This idea came from Nicolas, Eric Miao produced an initial version,
which was then rewritten into this.
Patch the physical to virtual translations at runtime. As we modify
the code, this makes it incompatible with XIP kernels, but allows us
to achieve this with minimal loss of performance.
As many translations are of the form:
physical = virtual + (PHYS_OFFSET - PAGE_OFFSET)
virtual = physical - (PHYS_OFFSET - PAGE_OFFSET)
we generate an 'add' instruction for __virt_to_phys(), and a 'sub'
instruction for __phys_to_virt(). We calculate at run time (PHYS_OFFSET
- PAGE_OFFSET) by comparing the address prior to MMU initialization with
where it should be once the MMU has been initialized, and place this
constant into the above add/sub instructions.
Once we have (PHYS_OFFSET - PAGE_OFFSET), we can calculate the real
PHYS_OFFSET as PAGE_OFFSET is a build-time constant, and save this for
the C-mode PHYS_OFFSET variable definition to use.
At present, we are unable to support Realview with Sparsemem enabled
as this uses a complex mapping function, and MSM as this requires a
constant which will not fit in our math instruction.
Add a module version magic string for this feature to prevent
incompatible modules being loaded.
Tested-by: Tony Lindgren <tony@atomide.com>
Reviewed-by: Nicolas Pitre <nicolas.pitre@linaro.org>
Tested-by: Nicolas Pitre <nicolas.pitre@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
These are needed for CONFIG_ARM_PATCH_PHYS_VIRT to work.
Signed-off-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
head.S makes use of PHYS_OFFSET. When it becomes a variable, the
assembler won't understand this. Compute PHYS_OFFSET by the following
method. This code is linked at its virtual address, but run at before
the MMU is enabled, so at his physical address.
1: .long .
.long PAGE_OFFSET
adr r0, 1b @ r0 = physical ','
ldmia r0, {r1, r2} @ r1 = virtual '.', r2 = PAGE_OFFSET
sub r1, r0, r1 @ r1 = physical-virtual
add r2, r2, r1 @ r2 = PAGE_OFFSET + physical-virtual
@ := PHYS_OFFSET.
Switch XIP users of PHYS_OFFSET to use PLAT_PHYS_OFFSET - we can't
use this method for XIP kernels as the code doesn't execute in RAM.
Tested-by: Tony Lindgren <tony@atomide.com>
Reviewed-by: Nicolas Pitre <nicolas.pitre@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
As PHYS_OFFSET will be becoming a variable, we can't have it used in
initializers nor assembly code. Replace those in generic code with
a run-time initialization. Replace those in platform code using the
individual platform specific PLAT_PHYS_OFFSET.
Acked-by: Nicolas Pitre <nicolas.pitre@linaro.org>
Acked-by: Tony Lindgren <tony@atomide.com>
Acked-by: Kukjin Kim <kgene.kim@samsung.com>
Acked-by: David Brown <davidb@codeaurora.org>
Acked-by: Eric Miao <eric.y.miao@gmail.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
This uncouple PHYS_OFFSET from the platform definitions, thereby
facilitating run-time computation of the physical memory offset.
Acked-by: Nicolas Pitre <nicolas.pitre@linaro.org>
Acked-by: Viresh Kumar <viresh.kumar@st.com>
Acked-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Acked-by: Magnus Damm <damm@opensource.se>
Acked-by: Tony Lindgren <tony@atomide.com>
Acked-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Acked-by: Wan ZongShun <mcuos.com@gmail.com>
Acked-by: Kukjin Kim <kgene.kim@samsung.com>
Acked-by: Eric Miao <eric.y.miao@gmail.com>
Acked-by: Jiandong Zheng <jdzheng@broadcom.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Since the debug macros no longer depend on the machine type information,
the machine type lookup can be deferred to setup_arch() in setup.c which
simplifies the code somewhat.
We also move the __error_a functionality into setup.c for displaying a
message when a bad machine ID is passed to the kernel via the LL debug
code. We also log this into the kernel ring buffer which makes it
possible to retrieve the message via a debugger.
Original idea from Grant Likely.
Acked-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
* 'omap-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap-2.6:
arch/arm/mach-omap2/dma.c: Convert IS_ERR result to PTR_ERR
arm: omap2: mux: fix compile warning
omap1: Simplify use of omap_irq_flags
omap2+: Fix unused variable warning for omap_irq_base
Fix kernel-doc warning in kernel.h from commit 7ef88ad561
("BUILD_BUG_ON: make it handle more cases"):
Warning(include/linux/kernel.h:605): No description found for parameter 'condition'
Warning(include/linux/kernel.h:605): Excess function parameter 'cond' description in 'BUILD_BUG_ON'
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Patches should keep coming through Rusty but it helps if I'm Cc'd as
well.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* 'usb-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6: (42 commits)
usb: gadget: composite: avoid access beyond array max length
USB: serial: handle Data Carrier Detect changes
USB: gadget: Fix endpoint representation in ci13xxx_udc
USB: gadget: Fix error path in ci13xxx_udc gadget probe function
usb: pch_udc: Fix the worning log issue at gadget driver remove
USB: serial: Updated support for ICOM devices
USB: ehci-mxc: add work-around for efika mx/sb bug
USB: unbreak ehci-mxc on otg port of i.MX27
drivers: update to pl2303 usb-serial to support Motorola cables
USB: adding USB support for Cinterion's HC2x, EU3 and PH8 products
USB serial: add missing .usb_driver field in serial drivers
USB: ehci-fsl: Fix 'have_sysif_regs' detection
USB: g_printer: fix bug in module parameter definitions
USB: g_printer: fix bug in unregistration
USB: uss720: remove duplicate USB device
MAINTAINERS: add ueagle-atm entry
USB: EHCI: fix DMA deallocation bug
USB: pch_udc: support new device ML7213 IOH
usb: pch_udc: Fixed issue which does not work with g_serial
usb: set ep_dev async suspend should be later than device_initialize
...
* 'staging-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging-2.6: (26 commits)
staging: r8712u: Add new device IDs
staging: brcm80211: fix suspend/resume issue in brcmsmac
staging: brcm80211: remove assert to avoid panic since 2.6.37 kernel
Staging: iio: Aditional fixpoint formatted output bugfix
staging: usbip: vhci: use urb->dev->portnum to find port
staging: usbip: vhci: handle EAGAIN from SO_RCVTIMEO
staging: usbip: vhci: friendly log messages for connection errors
staging: usbip: vhci: refuse to enqueue for dead connections
staging: usbip: vhci: give back URBs from in-flight unlink requests
staging: usbip: vhci: update reference count for usb_device
staging: usbip: stub: update refcounts for devices and interfaces
staging: tidspbridge: replace mbox callback with notifier_call
staging: comedi: ni_labpc: Use shared IRQ for PCMCIA card
Staging: speakup: &&/|| confusion in silent_store()
iio: Fixpoint formatted output bugfix
staging: rt2860: Fix incorrect netif_stop_queue usage warning
staging: r8712u: Fix memory leak in firmware loading
staging: tidspbridge: configure full L1 MMU range
staging: rt2870sta: Add ID for Linksys WUSB100v2
Staging: xgfib: put parenthesis in the right place
...
* 'tty-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty-2.6:
tty/serial: fix apbuart build
n_hdlc: fix read and write locking
serial: unbreak billionton CF card
tty: use for_each_console() and WARN() on sysfs failures
vt: fix issue when fbcon wants to takeover a second time.
Fix up trivial conflict in drivers/tty/tty_io.c
Allow non-ARM SMP processors to use the SMP_ON_UP feature. CPUs
supporting SMP must have the new CPU ID format, so check for this first.
Then check for ARM11MPCore, which fails the MPIDR check. Lastly check
the MPIDR reports multiprocessing extensions and that the CPU is part of
a multiprocessing system.
Cc: <stable@kernel.org>
Reported-and-Tested-by: Stephen Boyd <sboyd@codeaurora.org>
Acked-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Ensure that the ISA/PCI IO space accessors are properly ordered on
ARMv6+ architectures. These should always be ordered with respect to
all other accesses.
This also fixes __iormb() and __iowmb() not being visible to ioread/
iowrite if a platform defines its own MMIO accessors.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Disable the initrd if the passed address already overlaps the reserved
region. This avoids oopses on Netwinders when NeTTrom tells the kernel
that an initrd is located at mem+4MB, but this overlaps the BSS,
resulting in the kernels in-use BSS being freed.
This should be applied to v2.6.37-stable.
Cc: <stable@kernel.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
0ea1293 (arm: return both physical and virtual addresses from addruart)
changed the way the 'addruart' worked, making it return both the virt
and phys addresses. Unfortunately, for footbridge, these were reversed.
Fix that. Tested on Netwinder.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
We should not report incomplete blocks on error. Return the number of
bytes successfully transferred, rounded down to the nearest block.
Acked-by: Linus Walleij <linus.walleij@stericsson.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
When we encounter an error, make sure we complete the transaction
otherwise we'll leave the request dangling.
Acked-by: Linus Walleij <linus.walleij@stericsson.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Do not respond with -EINVAL to EVIOCGKEYCODE for not-yet-mapped
scancodes, but rather return KEY_RESERVED.
This fixes breakage with Ubuntu's input-kbd utility that stopped
returning full keymaps for remote controls.
Tested-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Tested-by: Mark Lord <kernel@teksavvy.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Since check_prlimit_permission always fails in the case of SUID/GUID
processes, such processes are not able to read or set their own limits.
This commit changes this by assuming that process can always read/change
its own limits.
Signed-off-by: Kacper Kornet <kornet@camk.edu.pl>
Acked-by: Jiri Slaby <jslaby@suse.cz>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
In ntfs_mft_record_alloc() when mapping the new extent mft record with
map_extent_mft_record() we overwrite @m with the return value and on
error, we then try to use the old @m but that is no longer there as @m
now contains an error code instead so we crash when dereferencing the
error code as if it were a pointer.
The simple fix is to use a temporary variable to store the return value
thus preserving the original @m for later use. This is a backport from
the commercial Tuxera-NTFS driver and is well tested...
Thanks go to Julia Lawall for pointing this out (whilst I had fixed it
in the commercial driver I had failed to fix it in the Linux kernel).
Signed-off-by: Anton Altaparmakov <anton@tuxera.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/cmarinas/linux-2.6-cm:
kmemleak: Allow kmemleak metadata allocations to fail
kmemleak: remove memset by using kzalloc
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6:
ALSA: HDA: Fix automute on Thinkpad L412/L512
ALSA: HDA: Fix dmesg output of HDMI supported bits
ALSA: fix invalid hardware.h include in ac97c for AVR32 architecture
ASoC: correct link specifications for corgi, poodle and spitz
ASoC: Samsung: Fix outdated cpu_dai_name for s3c24xx i2s
ASoC: Fix codec device id format used by some dai_links
ALSA: azt3328 - fix broken AZF_FMT_XLATE macro
ALSA: Xonar, CS43xx: Don't overrun static array
ASoC: Handle low measured DC offsets for wm_hubs devices
ASoC: da8xx/omap-l1xx: match codec_name with i2c ids
ASoC: WM8994: fix wrong value in tristate function
ASoC: WM8995: Fix incorrect use of snd_soc_update_bits()
* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev:
libata: set queue DMA alignment to sector size for ATAPI too
libata: DVR-212D can't do SETXFER DVD-RW DVR-212D
ahci: add HFLAG_YES_FBS and apply it to 88SE9128
pata_hpt37x: inherit prereset() method for HPT374
ahci: AHCI mode SATA patch for Intel DH89xxCC DeviceIDs
pata_hpt37x: fold 'if' statement into 'switch'
pata_hpt{37x|3x2n}: use pr_*(DRV_NAME ...) instead of printk(KERN_* ...)
pata_hpt{366|37x}: use pr_warning(...) instead of printk(KERN_WARNING ...)
pata_mpc52xx: inherit from ata_bmdma_port_ops
This reverts commit a121f64399.
Unfortunately, this commit breaks UBIFS backward compatibility and
makes new UBIFS refuse older UBIFS-formatted media:
UBIFS error: validate_sb: min. I/O unit mismatch: 8 in superblock, 64 real
Thus, we have to revert this patch and work on a better solution.
Reported-by: Holger Brunck <holger.brunck@keymile.com>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
On some machines, like Dell Studio XPS 16 (1640), touchpad fails to
respond to the standard query after first reset but may start
responding later, so let's repeat reset sequence several (3) times.
Signed-off-by: Alexandre Peixoto Ferreira <alexandref75@gmail.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
synaptics_set_advanced_gesture_mode() affect capabilities bits we should
perform comparison after calling this function, otherwise they will never
match and we will be forced to perform full reconnect.
Signed-off-by: Alexandre Peixoto Ferreira <alexandref75@gmail.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
On recent 2.6.38-rc kernels, connectathon basic test 6 fails on
NFSv4 mounts of OpenSolaris with something like:
> ./test6: readdir
> ./test6: (/mnt/klimt/matisse.test) didn't read expected 'file.12' dir entry, pass 0
> ./test6: (/mnt/klimt/matisse.test) didn't read expected 'file.82' dir entry, pass 0
> ./test6: (/mnt/klimt/matisse.test) didn't read expected 'file.164' dir entry, pass 0
> ./test6: (/mnt/klimt/matisse.test) Test failed with 3 errors
> basic tests failed
> Tests failed, leaving /mnt/klimt mounted
> [cel@matisse cthon04]$
I narrowed the problem down to nfs4_decode_dirent() reporting that the
decode buffer had overflowed while decoding the entries for those
missing files.
verify_attr_len() assumes both it's pointer arguments reside on the
same page. When these arguments point to locations on two different
pages, verify_attr_len() can report false errors. This can happen now
that a large NFSv4 readdir result can span pages.
We have reasonably good checking in nfs4_decode_dirent() anyway, so
it should be safe to simply remove the extra checking.
At a guess, this was introduced by commit 6650239a, "NFS: Don't use
vm_map_ram() in readdir".
Cc: stable@kernel.org [2.6.37]
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Make the decoding of NFSv4 directory entries slightly more efficient
by:
1. Avoiding unnecessary byte swapping when checking XDR booleans,
and
2. Not bumping "p" when its value will be immediately replaced by
xdr_inline_decode()
This commit makes nfs4_decode_dirent() consistent with similar logic
in the other two decode_dirent() functions.
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
There is no reason to be freeing the delegation cred in the rcu callback,
and doing so is resulting in a lockdep complaint that rpc_credcache_lock
is being called from both softirq and non-softirq contexts.
Reported-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: stable@kernel.org
When filling in the middle of a previous delayed allocation in
xfs_bmap_add_extent_delay_real, set br_startblock of the new delay
extent to the right to nullstartblock instead of 0 before inserting
the extent into the ifork (xfs_iext_insert), rather than setting
br_startblock afterward.
Adding the extent into the ifork with br_startblock=0 can lead to
the extent being copied into the btree by xfs_bmap_extent_to_btree
if we happen to convert from extents format to btree format before
updating br_startblock with the correct value. The unexpected
addition of this delay extent to the btree can cause subsequent
XFS_WANT_CORRUPTED_GOTO filesystem shutdown in several
xfs_bmap_add_extent_delay_real cases where we are converting a delay
extent to real and unexpectedly find an extent already inserted.
For example:
911 case BMAP_LEFT_FILLING:
912 /*
913 * Filling in the first part of a previous delayed allocation.
914 * The left neighbor is not contiguous.
915 */
916 trace_xfs_bmap_pre_update(ip, idx, state, _THIS_IP_);
917 xfs_bmbt_set_startoff(ep, new_endoff);
918 temp = PREV.br_blockcount - new->br_blockcount;
919 xfs_bmbt_set_blockcount(ep, temp);
920 xfs_iext_insert(ip, idx, 1, new, state);
921 ip->i_df.if_lastex = idx;
922 ip->i_d.di_nextents++;
923 if (cur == NULL)
924 rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
925 else {
926 rval = XFS_ILOG_CORE;
927 if ((error = xfs_bmbt_lookup_eq(cur, new->br_startoff,
928 new->br_startblock, new->br_blockcount,
929 &i)))
930 goto done;
931 XFS_WANT_CORRUPTED_GOTO(i == 0, done);
With the bogus extent in the btree we shutdown the filesystem at
931. The conversion from extents to btree format happens when the
number of extents in the inode increases above ip->i_df.if_ext_max.
xfs_bmap_extent_to_btree copies extents from the ifork into the
btree, ignoring all delalloc extents which are denoted by
br_startblock having some value of nullstartblock.
SGI-PV: 1013221
Signed-off-by: Ben Myers <bpm@sgi.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
Commit 368e136 ("xfs: remove duplicate code from dquot reclaim") fails
to unlock the dquot freelist when the number of loop restarts is
exceeded in xfs_qm_dqreclaim_one(). This causes hangs in memory
reclaim.
Rework the loop control logic into an unwind stack that all the
different cases jump into. This means there is only one set of code
that processes the loop exit criteria, and simplifies the unlocking
of all the items from different points in the loop. It also fixes a
double increment of the restart counter from the qi_dqlist_lock
case.
Reported-by: Malcolm Scott <lkml@malc.org.uk>
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Alex Elder <aelder@sgi.com>
Failure to commit a transaction into the CIL is not handled
correctly. This currently can only happen when racing with a
shutdown and requires an explicit shutdown check, so it rare and can
be avoided. Remove the shutdown check and make the CIL commit a void
function to indicate it will always succeed, thereby removing the
incorrectly handled failure case.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Alex Elder <aelder@sgi.com>
The extent size hint can be set to larger than an AG. This means
that the alignment process can push the range to be allocated
outside the bounds of the AG, resulting in assert failures or
corrupted bmbt records. Similarly, if the extsize is larger than the
maximum extent size supported, the alignment process will produce
extents that are too large to fit into the bmbt records, resulting
in a different type of assert/corruption failure.
Fix this by limiting extsize at the time іt is set firstly to be
less than MAXEXTLEN, then to be a maximum of half the size of the
AGs in the filesystem for non-realtime inodes. Realtime inodes do
not allocate out of AGs, so don't have to be restricted by the size
of AGs.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Alex Elder <aelder@sgi.com>
When doing delayed allocation, if the allocation size is for a
maximally sized extent, extent size alignment can push it over this
limit. This results in an assert failure in xfs_bmbt_set_allf() as
the extent length is too large to find in the extent record.
Fix this by ensuring that we allow for space that extent size
alignment requires (up to 2 * (extsize -1) blocks as we have to
handle both head and tail alignment) when limiting the maximum size
of the extent.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Alex Elder <aelder@sgi.com>
Delayed allocation extents can be larger than AGs, so when trying to
convert a large range we may scan every AG inside
xfs_bmap_alloc_nullfb() trying to find an AG with a size larger than
an AG. We should stop when we find the first AG with a maximum
possible allocation size. This causes excessive CPU usage when there
are lots of AGs.
The same problem occurs when doing preallocation of a range larger
than an AG.
Fix the problem by limiting real allocation lengths to the maximum
that an AG can support. This means if we have empty AGs, we'll stop
the search at the first of them. If there are no empty AGs, we'll
still scan them all, but that is a different problem....
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Alex Elder <aelder@sgi.com>
rounddown_power_of_2() returns an undefined result when passed a
value of zero. The specualtive delayed allocation code is doing this
when the inode is zero length. Hence occasionally the preallocation
is much, much larger than is necessary (e.g. 8GB for a 270 _byte_
file). Ensure we don't even pass a zero value to this function so
the result of preallocation is always the desired size.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Alex Elder <aelder@sgi.com>
After test 139, kmemleak shows:
unreferenced object 0xffff880078b405d8 (size 400):
comm "xfs_io", pid 4904, jiffies 4294909383 (age 1186.728s)
hex dump (first 32 bytes):
60 c1 17 79 00 88 ff ff 60 c1 17 79 00 88 ff ff `..y....`..y....
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
backtrace:
[<ffffffff81afb04d>] kmemleak_alloc+0x2d/0x60
[<ffffffff8115c6cf>] kmem_cache_alloc+0x13f/0x2b0
[<ffffffff814aaa97>] kmem_zone_alloc+0x77/0xf0
[<ffffffff814aab2e>] kmem_zone_zalloc+0x1e/0x50
[<ffffffff8147cd6b>] xfs_efi_init+0x4b/0xb0
[<ffffffff814a4ee8>] xfs_trans_get_efi+0x58/0x90
[<ffffffff81455fab>] xfs_bmap_finish+0x8b/0x1d0
[<ffffffff814851b4>] xfs_itruncate_finish+0x2c4/0x5d0
[<ffffffff814a970f>] xfs_setattr+0x8df/0xa70
[<ffffffff814b5c7b>] xfs_vn_setattr+0x1b/0x20
[<ffffffff8117dc00>] notify_change+0x170/0x2e0
[<ffffffff81163bf6>] do_truncate+0x66/0xa0
[<ffffffff81163d0b>] sys_ftruncate+0xdb/0xe0
[<ffffffff8103a002>] system_call_fastpath+0x16/0x1b
[<ffffffffffffffff>] 0xffffffffffffffff
The cause of the leak is that the "remove" parameter of IOP_UNPIN()
is never set when a CIL push is aborted. This means that the EFI
item is never freed if it was in the push being cancelled. The
problem is specific to delayed logging, but has uncovered a couple
of problems with the handling of IOP_UNPIN(remove).
Firstly, we cannot safely call xfs_trans_del_item() from IOP_UNPIN()
in the CIL commit failure path or the iclog write failure path
because for delayed loging we have no transaction context. Hence we
must only call xfs_trans_del_item() if the log item being unpinned
has an active log item descriptor.
Secondly, xfs_trans_uncommit() does not handle log item descriptor
freeing during the traversal of log items on a transaction. It can
reference a freed log item descriptor when unpinning an EFI item.
Hence it needs to use a safe list traversal method to allow items to
be removed from the transaction during IOP_UNPIN().
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Alex Elder <aelder@sgi.com>
Russell King