reconfig.c: pSeries_reconfig_chain
arch/s390/kernel/process.c: idle_chain
arch/x86_64/kernel/process.c idle_notifier
drivers/base/memory.c: memory_chain
drivers/cpufreq/cpufreq.c cpufreq_policy_notifier_list
drivers/cpufreq/cpufreq.c cpufreq_transition_notifier_list
drivers/macintosh/adb.c: adb_client_list
drivers/macintosh/via-pmu.c sleep_notifier_list
drivers/macintosh/via-pmu68k.c sleep_notifier_list
drivers/macintosh/windfarm_core.c wf_client_list
drivers/usb/core/notify.c usb_notifier_list
drivers/video/fbmem.c fb_notifier_list
kernel/cpu.c cpu_chain
kernel/module.c module_notify_list
kernel/profile.c munmap_notifier
kernel/profile.c task_exit_notifier
kernel/sys.c reboot_notifier_list
net/core/dev.c netdev_chain
net/decnet/dn_dev.c: dnaddr_chain
net/ipv4/devinet.c: inetaddr_chain
It's possible that some of these classifications are wrong. If they are,
please let us know or submit a patch to fix them. Note that any chain that
gets called very frequently should be atomic, because the rwsem read-locking
used for blocking chains is very likely to incur cache misses on SMP systems.
(However, if the chain's callout routines may sleep then the chain cannot be
atomic.)
The patch set was written by Alan Stern and Chandra Seetharaman, incorporating
material written by Keith Owens and suggestions from Paul McKenney and Andrew
Morton.
[jes@sgi.com: restructure the notifier chain initialization macros]
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Chandra Seetharaman <sekharan@us.ibm.com>
Signed-off-by: Jes Sorensen <jes@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Modify well over a dozen mempool users to call mempool_create_slab_pool()
rather than calling mempool_create() with extra arguments, saving about 30
lines of code and increasing readability.
Signed-off-by: Matthew Dobson <colpatch@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch changes several mempool users, all of which are basically just
wrappers around kmalloc(), to use the common mempool_kmalloc/kfree, rather
than their own wrapper function, removing a bunch of duplicated code.
Signed-off-by: Matthew Dobson <colpatch@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
drivers/scsi/sd.c: In function `sd_store_cache_type':
drivers/scsi/sd.c:193: warning: comparison of distinct pointer types lacks a cast
Cc: James Bottomley <James.Bottomley@steeleye.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
MODULE_PARM was actually breaking: recent gcc version optimize them out as
unused. It's time to replace the last users, which are generally in the
most unloved drivers anyway.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/libata-dev:
[PATCH] libata: Remove dependence on host_set->dev for SAS
[PATCH] libata: ata_scsi_ioctl cleanup
[PATCH] libata: ata_scsi_queuecmd cleanup
[libata] export ata_dev_pair; trim trailing whitespace
[PATCH] libata: add ata_dev_pair helper
[PATCH] Make libata not powerdown drivers on PM_EVENT_FREEZE.
[PATCH] libata: make ata_set_mode() responsible for failure handling
[PATCH] libata: use ata_dev_disable() in ata_bus_probe()
[PATCH] libata: implement ata_dev_disable()
[PATCH] libata: check if port is disabled after internal command
[PATCH] libata: make per-dev transfer mode limits per-dev
[PATCH] libata: add per-dev pio/mwdma/udma_mask
[PATCH] libata: implement ata_unpack_xfermask()
[libata] Move some bmdma-specific code to libata-bmdma.c
[libata sata_uli] kill scr_addr abuse
[libata sata_nv] eliminate duplicate codepaths with iomap
[libata sata_nv] cleanups: convert #defines to enums; remove in-file history
[libata sata_sil24] cleanups: use pci_iomap(), kzalloc()
switched to ioremap(), cleaned the probing up a bit.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Remove some of the dependence on the host_set struct
in preparation for supporting SAS HBAs. Adds a struct device
pointer to the ata_port struct.
Signed-off-by: Brian King <brking@us.ibm.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
In preparation for SAS, kill some unnecessary code in ata_scsi_ioctl
to find the ATA port and device given the scsi_device. Neither local
is used in the function.
Signed-off-by: Brian King <brking@us.ibm.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Encapsulate part of ata_scsi_queuecmd so that it can be
reused by future SAS patches.
Signed-off-by: Brian King <brking@us.ibm.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
At the moment libata doesn't pass pm_message_t down ata_device_suspend.
This causes drives to be powered down when we just want a freeze,
causing unnecessary wear and tear. This patch gets pm_message_t passed
down so that it can be used to determine whether to power down the
drive.
Signed-off-by: Nigel Cunningham <nigel@suspend2.net>
drivers/scsi/libata-core.c | 5 +++--
drivers/scsi/libata-scsi.c | 4 ++--
drivers/scsi/scsi_sysfs.c | 2 +-
include/linux/libata.h | 4 ++--
include/scsi/scsi_host.h | 2 +-
5 files changed, 9 insertions(+), 8 deletions(-)
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Make ata_set_mode() responsible for determining whether to take port
or device offline on failure. ata_dev_set_xfermode() and
ata_dev_set_mode() indicate error to the caller instead of disabling
port directly on failure. Also, for consistency, ata_dev_present()
check is done in ata_set_mode() instead of ata_dev_set_mode().
Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
We may or may not disable a device after ata_dev_configure() fails.
Kill 'not supported, ignoring' message in ata_dev_configure() and use
ata_dev_disable() in ata_bus_probe().
Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
This patch implements ata_dev_disable() which prints a warning message
and takes @dev offline. Currently, this is done by explicitly
incrementing dev->class with case-by-case warning messages. Giving
user clear indication when libata gives up will be more important as
libata will be doing more retries.
Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
libata core is being changed to disallow port/device disable on lower
layers. However, some LLDDs (sata_mv) directly disable port on
command failure. This patch makes ata_exec_internal() check whether a
port got disabled after an internal command. If it is, AC_ERR_SYSTEM
is added to err_mask and the port gets re-enabled.
As internal command failure results in device disable for drivers
which don't implement newer reset/EH callbacks, this change results in
no behavior change for single device per port controllers. For
slave-possible LLDDs which disable port on command failure, (1) such
drivers don't exist currently, (2) issuing command to the other device
of once-disabled port shouldn't result in catastrophe even if such
driver exists. So, this should be enough as a temporary measure.
Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Now that each ata_device has xfer masks, per-dev limits can be made
per-dev instead of per-port. Make per-dev limits per-dev.
Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Add per-dev pio/mwdma/udma_mask. All transfer mode limits used to be
applied to ap->*_mask which unnecessarily restricted other devices
sharing the port. This change will also benefit later EH speed down
and hotplug.
Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Semaphore to mutex conversion.
The conversion was generated via scripts, and the result was validated
automatically via a script as well.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Jens Axboe <axboe@suse.de>
Cc: Bartlomiej Zolnierkiewicz <B.Zolnierkiewicz@elka.pw.edu.pl>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Since early 2.4.x all cdrom drivers implement the block_device methods
themselves, so they can handle additional ioctls directly instead of going
through the cdrom layer.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: Jens Axboe <axboe@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
sata_uli was storing PCI config addresses in a variable intended for
port addresses, a variable soon to become void __iomem *.
Update the driver to store the SCR address, found in PCI config space,
in the driver-private data area.
If the debug driver is built-in, link it in last, so that any real
drivers will probe first, rather than having the debug driver pick the
first scsi slots..
Signed-off-by: Douglas Gilbert <dougg@torque.net>
Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Acked-by: James Bottomley <James.Bottomley@SteelEye.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Now that libata is smart enought to handle both soft and hard resets,
add softreset method.
Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Not much to say here except that some drives have fixed and bad firmware
Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
The current code follows the spec but uses an overlong delay. This would
be great if the hardware did. Several vendors however forget the D7
pulldown. Fortunately 0xFF isnt a sane reset state so we can use it to
skip detection as is done in drivers/ide. (ie this is a tested solution
over a long time)
Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
sg increments the refcount of constituent pages in its higher order memory
allocations when they are about to be mapped by userspace. This is done so
the subsequent get_page/put_page when doing the mapping and unmapping does not
free the page.
Move over to the preferred way, that is, using compound pages instead. This
fixes a whole class of possible obscure bugs where a get_user_pages on a
constituent page may outlast the user mappings or even the driver.
Signed-off-by: Nick Piggin <npiggin@suse.de>
Cc: Hugh Dickins <hugh@veritas.com>
Cc: Douglas Gilbert <dougg@torque.net>
Cc: James Bottomley <James.Bottomley@steeleye.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
SLAB_NO_REAP is documented as an option that will cause this slab not to be
reaped under memory pressure. However, that is not what happens. The only
thing that SLAB_NO_REAP controls at the moment is the reclaim of the unused
slab elements that were allocated in batch in cache_reap(). Cache_reap()
is run every few seconds independently of memory pressure.
Could we remove the whole thing? Its only used by three slabs anyways and
I cannot find a reason for having this option.
There is an additional problem with SLAB_NO_REAP. If set then the recovery
of objects from alien caches is switched off. Objects not freed on the
same node where they were initially allocated will only be reused if a
certain amount of objects accumulates from one alien node (not very likely)
or if the cache is explicitly shrunk. (Strangely __cache_shrink does not
check for SLAB_NO_REAP)
Getting rid of SLAB_NO_REAP fixes the problems with alien cache freeing.
Signed-off-by: Christoph Lameter <clameter@sgi.com>
Cc: Pekka Enberg <penberg@cs.helsinki.fi>
Cc: Manfred Spraul <manfred@colorfullife.com>
Cc: Mark Fasheh <mark.fasheh@oracle.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Alan noted: "bmdma may be zero but the bmdma_irq_clear function gets
called even in this case during pure PIO operation. Check we have a
bmdma before we use it."
I fixed this by adding a check for zero. While was I there, I fixed the
non-standard indentation of the small function's code.
Signed-off-by: Jeff Garzik <jeff@garzik.org>
The first performs the simplex clearing relevant to some chipsets that
report simplex by default but can in fact do more if poked. The second
is used to strip DMA modes from a PCI control with no BAR4 allocation.
Signed-off-by: Jeff Garzik <jeff@garzik.org>
