Reported by a Fedora user when they tried to build some out of tree module..
Signed-off-by: Dave Jones <davej@redhat.com>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
The original errormessage was just plain unreadable.
Sample error message after this update (not for real - I provoked it):
FATAL: drivers/net/s2io: sizeof(struct pci_device_id)=33 is not a modulo of the
size of section __mod_pci_device_table=160.
Fix definition of struct pci_device_id in mod_devicetable.h
Before a warning was generated - this is now a fatal error.
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
When we introduced -rR then aic7xxx no loger could pick up definition
of YACC&LEX from make - so do it explicit now.
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Note how any error from acpi_processor_preregister_performance is ignored.
From: bert hubert <bert.hubert@netherlabs.nl>
Signed-off-by: Dave Jones <davej@redhat.com>
Adds a __find_governor() helper function to look up a governor by
name. Also restructures some error handling to conform to the
"single-exit" model which is generally preferred for kernel code.
Signed-off-by: Jeremy Fitzhardinge <jeremy@goop.org>
Signed-off-by: Dave Jones <davej@redhat.com>
This table is only used by Ezra-T CPUs currently, and has values
for some other CPU. Fix them to match the values used by that CPU,
and for now make it clearer by renaming the variable.
Signed-off-by: Rafa³ Bilski <rafalbilski@interia.pl>
Signed-off-by: Dave Jones <davej@redhat.com>
This is changing "always true" test to something usefull.
Signed-off-by: Rafa³ Bilski <rafalbilski@interia.pl>
Signed-off-by: Dave Jones <davej@redhat.com>
This patch makes the needlessly global longhaul_walk_callback() static.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Dave Jones <davej@redhat.com>
It seems commit 32ee8c3e47 accidentially
reverted cdc9cc1d74, IOW, it reintroduced
the following compile error with CONFIG_PCI=n:
<-- snip -->
...
CC arch/i386/kernel/cpu/cpufreq/gx-suspmod.o
arch/i386/kernel/cpu/cpufreq/gx-suspmod.c: In function ‘gx_detect_chipset’:
arch/i386/kernel/cpu/cpufreq/gx-suspmod.c:193: error: implicit declaration of function ‘pci_match_id’
arch/i386/kernel/cpu/cpufreq/gx-suspmod.c:193: warning: comparison between pointer and integer
make[3]: *** [arch/i386/kernel/cpu/cpufreq/gx-suspmod.o] Error 1
<-- snip -->
This patch therefore re-adds the dependency of X86_GX_SUSPMOD on PCI.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Dave Jones <davej@redhat.com>
Without this longhaul will always fail when compiled into kernel,
as it needs to initialise after the ACPI processor module.
I lost this when I was splitting patches. Sorry.
Signed-off-by: Rafa³ Bilski <rafalbilski@interia.pl>
Signed-off-by: Dave Jones <davej@redhat.com>
There is no need to worry about local APIC.
There is need to worry about I/O APIC, because I/O APIC
is replacing good old 8259. According to Nehemiah datasheet VIA is
using 3-wire bus to connect local APIC to I/O APIC.
"[...] When IA32_APIC_BASE[11] is set to 0, processor APICs based on the 3-wire APIC
bus cannot be generally re-enabled until a system hardware reset. The 3-wire bus
looses track of arbitration that would be necessary for complete re-enabling. Certain
(local) APIC functionality can be enabled. [...]"
So we must set disable bit for each interrupt in I/O APIC registers.
Same situation as for PIC - we must poke registers direcly.
How to do this? I don't know. So at the moment it is better to fail.
Signed-off-by: Rafa³ Bilski <rafalbilski@interia.pl>
Signed-off-by: Dave Jones <davej@redhat.com>
Minimal change necessary for hardware support.
Changes in longhaul.c:
- most important - now C3 state is causing transition,
- code responsible for clearing "bus master" bit removed,
- protect bcr2 transition in the same way as longhaul.
Signed-off-by: Rafa³ Bilski <rafalbilski@interia.pl>
Signed-off-by: Dave Jones <davej@redhat.com>
I just stumbled on this bug/feature, this is how to reproduce it:
# echo 450000 > /sys/devices/system/cpu/cpu0/cpufreq/scaling_min_freq
# echo 450000 > /sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq
# echo powersave > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor
# cpufreq-info -p
450000 450000 powersave
# echo 1800000 > /sys/devices/system/cpu/cpu0/cpufreq/scaling_min_freq ; echo $?
0
# cpufreq-info -p
450000 450000 powersave
Here it is. The kernel refuses to set a min_freq higher than the
max_freq but it allows a max_freq lower than min_freq (lowering min_freq
also).
This behaviour is pretty straightforward (but undocumented) and it
doesn't return an error altough failing to accomplish the requested
action (set min_freq).
The problem (IMO) is basically that userspace is not allowed to set a
full policy atomically while the kernel always does that thus it must
enforce an ordering on operations.
The attached patch returns -EINVAL if trying to increase frequencies
starting from scaling_min_freq and documents the correct ordering of writes.
Signed-off-by: Mattia Dongili <malattia@linux.it>
Signed-off-by: Dominik Brodowski <linux at dominikbrodowski.net>
Signed-off-by: Dave Jones <davej@redhat.com>
--
* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc:
[POWERPC] Minor comment fix for misc_64.S
[POWERPC] Use H_CEDE on non-SMT
[POWERPC] force 64bit mode in fwnmi handlers to workaround firmware bugs
[POWERPC] PMAC_APM_EMU should depend on ADB_PMU
[POWERPC] Fix new interrupt code (MPIC detection)
[POWERPC] Fix new interrupt code (MPIC endianness)
[POWERPC] Add cpufreq support for Xserve G5
[POWERPC] Xserve G5 thermal control fixes
[POWERPC] Fix mem= handling when the memory limit is > RMO size
[POWERPC] More offb/bootx fixes
[POWERPC] Fix legacy_serial.c error handling on 32 bits
[POWERPC] Fix default clock for udbg_16550
[POWERPC] Fix non-MPIC CHRPs with CONFIG_SMP set
[POWERPC] Fix 32 bits warning in prom_init.c
[POWERPC] Workaround Pegasos incorrect ISA "ranges"
[POWERPC] fix up front-LED Kconfig
Since we now use the generic backlight infrastructure, I think we need to
call rivafb_bl_init before calling register_framebuffer since otherwise
rivafb_bl_init might race with the framebuffer layer already opening the
device and setting up the video mode. In this case we might end up with a
not yet fully intialized backlight (info->bl_dev still NULL) when calling
riva_bl_set_power via rivafb_set_par/rivafb_load_video_mode and the kernel
dies without any further notice during boot.
This fixes booting current git on a PB 6,1. In this case radeonfb/atyfb
would be affected too - I can fix that too but don't have any hardware to
test this on.
Cc: "Antonino A. Daplas" <adaplas@pol.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
CONFIG_FB_NVIDIA already depends on CONFIG_PCI in drivers/video/Kconfig.
Driver does an extra ``sanity check'' which is then redundant.
Signed-off-by: Arthur Othieno <apgo@patchbomb.org>
Cc: Antonino Daplas <adaplas@pol.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch fixes several problems:
- The legacy backlight value might be set at interrupt time. Introduced
a worker to prevent it from directly calling the backlight code.
- via-pmu allows the backlight to be grabbed, in which case we need to
prevent other kernel code from changing the brightness.
- Don't send PMU requests in via-pmu-backlight when the machine is about
to sleep or waking up.
- More Kconfig fixes.
Signed-off-by: Michael Hanselmann <linux-kernel@hansmi.ch>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: "Antonino A. Daplas" <adaplas@pol.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Many IBM Thinkpad T4* models and some R* and X* with radeon video cards draw
too much power when suspended to RAM, reducing drastically the battery
lifetime. The solution is to enable suspend-to-D2 on these machines. They
are whitelisted through their subsystem vendor/device ID. This fixes
http://bugzilla.kernel.org/show_bug.cgi?id=3022
The patch introduces a framework to alter the pm_mode and reinit_func fields
of the radeonfb_info structure based on a whitelist. This should facilitate
future hardware-dependent workarounds. The workaround for the Samsung P35
that is already in the radeonfb code has been rewritten using this framework.
The behavior can be overridden with module options:
i) video=radeonfb:force_sleep=1
enable suspend-to-D2 also on non-whitelisted machines (useful for
testing new notebook models),
ii) video=radeonfb:ignore_devlist=1
Disable checking the whitelist and do not apply any workarounds.
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: "Antonino A. Daplas" <adaplas@pol.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The backlight and lcd subsystems can be notified by the framebuffer layer
of blanking events. However, these subsystems, as a whole, can function
independently from the framebuffer layer. But in order to enable to the
lcd and backlight subsystems, the framebuffer has to be compiled also,
effectively sucking in a huge amount of unneeded code.
To prevent dependency problems, separate out the framebuffer notification
mechanism from the framebuffer layer and permanently link it to the kernel.
Signed-off-by: Antonino Daplas <adaplas@pol.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Based on a bug report from Russ Ross <russruss@gmail.com>
According to the spec:
"The remove request asks the file server both to remove the file
represented by fid and to clunk the fid, even if the remove fails."
but the Linux client seems to expect the fid to be valid after a failed
remove attempt. Specifically, I'm getting this behavior when attempting to
remove a non-empty directory.
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Use preferred email address. Remove sf.net project reference. It is no
longer used.
Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Signed-off-by: Russ Ross <russross@gmail.com>
Signed-off-by: Eric Van Hensbergen <ericvh@gmail.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
For files other than IFREG, nobh option doesn't make sense. Modifications
to them are journalled and needs buffer heads to do that. Without this
patch, we get kernel oops in page_buffers().
Signed-off-by: Badari Pulavarty <pbadari@us.ibm.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
kernel/timer.c defines a (per-cpu) pointer to tvec_base_t, but initializes
it using { &a_tvec_base_t }, which sparse warns about; change this to just
&a_tvec_base_t.
Signed-off-by: Josh Triplett <josh@freedesktop.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Commit 7b2fd697427e73c81d5fa659efd91bd07d303b0e in the historical GIT tree
stopped calling the readdir member of a file_operations struct with the big
kernel lock held, and fixed up all the readdir functions to do their own
locking. However, that change added calls to unlock_kernel() in
vxfs_readdir, but no call to lock_kernel(). Fix this by adding a call to
lock_kernel().
Signed-off-by: Josh Triplett <josh@freedesktop.org>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
I spent a long time the other day trying to examine an initrd image on a
fedora core 5 system because the initrd.txt file is apparently obsolete.
Here is a patch which I hope will reduce future confusion for others.
Signed-off-by: Thomas Horsley <tom.horsley@ccur.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Clean up ipc/msg.c to conform to Documentation/CodingStyle. (before it was
an inconsistent hodepodge of various coding styles)
Verified that the before/after .o's are identical.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
It is entirely possible (though rare) that jiffies half-wraps around, while a
dentry/inode remains in the cache. This could mean that the dentry/inode is
not invalidated for another half wraparound-time.
To get around this problem, use 64-bit jiffies. The only problem with this is
that dentry->d_time is 32 bits on 32-bit archs. So use d_fsdata as the high
32 bits. This is an ugly hack, but far simpler, than having to allocate
private data just for this purpose.
Since 64-bit jiffies can be assumed never to wrap around, simple comparison
can be used, and a zero time value can represent "invalid".
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
An attribute and entry timeout of zero should mean, that the entity is
invalidated immediately after the operation. Previously invalidation only
happened at the next clock tick.
Reported and tested by Craig Davies.
Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
CFA needs to be adjusted upwards for push, and downwards for pop.
arch/i386/kernel/entry.S gets it wrong in one place.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Acked-by: Jan Beulich <jbeulich@novell.com>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The latest toolchains can produce a new ELF section in DSOs and
dynamically-linked executables. The new section ".gnu.hash" replaces
".hash", and allows for more efficient runtime symbol lookups by the
dynamic linker. The new ld option --hash-style={sysv|gnu|both} controls
whether to produce the old ".hash", the new ".gnu.hash", or both. In some
new systems such as Fedora Core 6, gcc by default passes --hash-style=gnu
to the linker, so that a standard invocation of "gcc -shared" results in
producing a DSO with only ".gnu.hash". The new ".gnu.hash" sections need
to be dealt with the same way as ".hash" sections in all respects; only the
dynamic linker cares about their contents. To work with older dynamic
linkers (i.e. preexisting releases of glibc), a binary must have the old
".hash" section. The --hash-style=both option produces binaries that a new
dynamic linker can use more efficiently, but an old dynamic linker can
still handle.
The new section runs afoul of the custom linker scripts used to build vDSO
images for the kernel. On ia64, the failure mode for this is a boot-time
panic because the vDSO's PT_IA_64_UNWIND segment winds up ill-formed.
This patch addresses the problem in two ways.
First, it mentions ".gnu.hash" in all the linker scripts alongside ".hash".
This produces correct vDSO images with --hash-style=sysv (or old tools),
with --hash-style=gnu, or with --hash-style=both.
Second, it passes the --hash-style=sysv option when building the vDSO
images, so that ".gnu.hash" is not actually produced. This is the most
conservative choice for compatibility with any old userland. There is some
concern that some ancient glibc builds (though not any known old production
system) might choke on --hash-style=both binaries. The optimizations
provided by the new style of hash section do not really matter for a DSO
with a tiny number of symbols, as the vDSO has. If someone wants to use
=gnu or =both for their vDSO builds and worry less about that
compatibility, just change the option and the linker script changes will
make any choice work fine.
Signed-off-by: Roland McGrath <roland@redhat.com>
Cc: "Luck, Tony" <tony.luck@intel.com>
Cc: Kyle McMartin <kyle@mcmartin.ca>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Jeff Dike <jdike@addtoit.com>
Cc: Andi Kleen <ak@muc.de>
Cc: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The geode hwrng leaks an iomapped resource, if hwrng_register() fails.
This fixes it.
Signed-off-by: Michael Buesch <mb@bu3sch.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The intel hwrng leaks an iomapped resource, if hwrng_register() failes.
This fixes it.
Signed-off-by: Michael Buesch <mb@bu3sch.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
kmem_cache_alloc() was documented twice, but kmem_cache_zalloc() never.
Fix this obvious typo to get things right.
Signed-off-by: Rolf Eike Beer <eike-kernel@sf-tec.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
In order to prevent Doc Rot, this patch adds a reference to the design
document for rtmutex.c in rtmutex.c. So when someone needs to update or
change the design of that file they will know that a document actually
exists that explains the design (helping them change it), and hopefully
that they will update the document if they too change the design.
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Acked-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
There is currently no affected user in the tree, but usage is less
surprising that way.
Signed-off-by: Uwe Zeisberger <Uwe_Zeisberger@digi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The recent changes from irqtrace feature has added overheads to
local_bh_disable and local_bh_enable that reduces UDP performance across
x86_64 and IA64, even though IA64 does not support the irqtrace feature.
Patch in question is
[PATCH]lockdep: irqtrace subsystem, core
http://www.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=c
ommit;h=de30a2b355ea85350ca2f58f3b9bf4e5bc007986
Prior to this patch, local_bh_disable was a short macro. Now it is a
function which calls __local_bh_disable with added irq flags save and
restore. The irq flags save and restore were also added to
local_bh_enable, probably for injecting the trace irqs code.
This overhead is on the generic code path across all architectures. On a
IA_64 test machine (Itanium-2 1.6 GHz) running a benchmark like netperf's
UDP streaming test, the added overhead results in a drop of 3% in
throughput, as udp_sendmsg calls the local_bh_enable/disable several times.
Other workloads that have heavy usages of local_bh_enable/disable could
also be affected. The patch ideally should not have affected IA-64
performance as it does not have IRQ tracing support. A significant portion
of the overhead is in the added irq flags save and restore, which I think
is not needed if IRQ tracing is unused. A suggested patch is attached
below that recovers the lost performance. However, the "ifdef"s in the
patch are a bit ugly.
Signed-off-by: Tim Chen <tim.c.chen@intel.com>
Acked-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Change "Thrid" into "Third", and realign similarly to other entries.
Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Cc: <device@lanana.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
ufs_symlink, in one of its error paths, calls unlock_kernel without ever
having called lock_kernel(); fix this by creating and jumping to a new
label out_notlocked rather than the out label used after calling
lock_kernel().
Signed-off-by: Josh Triplett <josh@freedesktop.org>
Cc: Evgeniy Dushistov <dushistov@mail.ru>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The EFS filesystem does not have an entry in MAINTAINERS; add one, giving
the EFS filesystem and listing the status as Orphan, per the note on that
page saying "I'm no longer actively maintaining EFS".
Signed-off-by: Josh Triplett <josh@freedesktop.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
If efs_symlink_readpage hits the -ENAMETOOLONG error path, it will call
unlock_kernel without ever having called lock_kernel(); fix this by
creating and jumping to a new label fail_notlocked rather than the fail
label used after calling lock_kernel().
Signed-off-by: Josh Triplett <josh@freedesktop.org>
Cc: Marcelo Tosatti <marcelo.tosatti@cyclades.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Commit 398c53a757702e1e3a7a2c24860c7ad26acb53ed (in the historical GIT
tree) moved the lock_kernel() in coda_open after the allocation of a
coda_file_info struct, but left an unlock_kernel() in the allocation
failure error path; remove it.
Signed-off-by: Josh Triplett <josh@freedesktop.org>
Acked-by: Jan Harkes <jaharkes@cs.cmu.edu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
swsusp is unable to suspend my machine (DTK FortisPro TOP-5A notebook) with
kernel 2.6.17.5 because it's unable to suspend PNP device 00:16 (mouse).
The problem is in PNP BIOS. pnp_bus_suspend() calls pnp_stop_dev() for the
device if the device can be disabled according to pnp_can_disable(). The
problem is that pnpbios_disable_resources() returns -EPERM if the device is
not dynamic (!pnpbios_is_dynamic()) but insert_device() happily sets
PNP_DISABLE capability/flag even if the device is not dynamic. So we try
to disable non-dynamic devices which will fail. This patch prevents
insert_device() from setting PNP_DISABLE if the device is not dynamic and
fixes suspend on my system.
Signed-off-by: Ondrej Zary <linux@rainbow-software.org>
Cc: Pavel Machek <pavel@ucw.cz>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This is a real deadlock, a nice complex one:
(warning: long explanation follows so that Andrew can have a complete
patch description)
it's an ABCDA deadlock:
A iprune_mutex
B inode->inotify_mutex
C ih->mutex
D dev->ev_mutex
The AB relationship comes straight from invalidate_inodes()
int invalidate_inodes(struct super_block * sb)
{
int busy;
LIST_HEAD(throw_away);
mutex_lock(&iprune_mutex);
spin_lock(&inode_lock);
inotify_unmount_inodes(&sb->s_inodes);
where inotify_umount_inodes() takes the
mutex_lock(&inode->inotify_mutex);
The BC relationship comes directly from inotify_find_update_watch():
s32 inotify_find_update_watch(struct inotify_handle *ih, struct inode *inode,
u32 mask)
{
...
mutex_lock(&inode->inotify_mutex);
mutex_lock(&ih->mutex);
The CD relationship comes from inotify_rm_wd:
inotify_rm_wd does
mutex_lock(&inode->inotify_mutex);
mutex_lock(&ih->mutex)
and then calls inotify_remove_watch_locked() which calls
notify_dev_queue_event() which does
mutex_lock(&dev->ev_mutex);
(this strictly is a BCD relationship)
The DA relationship comes from the most interesting part:
[<ffffffff8022d9f2>] shrink_icache_memory+0x42/0x270
[<ffffffff80240dc4>] shrink_slab+0x11d/0x1c9
[<ffffffff802b5104>] try_to_free_pages+0x187/0x244
[<ffffffff8020efed>] __alloc_pages+0x1cd/0x2e0
[<ffffffff8025e1f8>] cache_alloc_refill+0x3f8/0x821
[<ffffffff8020a5e5>] kmem_cache_alloc+0x85/0xcb
[<ffffffff802db027>] kernel_event+0x2e/0x122
[<ffffffff8021d61c>] inotify_dev_queue_event+0xcc/0x140
inotify_dev_queue_event schedules a kernel_event which does a
kmem_cache_alloc( , GFP_KERNEL) which may try to shrink slabs, including
the inode cache .. which then takes iprune_mutex.
And voila, there is an AB, a BC, a CD relationship (even a direct BCD),
and also now a DA relationship -> a circular type AB-BA deadlock but
involving 4 locks.
The solution is simple: kernel_event() is NOT allowed to use GFP_KERNEL,
but must use GFP_NOFS to not cause recursion into the VFS.
Signed-off-by: Arjan van de Ven <arjan@linux.intel.com>
Acked-by: Ingo Molnar <mingo@elte.hu>
Acked-by: Robert Love <rml@novell.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
