The /proc/fs/jbd2/<dev>/history was maintained manually; by using
tracepoints, we can get all of the existing functionality of the /proc
file plus extra capabilities thanks to the ftrace infrastructure. We
save memory as a bonus.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The /proc/fs/ext4/<dev>/mb_history was maintained manually, and had a
number of problems: it required a largish amount of memory to be
allocated for each ext4 filesystem, and the s_mb_history_lock
introduced a CPU contention problem.
By ripping out the mb_history code and replacing it with ftrace
tracepoints, and we get more functionality: timestamps, event
filtering, the ability to correlate mballoc history with other ext4
tracepoints, etc.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
If an ioctl-initiated transaction is open, we can't force a commit during
the free space checks in order to free up pinned extents or else we
deadlock. Just ENOSPC instead.
A more satisfying solution that reserves space for the entire user
transaction up front is forthcoming...
Signed-off-by: Sage Weil <sage@newdream.net>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
Fix leak of vfsmount write reference and open_ioctl_trans reference on
ENOMEM. Clean up the error paths while we're at it.
Signed-off-by: Sage Weil <sage@newdream.net>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
This patch fixes a embarrassing bug which was introduced by:
"[PATCH] ar9170: implement frequency calibration for one-stage/openfw"
The phy_data variable initialization has to done outside the for-loop
scope. This is because the for-loop uses u32 phy_data variable more
like a 4-byte field. But in each run only a single byte is calculated.
Therefore phy_data content needs to stay the same for at least 3 more
iterations, before the complete set can be uploaded.
Reported-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Thrustmaster FunAccess WIFI USB works with rt73usb with little
modification of rt73usb.c.
Tested with version 2.3.0 of driver.
Signed-off-by: Michal Szalata <szalat@gmail.com>
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
mac80211_hwsim does not start transmitting Beacon frames when hostapd
is started for the first time and restarting hostapd fixes this. The
issue is caused by the config() handler not being able to start
beacon_timer when beacon interval is not yet known and
bss_info_changed() handler not starting the timer. This can be fixed by
making the bss_info_changed() update the timer.
Signed-off-by: Jouni Malinen <j@w1.fi>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
On SDIO the PIO data register seems to be hardwired to LE. So
the MACCTL bit has no effect on the endianness.
So also use block-I/O for the last word of the packet. block-I/O is always LE.
Signed-off-by: Michael Buesch <mb@bu3sch.de>
Tested-by: Albert Herranz <albert_herranz@yahoo.es>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Consider the following step-by step:
1. A STA authenticates and associates with the AP and exchanges
traffic.
2. The STA reports to the AP that it is going to PS state.
3. Some time later the STA device goes to the stand-by mode (not only
its wi-fi card, but the device itself) and drops the association state
without sending a disassociation frame.
4. The STA device wakes up and begins authentication with an
Auth frame as it hasn't been authenticated/associated previously.
At the step 4 the AP "remembers" the STA and considers it is still in
the PS state, so the AP buffers frames, which it has to send to the STA.
But the STA isn't actually in the PS state and so it neither checks
TIM bits nor reports to the AP that it isn't power saving.
Because of that authentication/[re]association fails.
To fix authentication/[re]association stage of this issue, Auth, Assoc
Resp and Reassoc Resp frames are transmitted disregarding of STA's power
saving state.
N.B. This patch doesn't fix further data frame exchange after
authentication/[re]association. A patch in hostapd is required to fix
that.
Signed-off-by: Igor Perminov <igor.perminov@inbox.ru>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
There are a number of kernel printk's which are printed when an ext4
filesystem is mounted and unmounted. Disable them to economize space
in the system logs. In addition, disabling the mballoc stats by
default saves a number of unneeded atomic operations for every block
allocation or deallocation.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
We've already defined CONFIG_BTRFS_POSIX_ACL in Kconfig, but we're
currently not using it and are testing CONFIG_FS_POSIX_ACL instead.
CONFIG_FS_POSIX_ACL states "Never use this symbol for ifdefs".
Signed-off-by: Chris Ball <cjb@laptop.org>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
We currently set sb->s_flags |= MS_POSIXACL unconditionally, which is
incorrect -- it tells the VFS that it shouldn't set umask because we
will, yet we don't set it ourselves if we aren't using POSIX ACLs, so
the umask ends up ignored.
Signed-off-by: Chris Ball <cjb@laptop.org>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
This patch fixes a problem with handling nested calls to
ext4_journal_start/ext4_journal_stop, when there is no journal present.
Signed-off-by: Curt Wohlgemuth <curtw@google.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
This patch a problem that ext4_dirty_inode() was not calling
ext4_mark_inode_dirty() if the current_handle is not valid, which it
is the case in no journal mode.
It also removes a test for non-matching transaction which can never
happen.
Signed-off-by: Curt Wohlgemuth <curtw@google.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
This is a cleanup of commit 91ac6f4. Since ext4_mark_inode_dirty()
has already called ext4_mark_iloc_dirty(), which in turn calls
ext4_do_update_inode(), it's not necessary to have ext4_write_inode()
call ext4_do_update_inode() in no journal mode. Indeed, it would be
duplicated work.
Reviewed-by: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Frank Mayhar <fmayhar@google.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The i_dir_start_lookup field in nilfs_inode_info objects should be
cleared when the objects are allocated, but the the initialization was
missing in case of reading from disk. This adds the initialization.
Since the variable just gives a start page on directory lookups, the
bug was nonfatal until now.
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
This will fix file system corruption which infrequently happens after
mount. The problem was reported from users with the title "[NILFS
users] Fail to mount NILFS." (Message-ID:
<200908211918.34720.yuri@itinteg.net>), and so forth. I've also
experienced the corruption multiple times on kernel 2.6.30 and 2.6.31.
The problem turned out to be caused due to discordance between
mapping->nrpages of a btree node cache and the actual number of pages
hung on the cache; if the mapping->nrpages becomes zero even as it has
pages, truncate_inode_pages() returns without doing anything. Usually
this is harmless except it may cause page leak, but garbage collection
fairly infrequently sees a stale page remained in the btree node cache
of DAT (i.e. disk address translation file of nilfs), and induces the
corruption.
I identified a missing initialization in btree node caches was the
root cause. This corrects the bug.
I've tested this for kernel 2.6.30 and 2.6.31.
Reported-by: Yuri Chislov <yuri@itinteg.net>
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
Cc: stable <stable@kernel.org>
Convert the r520 asic support to new init path, change are smaller than
previous one as most of the architecture is now in place and more code
sharing can happen btw various asics.
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@linux.ie>
Convert the rv515 asic support to new init path also add an explanation
in radeon.h about the new init path. There is also few cleanups
associated with this change (others asic calling rv515 helper
functions).
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@linux.ie>
Compiling the radeon DRM driver with !CONFIG_DEBUG_FS
throws the following warnings:
drivers/gpu/drm/radeon/radeon_ttm.c: In function 'radeon_ttm_debugfs_init':
drivers/gpu/drm/radeon/radeon_ttm.c:714: warning: unused variable 'i'
drivers/gpu/drm/radeon/radeon_ttm.c: At top level:
drivers/gpu/drm/radeon/radeon_ttm.c:692: warning: 'radeon_mem_types_list' defined but not used
drivers/gpu/drm/radeon/radeon_ttm.c:693: warning: 'radeon_mem_types_names' defined but not used
Fix: move these variables inside the #if defined(CONFIG_DEBUG_FS)
block in radeon_ttm_debugsfs_init(), which is the only place using them.
Signed-off-by: Mikael Pettersson <mikpe@it.uu.se>
Signed-off-by: Dave Airlie <airlied@linux.ie>
Compiling DRM throws the following warning if MAGIC_SYSRQ is disabled:
drivers/gpu/drm/drm_fb_helper.c:101: warning: 'sysrq_drm_fb_helper_restore_op' defined but not used
Fix: place sysrq_drm_fb_helper_restore_op and associated
definitions inside #ifdef CONFIG_MAGIC_SYSRQ.
Signed-off-by: Mikael Pettersson <mikpe@it.uu.se>
Signed-off-by: Dave Airlie <airlied@linux.ie>
Warn and dump stack when percpu allocation fails. percpu allocator is
still young and unchecked NULL percpu pointer usage can result in
random memory corruption when combined with the pointer shifting in
access macros. Allocation failures should be rare and the warning
message will be disabled after certain times.
Signed-off-by: Tejun Heo <tj@kernel.org>
The parameters to pcpu_setup_first_chunk() come from different sources
depending on architecture and can be quite complex. The function runs
various sanity checks on the parameters and triggers BUG() if
something isn't right. However, this is very early during the boot
and not reporting exactly what the problem is makes debugging even
harder.
Add PCPU_SETUP_BUG() macro which prints out enough information about
the parameters. As the macro still puts separate BUG() for each
check, it won't lose any information even on the situations where only
the program counter can be retrieved.
While at it, also bump pcpu_dump_alloc_info() message to KERN_INFO so
that it's visible on the console if boot fails to complete.
Signed-off-by: Tejun Heo <tj@kernel.org>
Embedding first chunk allocator maintains the distances between units
in the vmalloc area and thus needs vmalloc space to be larger than the
maximum distances between units; otherwise, it wouldn't be able to
create any dynamic chunks. This patch makes the embedding first chunk
allocator check vmalloc space size and if the maximum distance between
units is larger than 75% of it, print warning and, if page mapping
allocator is available, fail initialization so that the system falls
back onto it.
This should work around percpu allocation failure problems on certain
sparc64 configurations where distances between NUMA nodes are larger
than the vmalloc area and makes percpu allocator more robust for
future configurations.
Signed-off-by: Tejun Heo <tj@kernel.org>
Implement page mapping percpu first chunk allocator as a fallback to
the embedding allocator. The next patch will make the embedding
allocator check distances between units to determine whether it fits
within the vmalloc area so that this fallback can be used on such
cases.
sparc64 currently has relatively small vmalloc area which makes it
impossible to create any dynamic chunks on certain configurations
leading to percpu allocation failures. This and the next patch should
allow those configurations to keep working until proper solution is
found.
While at it, mark pcpu_cpu_distance() with __init.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: David S. Miller <davem@davemloft.net>
pcpu_build_alloc_info() may be called multiple times when percpu is
falling back to different first chunk allocator. Make it clear static
buffers so that they don't contain values from previous runs.
Signed-off-by: Tejun Heo <tj@kernel.org>
pcpu_setup_first_chunk() incorrectly used NR_CPUS as the impossible
unit number while unit number can equal and go over NR_CPUS with
sparse unit map. This triggers BUG_ON() spuriously on machines which
have non-power-of-two number of cpus. Use UINT_MAX instead.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-and-tested-by: Tony Vroon <tony@linx.net>
Since 2.6.29 the PCI PM core have been restoring the standard
configuration registers of PCI devices in the early phase of
resume. In particular, PCI devices without drivers have been handled
this way since commit 355a72d75b
(PCI: Rework default handling of suspend and resume). Unfortunately,
this leads to post-resume problems with CardBus devices which cannot
be accessed in the early phase of resume, because the sockets they
are on have not been woken up yet at that point.
To solve this problem, move the yenta socket resume to the early
phase of resume and, analogously, move the suspend of it to the late
phase of suspend. Additionally, remove some unnecessary PCI code
from the yenta socket's resume routine.
Fixes http://bugzilla.kernel.org/show_bug.cgi?id=13092, which is a
post-2.6.28 regression.
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Reported-by: Florian <fs-kernelbugzilla@spline.de>
Cc: stable@kernel.org
pcmcia_socket_dev_suspend() doesn't use its second argument, so it
may be dropped safely.
This change is necessary for the subsequent yenta suspend/resume fix.
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Cc: stable@kernel.org
Without this, the hard-blocked state will be reported incorrectly if
the hardware switch is changed while the laptop is suspended.
Signed-off-by: Alan Jenkins <alan-jenkins@tuffmail.co.uk>
Tested-by: Norbert Preining <preining@logic.at>
Acked-by: Mattia Dongili <malattia@linux.it>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
"I recently (on a flight) I found out that when I boot with the hard-switch
activated, so turning off all wireless activity on my laptop, the state
is not correctly announced in /dev/rfkill (reading it with rfkill command,
or my own gnome applet)...
After turning off and on again the hard-switch the events were right."
We can fix this by querying the firmware at load time and calling
rfkill_set_hw_state().
Signed-off-by: Alan Jenkins <alan-jenkins@tuffmail.co.uk>
Tested-by: Norbert Preining <preining@logic.at>
Acked-by: Johannes Berg <johannes@sipsolutions.net>
Acked-by: Mattia Dongili <malattia@linux.it>
CC: stable@kernel.org
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The move away from having drivers assign wireless handlers,
in favour of making cfg80211 assign them, broke the sysfs
registration (the wireless/ dir went missing) because the
handlers are now assigned only after registration, which is
too late.
Fix this by special-casing cfg80211-based devices, all
of which are required to have an ieee80211_ptr, in the
sysfs code, and also using get_wireless_stats() to have
the same values reported as in procfs.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Reported-by: Hugh Dickins <hugh.dickins@tiscali.co.uk>
Tested-by: Hugh Dickins <hugh.dickins@tiscali.co.uk>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The wireless extensions have a copy_from_user to a local stack
array "essid", but both me and gcc have failed to find where
the bounds for this copy are located in the code.
This patch adds some basic sanity checks for the copy length
to make sure that we don't overflow the stack buffer.
Signed-off-by: Arjan van de Ven <arjan@linux.intel.com>
Cc: linux-wireless@vger.kernel.org
Signed-off-by: John W. Linville <linville@tuxdriver.com>
It's useful to know the MAC address when being
disassociated; fix a typo (missing colon) and
move some messages so we get them only when they
are actually taking effect.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Multiple problems were reported due to interaction
between wpa_supplicant and the wext compat code in
cfg80211, which appear to be due to it not getting
any bss pointer here when wpa_supplicant sets all
parameters -- do that now. We should still get the
bss after doing an extra scan, but that appears to
increase the time we need for connecting enough to
sometimes cause timeouts.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Tested-by: Hin-Tak Leung <hintak.leung@gmail.com>,
Signed-off-by: John W. Linville <linville@tuxdriver.com>
When hardware or uCode problem occurs driver captures significant
information from device to enable debugging. The format of this information
is different between 3945 and 4965 and later devices, yet currently the
3945 uses the 4965 and later format. Fix this by adding a new library call
that is initialized to the correct formatting routine based on device.
This moves the iwlagn event and error log handling back to iwl-agn.c to
make it part of iwlagn module.
Also remove the 3945 sysfs file that triggers dump of event log - there is
already a debugfs file that can do it for all drivers.
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Also free the array of command pointers and meta data of each
command buffer when command queue is freed.
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
We keep track of where to write into a buffer by keeping a count of how
much has been written so far. When writing to the buffer we thus take the
buffer pointer and adding the count of what has been written so far.
Keeping track of what has been written so far is done by incrementing
this number every time something is written to the buffer with how much has
been written at that time.
Currently this number is incremented incorrectly when using the
"hex_dump_to_buffer" call to add data to the buffer. Fix this by only
adding what has been added to the buffer in that call instead of what has
been added since beginning of buffer.
Issue was discovered and discussed during testing of
https://bugzilla.redhat.com/show_bug.cgi?id=464598 .
When a user views any of these files they will see something like:
[ 179.355202] ------------[ cut here ]------------
[ 179.355209] WARNING: at ../lib/vsprintf.c:989 vsnprintf+0x5ec/0x5f0()
[ 179.355212] Hardware name: VGN-Z540N
[ 179.355213] Modules linked in: i915 drm i2c_algo_bit i2c_core ipv6 acpi_cpufreq cpufreq_userspace cpufreq_powersave cpufreq_ondemand cpufreq_conservative cpufreq_stats freq_table container sbs sbshc arc4 ecb iwlagn iwlcore joydev led_class mac80211 af_packet pcmcia psmouse sony_laptop cfg80211 iTCO_wdt iTCO_vendor_support pcspkr serio_raw rfkill intel_agp video output tpm_infineon tpm tpm_bios button battery yenta_socket rsrc_nonstatic pcmcia_core processor ac evdev ext3 jbd mbcache sr_mod sg cdrom sd_mod ahci libata scsi_mod ehci_hcd uhci_hcd usbcore thermal fan thermal_sys
[ 179.355262] Pid: 5449, comm: cat Not tainted 2.6.31-wl-54419-ge881071 #62
[ 179.355264] Call Trace:
[ 179.355267] [<ffffffff811ad14c>] ? vsnprintf+0x5ec/0x5f0
[ 179.355271] [<ffffffff81041348>] warn_slowpath_common+0x78/0xd0
[ 179.355275] [<ffffffff810413af>] warn_slowpath_null+0xf/0x20
[ 179.355277] [<ffffffff811ad14c>] vsnprintf+0x5ec/0x5f0
[ 179.355280] [<ffffffff811ad23d>] ? scnprintf+0x5d/0x80
[ 179.355283] [<ffffffff811ad23d>] scnprintf+0x5d/0x80
[ 179.355286] [<ffffffff811aed29>] ? hex_dump_to_buffer+0x189/0x340
[ 179.355290] [<ffffffff810e91d7>] ? __kmalloc+0x207/0x260
[ 179.355303] [<ffffffffa02a02f8>] iwl_dbgfs_nvm_read+0xe8/0x220 [iwlcore]
[ 179.355306] [<ffffffff811a9b62>] ? __up_read+0x92/0xb0
[ 179.355310] [<ffffffff810f0988>] vfs_read+0xc8/0x1a0
[ 179.355313] [<ffffffff810f0b50>] sys_read+0x50/0x90
[ 179.355316] [<ffffffff8100bd6b>] system_call_fastpath+0x16/0x1b
[ 179.355319] ---[ end trace 2383d0d5e0752ca0 ]---
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
When wpa_supplicant is used to connect to open networks,
it causes the wdev->wext.keys to point to key memory, but
that key memory is all empty. Only use privacy when there
is a default key to be used.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Tested-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Tested-by: Kalle Valo <kalle.valo@iki.fi>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Currently, cfg80211's SIOCGIWAP implementation returns
the BSSID that the user set, even if the connection has
since been dropped due to other changes. It only should
return the current BSSID when actually connected.
Also do a small code cleanup.
Reported-by: Thomas H. Guenther <thomas.h.guenther@intel.com>
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Tested-by: Thomas H. Guenther <thomas.h.guenther@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
At the start of a transaction we do a btrfs_reserve_metadata_space() and
specify how many items we plan on modifying. Then once we've done our
modifications and such, just call btrfs_unreserve_metadata_space() for
the same number of items we reserved.
For keeping track of metadata needed for data I've had to add an extent_io op
for when we merge extents. This lets us track space properly when we are doing
sequential writes, so we don't end up reserving way more metadata space than
what we need.
The only place where the metadata space accounting is not done is in the
relocation code. This is because Yan is going to be reworking that code in the
near future, so running btrfs-vol -b could still possibly result in a ENOSPC
related panic. This patch also turns off the metadata_ratio stuff in order to
allow users to more efficiently use their disk space.
This patch makes it so we track how much metadata we need for an inode's
delayed allocation extents by tracking how many extents are currently
waiting for allocation. It introduces two new callbacks for the
extent_io tree's, merge_extent_hook and split_extent_hook. These help
us keep track of when we merge delalloc extents together and split them
up. Reservations are handled prior to any actually dirty'ing occurs,
and then we unreserve after we dirty.
btrfs_unreserve_metadata_for_delalloc() will make the appropriate
unreservations as needed based on the number of reservations we
currently have and the number of extents we currently have. Doing the
reservation outside of doing any of the actual dirty'ing lets us do
things like filemap_flush() the inode to try and force delalloc to
happen, or as a last resort actually start allocation on all delalloc
inodes in the fs. This has survived dbench, fs_mark and an fsx torture
test.
Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
Move the check to make sure the original and donor inodes are
different earlier, to avoid a potential deadlock by trying to lock the
same inode twice.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The sys_socketcall() function has a very clever system for the copy
size of its arguments. Unfortunately, gcc cannot deal with this in
terms of proving that the copy_from_user() is then always in bounds.
This is the last (well 9th of this series, but last in the kernel) such
case around.
With this patch, we can turn on code to make having the boundary provably
right for the whole kernel, and detect introduction of new security
accidents of this type early on.
Signed-off-by: Arjan van de Ven <arjan@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There is a potential double-kfree in net/bridge/br_if.c. If br_fdb_insert
fails, then the kobject is put back (which calls kfree due to the kobject
release), and then kfree is called again on the net_bridge_port. This
patch fixes the crash.
Thanks to Stephen Hemminger for the one-line fix.
Signed-off-by: Jeff Hansen <x@jeffhansen.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
For async direct IO that covers holes or fallocate, the end_io
callback function now queued the convertion work on workqueue but
don't flush the work rightaway as it might take too long to afford.
But when fsync is called after all the data is completed, user expects
the metadata also being updated before fsync returns.
Thus we need to flush the conversion work when fsync() is called.
This patch keep track of a listed of completed async direct io that
has a work queued on workqueue. When fsync() is called, it will go
through the list and do the conversion.
Signed-off-by: Mingming Cao <cmm@us.ibm.com>
Currently the DIO VFS code passes create = 0 when writing to the
middle of file. It does this to avoid block allocation for holes, so
as not to expose stale data out when there is a parallel buffered read
(which does not hold the i_mutex lock). Direct I/O writes into holes
falls back to buffered IO for this reason.
Since preallocated extents are treated as holes when doing a
get_block() look up (buffer is not mapped), direct IO over fallocate
also falls back to buffered IO. Thus ext4 actually silently falls
back to buffered IO in above two cases, which is undesirable.
To fix this, this patch creates unitialized extents when a direct I/O
write into holes in sparse files, and registering an end_io callback which
converts the uninitialized extent to an initialized extent after the
I/O is completed.
Singed-Off-By: Mingming Cao <cmm@us.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
When writing into an unitialized extent via direct I/O, and the direct
I/O doesn't exactly cover the unitialized extent, split the extent
into uninitialized and initialized extents before submitting the I/O.
This avoids needing to deal with an ENOSPC error in the end_io
callback that gets used for direct I/O.
When the IO is complete, the written extent will be marked as initialized.
Singed-Off-By: Mingming Cao <cmm@us.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
ext4_da_reserve_space() can reserve quota blocks multiple times if
ext4_claim_free_blocks() fail and we retry the allocation. We should
release the quota reservation before restarting.
Bug found by Jan Kara.
Signed-off-by: Mingming Cao <cmm@us.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Theodore Ts'o
Sage Weil
Christian Lamparter
Michal Szalata
Jouni Malinen
Michael Buesch
Igor Perminov
Chris Ball
Julia Lawall
Curt Wohlgemuth
Frank Mayhar
Ryusuke Konishi
Jerome Glisse
Mikael Pettersson
Tejun Heo
Rafael J. Wysocki
David S. Miller
Alan Jenkins
Johannes Berg
Arjan van de Ven
Reinette Chatre
Josef Bacik
Jeff Hansen
Mingming Cao