On LPAE, each level 1 (pgd) page table entry maps 1GiB, and the level 2
(pmd) entries map 2MiB.
When the identity mapping is created on LPAE, the pgd pointers are copied
from the swapper_pg_dir. If we find that we need to modify the contents
of a pmd, we allocate a new empty pmd table and insert it into the
appropriate 1GB slot, before then filling it with the identity mapping.
However, if the 1GB slot covers the kernel lowmem mappings, we obliterate
those mappings.
When replacing a PMD, first copy the old PMD contents to the new PMD, so
that we preserve the existing mappings, particularly the mappings of the
kernel itself.
[rewrote commit message and added code comment -- rmk]
Fixes: ae2de101739c ("ARM: LPAE: Add identity mapping support for the 3-level page table format") Signed-off-by: Konstantin Khlebnikov <k.khlebnikov@samsung.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Th AF_ALG socket was missing a security label (e.g. SELinux)
which means that socket was in "unlabeled" state.
This was recently demonstrated in the cryptsetup package
(cryptsetup v1.6.5 and later.)
See https://bugzilla.redhat.com/show_bug.cgi?id=1115120
This patch clones the sock's label from the parent sock
and resolves the issue (similar to AF_BLUETOOTH protocol family).
Signed-off-by: Milan Broz <gmazyland@gmail.com> Acked-by: Paul Moore <paul@paul-moore.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commits 9ec36ca (of/irq: do irq resolution in platform_get_irq)
and ad69674 (of/irq: do irq resolution in platform_get_irq_byname)
change the semantics of platform_get_irq and platform_get_irq_byname
to always rely on devicetree information if devicetree is enabled
and if a devicetree node is attached to the device. The functions
now return an error if the devicetree data does not include interrupt
information, even if the information is available as platform resource
data.
This causes mfd client drivers to fail if the interrupt number is
passed via platform resources. Therefore, if of_irq_get fails, try
platform_get_resource as method of last resort. This restores the
original functionality for drivers depending on platform resources
to get irq information.
Cc: Russell King <linux@arm.linux.org.uk> Cc: Tony Lindgren <tony@atomide.com> Cc: Grant Likely <grant.likely@linaro.org> Cc: Grygorii Strashko <grygorii.strashko@ti.com> Signed-off-by: Guenter Roeck <linux@roeck-us.net> Acked-by: Rob Herring <robh@kernel.org>
[ Guenter Roeck: backported to 3.15 ] Signed-off-by: Guenter Roeck <linux@roeck-us.net>
The PE/COFF headers currently describe only the initialised-data
portions of the image, and result in no space being allocated for the
uninitialised-data portions. Consequently, the EFI boot stub will end
up overwriting unexpected areas of memory, with unpredictable results.
Fix by including a .bss section in the PE/COFF headers (functionally
equivalent to the init_size field in the bzImage header).
Signed-off-by: Michael Brown <mbrown@fensystems.co.uk> Cc: Thomas Bächler <thomas@archlinux.org> Cc: Josh Boyer <jwboyer@fedoraproject.org> Signed-off-by: Matt Fleming <matt.fleming@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Michel Dänzer and a couple of other people reported inexplicable random
oopses in the scheduler, and the cause turns out to be gcc mis-compiling
the load_balance() function when debugging is enabled. The gcc bug
apparently goes back to gcc-4.5, but slight optimization changes means
that it now showed up as a problem in 4.9.0 and 4.9.1.
The instruction scheduling problem causes gcc to schedule a spill
operation to before the stack frame has been created, which in turn can
corrupt the spilled value if an interrupt comes in. There may be other
effects of this bug too, but that's the code generation problem seen in
Michel's case.
This is fixed in current gcc HEAD, but the workaround as suggested by
Markus Trippelsdorf is pretty simple: use -fno-var-tracking-assignments
when compiling the kernel, which disables the gcc code that causes the
problem. This can result in slightly worse debug information for
variable accesses, but that is infinitely preferable to actual code
generation problems.
Doing this unconditionally (not just for CONFIG_DEBUG_INFO) also allows
non-debug builds to verify that the debug build would be identical: we
can do
export GCC_COMPARE_DEBUG=1
to make gcc internally verify that the result of the build is
independent of the "-g" flag (it will make the compiler build everything
twice, toggling the debug flag, and compare the results).
Without the "-fno-var-tracking-assignments" option, the build would fail
(even with 4.8.3 that didn't show the actual stack frame bug) with a gcc
compare failure.
Commit 4a705fef9862 ("hugetlb: fix copy_hugetlb_page_range() to handle
migration/hwpoisoned entry") changed the order of
huge_ptep_set_wrprotect() and huge_ptep_get(), which leads to breakage
in some workloads like hugepage-backed heap allocation via libhugetlbfs.
This patch fixes it.
Fixes 4a705fef9862 ("hugetlb: fix copy_hugetlb_page_range() to handle
migration/hwpoisoned entry"), so is applicable to -stable kernels which
include it.
We must mask out the overflow bit as well, otherwise
the wptr will never match the rptr again and the interrupt
handler will loop forever.
Signed-off-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Michel Dänzer <michel.daenzer@amd.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 554086d ("x86_32, entry: Do syscall exit work on badsys
(CVE-2014-4508)") introduced a regression in the x86_32 syscall entry
code, resulting in syscall() not returning proper errors for undefined
syscalls on CPUs supporting the sysenter feature.
The following code:
> int result = syscall(666);
> printf("result=%d errno=%d error=%s\n", result, errno, strerror(errno));
results in:
> result=666 errno=0 error=Success
Obviously, the syscall return value is the called syscall number, but it
should have been an ENOSYS error. When run under ptrace it behaves
correctly, which makes it hard to debug in the wild:
> result=-1 errno=38 error=Function not implemented
The %eax register is the return value register. For debugging via ptrace
the syscall entry code stores the complete register context on the
stack. The badsys handlers only store the ENOSYS error code in the
ptrace register set and do not set %eax like a regular syscall handler
would. The old resume_userspace call chain contains code that clobbers
%eax and it restores %eax from the ptrace registers afterwards. The same
goes for the ptrace-enabled call chain. When ptrace is not used, the
syscall return value is the passed-in syscall number from the untouched
%eax register.
Use %eax as the return value register in syscall_badsys and
sysenter_badsys, like a real syscall handler does, and have the caller
push the value onto the stack for ptrace access.
In this case mountpoint_last() gets an extra refcount on path->mnt
Signed-off-by: Vasily Averin <vvs@openvz.org> Acked-by: Ian Kent <raven@themaw.net> Acked-by: Jeff Layton <jlayton@primarydata.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Do not split the PARPORT-related symbols with the new kconfig
symbol ARCH_MIGHT_HAVE_PC_PARPORT. The split was causing incorrect
display of these symbols -- they were not being displayed together
as they should be.
Temperature limit clamps are applied after converting the temperature
from milli-degrees C to degrees C, so either the clamp limit needs
to be specified in degrees C, not milli-degrees C, or clamping must
happen before converting to degrees C. Use the latter method to avoid
overflows.
vrm is an u8, so the written value needs to be limited to [0, 255].
Cc: Axel Lin <axel.lin@ingics.com> Signed-off-by: Guenter Roeck <linux@roeck-us.net> Reviewed-by: Jean Delvare <jdelvare@suse.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The sa_restorer field in struct sigaction is obsolete and no longer in
the parisc implementation. However, the core code assumes the field is
present if SA_RESTORER is defined. So, the define needs to be removed.
Signed-off-by: John David Anglin <dave.anglin@bell.net> Signed-off-by: Helge Deller <deller@gmx.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 079148b919d0 ("coredump: factor out the setting of PF_DUMPCORE")
cleaned up the setting of PF_DUMPCORE by removing it from all the
linux_binfmt->core_dump() and moving it to zap_threads().But this ended
up clearing all the previously set flags. This causes issues during
core generation when tsk->flags is checked again (eg. for PF_USED_MATH
to dump floating point registers). Fix this.
Signed-off-by: Silesh C V <svellattu@mvista.com> Acked-by: Oleg Nesterov <oleg@redhat.com> Cc: Mandeep Singh Baines <msb@chromium.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We attempt to remove noise from coordinates reported by devices in
input_handle_abs_event(), unfortunately, unless we were dropping the
event altogether, we were ignoring the adjusted value and were passing
on the original value instead.
Reviewed-by: Andrew de los Reyes <adlr@chromium.org> Reviewed-by: Benson Leung <bleung@chromium.org> Reviewed-by: David Herrmann <dh.herrmann@gmail.com> Reviewed-by: Henrik Rydberg <rydberg@euromail.se> Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The patch 3e374919b314f20e2a04f641ebc1093d758f66a4 is supposed to fix the
problem where kmem_cache_create incorrectly reports duplicate cache name
and fails. The problem is described in the header of that patch.
However, the patch doesn't really fix the problem because of these
reasons:
* the logic to test for debugging is reversed. It was intended to perform
the check only if slub debugging is enabled (which implies that caches
with the same parameters are not merged). Therefore, there should be
#if !defined(CONFIG_SLUB) || defined(CONFIG_SLUB_DEBUG_ON)
The current code has the condition reversed and performs the test if
debugging is disabled.
* slub debugging may be enabled or disabled based on kernel command line,
CONFIG_SLUB_DEBUG_ON is just the default settings. Therefore the test
based on definition of CONFIG_SLUB_DEBUG_ON is unreliable.
This patch fixes the problem by removing the test
"!defined(CONFIG_SLUB_DEBUG_ON)". Therefore, duplicate names are never
checked if the SLUB allocator is used.
has wraparound problems when the system has been up more
than 1 hour 11 minutes and 34 seconds. It converts jiffies
to nanoseconds using:
(u64)jiffies_to_usecs(jiffy) * 1000ULL
but since jiffies_to_usecs() only returns a 32-bit value, it
truncates at 2^32 microseconds. An additional problem on 32-bit
systems is that the argument is "unsigned long", so fixing the
return value only helps until 2^32 jiffies (49.7 days on a HZ=1000
system).
Avoid these problems by using jiffies_64 as our basis, and
not converting to nanoseconds (we do convert to clock_t because
user facing API must not be dependent on internal kernel
HZ values).
While a queue is being destroyed, all the blkgs are destroyed and its
->root_blkg pointer is set to NULL. If someone else starts to drain
while the queue is in this state, the following oops happens.
776687bce42b ("block, blk-mq: draining can't be skipped even if
bypass_depth was non-zero") made it easier to trigger this bug by
making blk_queue_bypass_start() drain even when it loses the first
bypass test to blk_cleanup_queue(); however, the bug has always been
there even before the commit as blk_queue_bypass_start() could race
against queue destruction, win the initial bypass test but perform the
actual draining after blk_cleanup_queue() already destroyed all blkgs.
Fix it by skippping calling into policy draining if all the blkgs are
already gone.
The PSW mask check of the PTRACE_POKEUSR_AREA command is incorrect.
The PSW_MASK_USER define contains the PSW_MASK_ASC bits, the ptrace
interface accepts all combinations for the address-space-control
bits. To protect the kernel space the PSW mask check in ptrace needs
to reject the address-space-control bit combination for home space.
Fixes CVE-2014-3534
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
1871ee134b73 ("libata: support the ata host which implements a queue
depth less than 32") directly used ata_port->scsi_host->can_queue from
ata_qc_new() to determine the number of tags supported by the host;
unfortunately, SAS controllers doing SATA don't initialize ->scsi_host
leading to the following oops.
Fix it by introducing ata_host->n_tags which is initialized to
ATA_MAX_QUEUE - 1 in ata_host_init() for SAS controllers and set to
scsi_host_template->can_queue in ata_host_register() for !SAS ones.
As SAS hosts are never registered, this will give them the same
ATA_MAX_QUEUE - 1 as before. Note that we can't use
scsi_host->can_queue directly for SAS hosts anyway as they can go
higher than the libata maximum.
Signed-off-by: Tejun Heo <tj@kernel.org> Reported-by: Mike Qiu <qiudayu@linux.vnet.ibm.com> Reported-by: Jesse Brandeburg <jesse.brandeburg@gmail.com> Reported-by: Peter Hurley <peter@hurleysoftware.com> Reported-by: Peter Zijlstra <peterz@infradead.org> Tested-by: Alexey Kardashevskiy <aik@ozlabs.ru> Fixes: 1871ee134b73 ("libata: support the ata host which implements a queue depth less than 32") Cc: Kevin Hao <haokexin@gmail.com> Cc: Dan Williams <dan.j.williams@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The sata on fsl mpc8315e is broken after the commit 8a4aeec8d2d6
("libata/ahci: accommodate tag ordered controllers"). The reason is
that the ata controller on this SoC only implement a queue depth of
16. When issuing the commands in tag order, all the commands in tag
16 ~ 31 are mapped to tag 0 unconditionally and then causes the sata
malfunction. It makes no senses to use a 32 queue in software while
the hardware has less queue depth. So consider the queue depth
implemented by the hardware when requesting a command tag.
Fixes: 8a4aeec8d2d6 ("libata/ahci: accommodate tag ordered controllers") Signed-off-by: Kevin Hao <haokexin@gmail.com> Acked-by: Dan Williams <dan.j.williams@intel.com> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There is no inherent reason why the last put of a tag structure must be
the one for the Scsi_Host, as device model objects can be held for
arbitrary periods. Merge blk_free_tags and __blk_free_tags into a single
funtion that just release a references and get rid of the BUG() when the
host reference wasn't the last.
The big ACL switched nfs to use generic_listxattr, which calls all existing
->list handlers. Add a custom .listxattr implementation that only lists
the ACLs if they actually are present on the given inode.
Signed-off-by: Christoph Hellwig <hch@lst.de> Reported-by: Philippe Troin <phil@fifi.org> Tested-by: Philippe Troin <phil@fifi.org> Fixes: 013cdf1088d7 (nfs: use generic posix ACL infrastructure ...) Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Only supported modulation for DVB-S is QPSK. Modulation parameter
contains invalid value for DVB-S on some cases, which leads driver
refusing tuning attempt. Due to that, hard code modulation to QPSK
in case of DVB-S.
When the audio encoding is changed the driver calls hdpvr_set_audio
with the current opt->audio_input value. However, that should have
been opt->audio_input + 1. So changing the audio encoding inadvertently
changes the input as well. This bug has always been there.
The second bug was introduced in kernel 3.10 and that broke the
default_audio_input module option handling: the audio encoding was
never switched to AC3 if default_audio_input was set to 2 (SPDIF input).
In addition, since starting with 3.10 the audio encoding is always set
at the start the first bug now always happens when the driver is loaded.
In the past this bug would only surface if the user would change the
audio encoding after the driver was loaded.
Also fixes a small trivial typo (bufffer -> buffer).
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Reported-by: Scott Doty <scott@corp.sonic.net> Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The OMAP4 camera support depends on I2C and VIDEO_V4L2, both
of which can be loadable modules. This causes build failures
if we want the camera driver to be built-in.
This can be solved by turning the option into "tristate",
which unfortunately causes another problem, because the
driver incorrectly calls a platform-internal interface
for omap4_ctrl_pad_readl/omap4_ctrl_pad_writel.
Instead, this patch just forbids the invalid configurations
and ensures that the driver can only be built if all its
dependencies are built-in.
shrink_inactive_list() used to wait 0.1s to avoid congestion when all
the pages that were isolated from the inactive list were dirty but not
under active writeback. That makes no real sense, and apparently causes
major interactivity issues under some loads since 3.11.
The ostensible reason for it was to wait for kswapd to start writing
pages, but that seems questionable as well, since the congestion wait
code seems to trigger for kswapd itself as well. Also, the logic behind
delaying anything when we haven't actually started writeback is not
clear - it only delays actually starting that writeback.
We'll still trigger the congestion waiting if
(a) the process is kswapd, and we hit pages flagged for immediate
reclaim
(b) the process is not kswapd, and the zone backing dev writeback is
actually congested.
This probably needs to be revisited, but as it is this fixes a reported
regression.
Reported-by: Felipe Contreras <felipe.contreras@gmail.com> Pinpointed-by: Hillf Danton <dhillf@gmail.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Mel Gorman <mgorman@suse.de> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[mhocko@suse.cz: backport to 3.12 stable tree] Fixes: e2be15f6c3ee ('mm: vmscan: stall page reclaim and writeback pages based on dirty/writepage pages encountered') Reported-by: Felipe Contreras <felipe.contreras@gmail.com> Pinpointed-by: Hillf Danton <dhillf@gmail.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Mel Gorman <mgorman@suse.de> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Michal Hocko <mhocko@suse.cz> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Firmware folks seem say that this flag can make trouble.
Drop it. The advantage of CTS to self is that it slightly
reduces the cost of the protection, but make the protection
less reliable.
As of commit f8567a3845ac05bb28f3c1b478ef752762bd39ef it is now possible to
have put_reqs_available() called from irq context. While put_reqs_available()
is per cpu, it did not protect itself from interrupts on the same CPU. This
lead to aio_complete() corrupting the available io requests count when run
under a heavy O_DIRECT workloads as reported by Robert Elliott. Fix this by
disabling irq updates around the per cpu batch updates of reqs_available.
Many thanks to Robert and folks for testing and tracking this down.
Reported-by: Robert Elliot <Elliott@hp.com> Tested-by: Robert Elliot <Elliott@hp.com> Signed-off-by: Benjamin LaHaise <bcrl@kvack.org> Cc: Jens Axboe <axboe@kernel.dk>, Christoph Hellwig <hch@infradead.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
nr_switches is unsigned long and do_div truncates it to 32 bits, which
means it can test non-zero on e.g. x86-64 and be truncated to zero for
division.
Fix the problem by using div64_ul() instead.
As a side effect calculations of avg_atom for big nr_switches are now correct.
The optimistic spin code assumes regular stores and cmpxchg() play nice;
this is found to not be true for at least: parisc, sparc32, tile32,
metag-lock1, arc-!llsc and hexagon.
There is further wreckage, but this in particular seemed easy to
trigger, so blacklist this.
Opt in for known good archs.
Signed-off-by: Peter Zijlstra <peterz@infradead.org> Reported-by: Mikulas Patocka <mpatocka@redhat.com> Cc: David Miller <davem@davemloft.net> Cc: Chris Metcalf <cmetcalf@tilera.com> Cc: James Bottomley <James.Bottomley@hansenpartnership.com> Cc: Vineet Gupta <vgupta@synopsys.com> Cc: Jason Low <jason.low2@hp.com> Cc: Waiman Long <waiman.long@hp.com> Cc: "James E.J. Bottomley" <jejb@parisc-linux.org> Cc: Paul McKenney <paulmck@linux.vnet.ibm.com> Cc: John David Anglin <dave.anglin@bell.net> Cc: James Hogan <james.hogan@imgtec.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Davidlohr Bueso <davidlohr@hp.com> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Catalin Marinas <catalin.marinas@arm.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Will Deacon <will.deacon@arm.com> Cc: linux-arm-kernel@lists.infradead.org Cc: linux-kernel@vger.kernel.org Cc: linuxppc-dev@lists.ozlabs.org Cc: sparclinux@vger.kernel.org Link: http://lkml.kernel.org/r/20140606175316.GV13930@laptop.programming.kicks-ass.net Signed-off-by: Ingo Molnar <mingo@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The commit [247bc037: PM / Sleep: Mitigate race between the freezer
and request_firmware()] introduced the finer state control, but it
also leads to a new bug; for example, a bug report regarding the
firmware loading of intel BT device at suspend/resume:
https://bugzilla.novell.com/show_bug.cgi?id=873790
The root cause seems to be a small window between the process resume
and the clear of usermodehelper lock. The request_firmware() function
checks the UMH lock and gives up when it's in UMH_DISABLE state. This
is for avoiding the invalid f/w loading during suspend/resume phase.
The problem is, however, that usermodehelper_enable() is called at the
end of thaw_processes(). Thus, a thawed process in between can kick
off the f/w loader code path (in this case, via btusb_setup_intel())
even before the call of usermodehelper_enable(). Then
usermodehelper_read_trylock() returns an error and request_firmware()
spews WARN_ON() in the end.
This oneliner patch fixes the issue just by setting to UMH_FREEZING
state again before restarting tasks, so that the call of
request_firmware() will be blocked until the end of this function
instead of returning an error.
Fixes: 247bc0374254 (PM / Sleep: Mitigate race between the freezer and request_firmware()) Link: https://bugzilla.novell.com/show_bug.cgi?id=873790 Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The block size for the dm-cache's data device must remained fixed for
the life of the cache. Disallow any attempt to change the cache's data
block size.
Signed-off-by: Mike Snitzer <snitzer@redhat.com> Acked-by: Joe Thornber <ejt@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The block size for the thin-pool's data device must remained fixed for
the life of the thin-pool. Disallow any attempt to change the
thin-pool's data block size.
It should be noted that attempting to change the data block size via
thin-pool table reload will be ignored as a side-effect of the thin-pool
handover that the thin-pool target does during thin-pool table reload.
Here is an example outcome of attempting to load a thin-pool table that
reduced the thin-pool's data block size from 1024K to 512K.
Before:
kernel: device-mapper: thin: 253:4: growing the data device from 204800 to 409600 blocks
After:
kernel: device-mapper: thin metadata: changing the data block size (from 2048 to 1024) is not supported
kernel: device-mapper: table: 253:4: thin-pool: Error creating metadata object
kernel: device-mapper: ioctl: error adding target to table
Signed-off-by: Mike Snitzer <snitzer@redhat.com> Acked-by: Joe Thornber <ejt@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Mauro reported that his AMD X2 using the powernow-k8 cpufreq driver
locked up when doing cpu hotplug.
Because we called set_cyc2ns_scale() from the time_cpufreq_notifier()
unconditionally, it gets called multiple times for each freq change,
instead of only the once, when the tsc_khz value actually changes.
Because it gets called more than once, we run out of cyc2ns data slots
and stall, waiting for a free one, but because we're half way offline,
there's no consumers to free slots.
By placing the call inside the condition that actually changes tsc_khz
we avoid superfluous calls and avoid the problem.
Reported-by: Mauro <registosites@hotmail.com> Tested-by: Mauro <registosites@hotmail.com> Fixes: 20d1c86a5776 ("sched/clock, x86: Rewrite cyc2ns() to avoid the need to disable IRQs") Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net> Cc: Viresh Kumar <viresh.kumar@linaro.org> Cc: Bin Gao <bin.gao@intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Mika Westerberg <mika.westerberg@linux.intel.com> Cc: Paul Gortmaker <paul.gortmaker@windriver.com> Cc: Stefani Seibold <stefani@seibold.net> Cc: linux-kernel@vger.kernel.org Signed-off-by: Peter Zijlstra <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Sharvil noticed with the posix timer_settime interface, using the
CLOCK_REALTIME_ALARM or CLOCK_BOOTTIME_ALARM clockid, if the users
tried to specify a relative time timer, it would incorrectly be
treated as absolute regardless of the state of the flags argument.
This patch corrects this, properly checking the absolute/relative flag,
as well as adds further error checking that no invalid flag bits are set.
In some cases we fetch the edid in the detect() callback
in order to determine what sort of monitor is connected.
If that happens, don't fetch the edid again in the get_modes()
callback or we will leak the edid.
Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Return IRQ_NONE if it was not our irq. This is necessary for the case
when qxl is sharing irq line with a device A in a crash kernel. If qxl
is initialized before A and A's irq was raised during this gap,
returning IRQ_HANDLED in this case will cause this irq to be raised
again after EOI since kernel think it was handled but in fact it was
not.
Cc: Gerd Hoffmann <kraxel@redhat.com> Signed-off-by: Jason Wang <jasowang@redhat.com> Signed-off-by: Dave Airlie <airlied@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If the value in the scratch register is 0, set it to the
max level. This fixes an issue where the console fb blanking
code calls back into the backlight driver on unblank and then
sets the backlight level to 0 after the driver has already
set the mode and enabled the backlight.
Certain GIC implementation, namely those found on earlier, single
cluster, Exynos SoCs, have registers mapped without per-CPU banking,
which means that the driver needs to use different offset for each CPU.
Currently the driver calculates the offset by multiplying value returned
by cpu_logical_map() by CPU offset parsed from DT. This is correct when
CPU topology is not specified in DT and aforementioned function returns
core ID alone. However when DT contains CPU topology, the function
changes to return cluster ID as well, which is non-zero on mentioned
SoCs and so breaks the calculation in GIC driver.
This patch fixes this by masking out cluster ID in CPU offset
calculation so that only core ID is considered. Multi-cluster Exynos
SoCs already have banked GIC implementations, so this simple fix should
be enough.
Reported-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> Reported-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com> Signed-off-by: Tomasz Figa <t.figa@samsung.com> Fixes: db0d4db22a78d ("ARM: gic: allow GIC to support non-banked setups") Link: https://lkml.kernel.org/r/1405610624-18722-1-git-send-email-t.figa@samsung.com Signed-off-by: Jason Cooper <jason@lakedaemon.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 3ab72f9156bb "dt-bindings: add GIC-400 binding" added the
"arm,gic-400" compatible string, but the corresponding IRQCHIP_DECLARE
was never added to the gic driver.
Therefore add the missing irqchip declaration for it.
Signed-off-by: Suravee Suthikulpanit <Suravee.Suthikulpanit@amd.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Removed additional empty line and adapted commit message to mark it
as fixing an issue. Signed-off-by: Heiko Stuebner <heiko@sntech.de> Acked-by: Will Deacon <will.deacon@arm.com> Fixes: 3ab72f9156bb ("dt-bindings: add GIC-400 binding") Link: https://lkml.kernel.org/r/2621565.f5eISveXXJ@diego Signed-off-by: Jason Cooper <jason@lakedaemon.net>
Patch 0a68214b "ARM: DT: Add binding for GIC virtualization extentions (VGIC)" added
the "arm,cortex-a7-gic" compatible string, but the corresponding IRQCHIP_DECLARE
was never added to the gic driver.
To let real Cortex-A7 SoCs use it, add the necessary declaration to the device driver.
Signed-off-by: Matthias Brugger <matthias.bgg@gmail.com> Link: https://lkml.kernel.org/r/1404388732-28890-1-git-send-email-matthias.bgg@gmail.com Fixes: 0a68214b76ca ("ARM: DT: Add binding for GIC virtualization extentions (VGIC)") Signed-off-by: Jason Cooper <jason@lakedaemon.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
ring_buffer_poll_wait() should always put the poll_table to its wait_queue
even there is immediate data available. Otherwise, the following epoll and
read sequence will eventually hang forever:
1. Put some data to make the trace_pipe ring_buffer read ready first
2. epoll_ctl(efd, EPOLL_CTL_ADD, trace_pipe_fd, ee)
3. epoll_wait()
4. read(trace_pipe_fd) till EAGAIN
5. Add some more data to the trace_pipe ring_buffer
6. epoll_wait() -> this epoll_wait() will block forever
~ During the epoll_ctl(efd, EPOLL_CTL_ADD,...) call in step 2,
ring_buffer_poll_wait() returns immediately without adding poll_table,
which has poll_table->_qproc pointing to ep_poll_callback(), to its
wait_queue.
~ During the epoll_wait() call in step 3 and step 6,
ring_buffer_poll_wait() cannot add ep_poll_callback() to its wait_queue
because the poll_table->_qproc is NULL and it is how epoll works.
~ When there is new data available in step 6, ring_buffer does not know
it has to call ep_poll_callback() because it is not in its wait queue.
Hence, block forever.
Other poll implementation seems to call poll_wait() unconditionally as the very
first thing to do. For example, tcp_poll() in tcp.c.
Link: http://lkml.kernel.org/p/20140610060637.GA14045@devbig242.prn2.facebook.com Fixes: 2a2cc8f7c4d0 "ftrace: allow the event pipe to be polled" Reviewed-by: Chris Mason <clm@fb.com> Signed-off-by: Martin Lau <kafai@fb.com> Signed-off-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
It is observed that sometimes Tx packet is downloaded without
adding driver's txpd header. This results in firmware parsing
garbage data as packet length. Sometimes firmware is unable
to read the packet if length comes out as invalid. This stops
further traffic and timeout occurs.
The root cause is uninitialized fields in tx_info(skb->cb) of
packet used to get garbage values. In this case if
MWIFIEX_BUF_FLAG_REQUEUED_PKT flag is mistakenly set, txpd
header was skipped. This patch makes sure that tx_info is
correctly initialized to fix the problem.
Reported-by: Andrew Wiley <wiley.andrew.j@gmail.com> Reported-by: Linus Gasser <list@markas-al-nour.org> Reported-by: Michael Hirsch <hirsch@teufel.de> Tested-by: Xinming Hu <huxm@marvell.com> Signed-off-by: Amitkumar Karwar <akarwar@marvell.com> Signed-off-by: Maithili Hinge <maithili@marvell.com> Signed-off-by: Avinash Patil <patila@marvell.com> Signed-off-by: Bing Zhao <bzhao@marvell.com> Signed-off-by: John W. Linville <linville@tuxdriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currently, any NMI is falsely handled by a NMI handler of NMI watchdog
if CondChgd bit in MSR_CORE_PERF_GLOBAL_STATUS MSR is set.
For example, we use external NMI to make system panic to get crash
dump, but in this case, the external NMI is falsely handled do to the
issue.
This commit deals with the issue simply by ignoring CondChgd bit.
Here is explanation in detail.
On x86 NMI watchdog uses performance monitoring feature to
periodically signal NMI each time performance counter gets overflowed.
intel_pmu_handle_irq() is called as a NMI_LOCAL handler from a NMI
handler of NMI watchdog, perf_event_nmi_handler(). It identifies an
owner of a given NMI by looking at overflow status bits in
MSR_CORE_PERF_GLOBAL_STATUS MSR. If some of the bits are set, then it
handles the given NMI as its own NMI.
The problem is that the intel_pmu_handle_irq() doesn't distinguish
CondChgd bit from other bits. Unlike the other status bits, CondChgd
bit doesn't represent overflow status for performance counters. Thus,
CondChgd bit cannot be thought of as a mark indicating a given NMI is
NMI watchdog's.
As a result, if CondChgd bit is set, any NMI is falsely handled by the
NMI handler of NMI watchdog. Also, if type of the falsely handled NMI
is either NMI_UNKNOWN, NMI_SERR or NMI_IO_CHECK, the corresponding
action is never performed until CondChgd bit is cleared.
I noticed this behavior on systems with Ivy Bridge processors: Intel
Xeon CPU E5-2630 v2 and Intel Xeon CPU E7-8890 v2. On both systems,
CondChgd bit in MSR_CORE_PERF_GLOBAL_STATUS MSR has already been set
in the beginning at boot. Then the CondChgd bit is immediately cleared
by next wrmsr to MSR_CORE_PERF_GLOBAL_CTRL MSR and appears to remain
0.
On the other hand, on older processors such as Nehalem, Xeon E7540,
CondChgd bit is not set in the beginning at boot.
I'm not sure about exact behavior of CondChgd bit, in particular when
this bit is set. Although I read Intel System Programmer's Manual to
figure out that, the descriptions I found are:
In 18.9.1:
"The MSR_PERF_GLOBAL_STATUS MSR also provides a ¡sticky bit¢ to
indicate changes to the state of performancmonitoring hardware"
In Table 35-2 IA-32 Architectural MSRs
63 CondChg: status bits of this register has changed.
These are different from the bahviour I see on the actual system as I
explained above.
At least, I think ignoring CondChgd bit should be enough for NMI
watchdog perspective.
The context check in perf_event_context_sched_out allows
non-cloned context to be part of the optimized schedule
out switch.
This could move non-cloned context into another workload
child. Once this child exits, the context is closed and
leaves all original (parent) events in closed state.
Any other new cloned event will have closed state and not
measure anything. And probably causing other odd bugs.
Signed-off-by: Jiri Olsa <jolsa@kernel.org> Signed-off-by: Peter Zijlstra <peterz@infradead.org> Cc: Arnaldo Carvalho de Melo <acme@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com> Cc: David Ahern <dsahern@gmail.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Link: http://lkml.kernel.org/r/1403598026-2310-2-git-send-email-jolsa@kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There is a benign buffer overflow in ip_options_compile spotted by
AddressSanitizer[1] :
Its benign because we always can access one extra byte in skb->head
(because header is followed by struct skb_shared_info), and in this case
this byte is not even used.
Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
*_result[len] is parsed as *(_result[len]) which is not at all what we
want to touch here.
Signed-off-by: Ben Hutchings <ben@decadent.org.uk> Fixes: 84a7c0b1db1c ("dns_resolver: assure that dns_query() result is null-terminated") Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
dns_query() credulously assumes that keys are null-terminated and
returns a copy of a memory block that is off by one.
Signed-off-by: Manuel Schölling <manuel.schoelling@gmx.de> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Nothing cleans up the objects created by
vnet_new(), they are completely leaked.
vnet_exit(), after doing the vio_unregister_driver() to clean
up ports, should call a helper function that iterates over vnet_list
and cleans up those objects. This includes unregister_netdevice()
as well as free_netdev().
Signed-off-by: Sowmini Varadhan <sowmini.varadhan@oracle.com> Acked-by: Dave Kleikamp <dave.kleikamp@oracle.com> Reviewed-by: Karl Volz <karl.volz@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fixed a bug that was introduced by my GRE-GRO patch
(bf5a755f5e9186406bbf50f4087100af5bd68e40 net-gre-gro: Add GRE
support to the GRO stack) that breaks the forwarding path
because various GSO related fields were not set. The bug will
cause on the egress path either the GSO code to fail, or a
GRE-TSO capable (NETIF_F_GSO_GRE) NICs to choke. The following
fix has been tested for both cases.
Signed-off-by: H.K. Jerry Chu <hkchu@google.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Obvious copy/paste error when I converted the ad_select to the new
option API. "lacp_rate" there should be "ad_select" so we can get the
proper value.
CC: Jay Vosburgh <j.vosburgh@gmail.com> CC: Veaceslav Falico <vfalico@gmail.com> CC: Andy Gospodarek <andy@greyhouse.net> CC: David S. Miller <davem@davemloft.net> Fixes: 9e5f5eebe765 ("bonding: convert ad_select to use the new option
API") Reported-by: Karim Scheik <karim.scheik@prisma-solutions.at> Signed-off-by: Nikolay Aleksandrov <nikolay@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The PPP channel MTU is used with Multilink PPP when ppp_mp_explode() (see
ppp_generic module) tries to determine how big a fragment might be. According
to RFC 1661, the MTU excludes the 2-byte PPP protocol field, see the
corresponding comment and code in ppp_mp_explode():
/*
* hdrlen includes the 2-byte PPP protocol field, but the
* MTU counts only the payload excluding the protocol field.
* (RFC1661 Section 2)
*/
mtu = pch->chan->mtu - (hdrlen - 2);
However, the pppoe module *does* include the PPP protocol field in the channel
MTU, which is wrong as it causes the PPP payload to be 1-2 bytes too big under
certain circumstances (one byte if PPP protocol compression is used, two
otherwise), causing the generated Ethernet packets to be dropped. So the pppoe
module has to subtract two bytes from the channel MTU. This error only
manifests itself when using Multilink PPP, as otherwise the channel MTU is not
used anywhere.
In the following, I will describe how to reproduce this bug. We configure two
pppd instances for multilink PPP over two PPPoE links, say eth2 and eth3, with
a MTU of 1492 bytes for each link and a MRRU of 2976 bytes. (This MRRU is
computed by adding the two link MTUs and subtracting the MP header twice, which
is 4 bytes long.) The necessary pppd statements on both sides are "multilink
mtu 1492 mru 1492 mrru 2976". On the client side, we additionally need "plugin
rp-pppoe.so eth2" and "plugin rp-pppoe.so eth3", respectively; on the server
side, we additionally need to start two pppoe-server instances to be able to
establish two PPPoE sessions, one over eth2 and one over eth3. We set the MTU
of the PPP network interface to the MRRU (2976) on both sides of the connection
in order to make use of the higher bandwidth. (If we didn't do that, IP
fragmentation would kick in, which we want to avoid.)
Now we send a ICMPv4 echo request with a payload of 2948 bytes from client to
server over the PPP link. This results in the following network packet:
These 2976 bytes do not exceed the MTU of the PPP network interface, so the
IP packet is not fragmented. Now the multilink PPP code in ppp_mp_explode()
prepends one protocol byte (0x21 for IPv4), making the packet one byte bigger
than the negotiated MRRU. So this packet would have to be divided in three
fragments. But this does not happen as each link MTU is assumed to be two bytes
larger. So this packet is diveded into two fragments only, one of size 1489 and
one of size 1488. Now we have for that bigger fragment:
1489 (PPP payload)
+ 4 (MP header)
+ 2 (PPP protocol field for the MP payload (0x3d))
+ 6 (PPPoE header)
--------------------------
1501 (Ethernet payload)
This packet exceeds the link MTU and is discarded.
If one configures the link MTU on the client side to 1501, one can see the
discarded Ethernet frames with tcpdump running on the client. A
ping -s 2948 -c 1 192.168.15.254
leads to the smaller fragment that is correctly received on the server side:
And the ICMPv4 echo request is successfully received at the server side:
IP (tos 0x0, ttl 64, id 21925, offset 0, flags [DF], proto ICMP (1),
length 2976)
192.168.222.2 > 192.168.15.254: ICMP echo request, id 30530, seq 0,
length 2956
The bug was introduced in commit c9aa6895371b2a257401f59d3393c9f7ac5a8698
("[PPPOE]: Advertise PPPoE MTU") from the very beginning. This patch applies
to 3.10 upwards but the fix can be applied (with minor modifications) to
kernels as old as 2.6.32.
Signed-off-by: Christoph Schulz <develop@kristov.de> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
While working on some other SCTP code, I noticed that some
structures shared with user space are leaking uninitialized
stack or heap buffer. In particular, struct sctp_sndrcvinfo
has a 2 bytes hole between .sinfo_flags and .sinfo_ppid that
remains unfilled by us in sctp_ulpevent_read_sndrcvinfo() when
putting this into cmsg. But also struct sctp_remote_error
contains a 2 bytes hole that we don't fill but place into a skb
through skb_copy_expand() via sctp_ulpevent_make_remote_error().
Both structures are defined by the IETF in RFC6458:
* Section 5.3.2. SCTP Header Information Structure:
A remote peer may send an Operation Error message to its peer.
This message indicates a variety of error conditions on an
association. The entire ERROR chunk as it appears on the wire
is included in an SCTP_REMOTE_ERROR event. Please refer to the
SCTP specification [RFC4960] and any extensions for a list of
possible error formats. An SCTP error notification has the
following format:
Fix this by setting both to 0 before filling them out. We also
have other structures shared between user and kernel space in
SCTP that contains holes (e.g. struct sctp_paddrthlds), but we
copy that buffer over from user space first and thus don't need
to care about it in that cases.
While at it, we can also remove lengthy comments copied from
the draft, instead, we update the comment with the correct RFC
number where one can look it up.
Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If the 'next' pointer of the last fragment buffer in a message is not
zeroed before reassembly, we risk ending up with a corrupt message,
since the reassembly function itself isn't doing this.
Currently, when a buffer is retrieved from the deferred queue of the
broadcast link, the next pointer is not cleared, with the result as
described above.
This commit corrects this, and thereby fixes a bug that may occur when
long broadcast messages are transmitted across dual interfaces. The bug
has been present since 40ba3cdf542a469aaa9083fa041656e59b109b90 ("tipc:
message reassembly using fragment chain")
This commit should be applied to both net and net-next.
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
On BE3, if the clear-interrupt bit of the EQ doorbell is not set the first
time it is armed, ocassionally we have observed that the EQ doesn't raise
anymore interrupts even if it is in armed state.
This patch fixes this by setting the clear-interrupt bit when EQs are
armed for the first time in be_open().
Signed-off-by: Suresh Reddy <Suresh.Reddy@emulex.com> Signed-off-by: Sathya Perla <sathya.perla@emulex.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
netlink_dump() returns a negative errno value on error. Until now,
netlink_recvmsg() directly recorded that negative value in sk->sk_err, but
that's wrong since sk_err takes positive errno values. (This manifests as
userspace receiving a positive return value from the recv() system call,
falsely indicating success.) This bug was introduced in the commit that
started checking the netlink_dump() return value, commit b44d211 (netlink:
handle errors from netlink_dump()).
Multithreaded Netlink dumps are one way to trigger this behavior in
practice, as described in the commit message for the userspace workaround
posted here:
http://openvswitch.org/pipermail/dev/2014-June/042339.html
This commit also fixes the same bug in netlink_poll(), introduced in commit cd1df525d (netlink: add flow control for memory mapped I/O).
Signed-off-by: Ben Pfaff <blp@nicira.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This commit fixes the command value generated for CSUM calculation
when running in big endian mode. The Ethernet protocol ID for IP was
being unconditionally byte-swapped in the layer 3 protocol check (with
swab16), which caused the mvneta driver to not function correctly in
big endian mode. This patch byte-swaps the ID conditionally with
htons.
Cc: <stable@vger.kernel.org> # v3.13+ Signed-off-by: Thomas Fitzsimmons <fitzsim@fitzsim.org> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
As reported by Maggie Mae Roxas, the mvneta driver doesn't behave
properly in 10 Mbit/s mode. This is due to a misconfiguration of the
MVNETA_GMAC_AUTONEG_CONFIG register: bit MVNETA_GMAC_CONFIG_MII_SPEED
must be set for a 100 Mbit/s speed, but cleared for a 10 Mbit/s speed,
which the driver was not properly doing. This commit adjusts that by
setting the MVNETA_GMAC_CONFIG_MII_SPEED bit only in 100 Mbit/s mode,
and relying on the fact that all the speed related bits of this
register are cleared at the beginning of the mvneta_adjust_link()
function.
This problem exists since c5aff18204da0 ("net: mvneta: driver for
Marvell Armada 370/XP network unit") which is the commit that
introduced the mvneta driver in the kernel.
Cc: <stable@vger.kernel.org> # v3.8+ Fixes: c5aff18204da0 ("net: mvneta: driver for Marvell Armada 370/XP network unit") Reported-by: Maggie Mae Roxas <maggie.mae.roxas@gmail.com> Cc: Maggie Mae Roxas <maggie.mae.roxas@gmail.com> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Setting just skb->sk without taking its reference and setting a
destructor is invalid. However, in the places where this was done, skb
is used in a way not requiring skb->sk setting. So dropping the setting
of skb->sk.
Thanks to Eric Dumazet <eric.dumazet@gmail.com> for correct solution.
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=79441 Reported-by: Ed Martin <edman007@edman007.com> Signed-off-by: Andrey Utkin <andrey.krieger.utkin@gmail.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The undo code assumes that, upon entering loss recovery, TCP
1) always retransmit something
2) the retransmission never fails locally (e.g., qdisc drop)
so undo_marker is set in tcp_enter_recovery() and undo_retrans is
incremented only when tcp_retransmit_skb() is successful.
When the assumption is broken because TCP's cwnd is too small to
retransmit or the retransmit fails locally. The next (DUP)ACK
would incorrectly revert the cwnd and the congestion state in
tcp_try_undo_dsack() or tcp_may_undo(). Subsequent (DUP)ACKs
may enter the recovery state. The sender repeatedly enter and
(incorrectly) exit recovery states if the retransmits continue to
fail locally while receiving (DUP)ACKs.
The fix is to initialize undo_retrans to -1 and start counting on
the first retransmission. Always increment undo_retrans even if the
retransmissions fail locally because they couldn't cause DSACKs to
undo the cwnd reduction.
Signed-off-by: Yuchung Cheng <ycheng@google.com> Signed-off-by: Neal Cardwell <ncardwell@google.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
1) create socket, bind and then setsockopt for add mc group.
mreq.imr_multiaddr.s_addr = inet_addr("255.0.0.37");
mreq.imr_interface.s_addr = inet_addr("192.168.1.2");
setsockopt(sockfd, IPPROTO_IP, IP_ADD_MEMBERSHIP, &mreq, sizeof(mreq));
2) drop the mc group for this socket.
mreq.imr_multiaddr.s_addr = inet_addr("255.0.0.37");
mreq.imr_interface.s_addr = inet_addr("0.0.0.0");
setsockopt(sockfd, IPPROTO_IP, IP_DROP_MEMBERSHIP, &mreq, sizeof(mreq));
3) and then drop the socket, I found the mc group was still used by the dev:
netstat -g
Interface RefCnt Group
--------------- ------ ---------------------
eth2 1 255.0.0.37
Normally even though the IP_DROP_MEMBERSHIP return error, the mc group still need
to be released for the netdev when drop the socket, but this process was broken when
route default is NULL, the reason is that:
The ip_mc_leave_group() will choose the in_dev by the imr_interface.s_addr, if input addr
is NULL, the default route dev will be chosen, then the ifindex is got from the dev,
then polling the inet->mc_list and return -ENODEV, but if the default route dev is NULL,
the in_dev and ifIndex is both NULL, when polling the inet->mc_list, the mc group will be
released from the mc_list, but the dev didn't dec the refcnt for this mc group, so
when dropping the socket, the mc_list is NULL and the dev still keep this group.
v1->v2: According Hideaki's suggestion, we should align with IPv6 (RFC3493) and BSDs,
so I add the checking for the in_dev before polling the mc_list, make sure when
we remove the mc group, dec the refcnt to the real dev which was using the mc address.
The problem would never happened again.
Signed-off-by: Ding Tianhong <dingtianhong@huawei.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
A bug was introduced in NETDEV_CHANGE notifier sequence causing the
arp table to be sometimes spuriously cleared (including manual arp
entries marked permanent), upon network link carrier changes.
The changed argument for the notifier was applied only to a single
caller of NETDEV_CHANGE, missing among others netdev_state_change().
So upon net_carrier events induced by the network, which are
triggering a call to netdev_state_change(), arp_netdev_event() would
decide whether to clear or not arp cache based on random/junk stack
values (a kind of read buffer overflow).
Fixes: be9efd365328 ("net: pass changed flags along with NETDEV_CHANGE event") Fixes: 6c8b4e3ff81b ("arp: flush arp cache on IFF_NOARP change") Signed-off-by: Loic Prylli <loicp@google.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add two device IDs found in an out-of-tree driver downloadable
from Netgear.
Signed-off-by: Bjørn Mork <bjorn@mork.no> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Some older router implementations still send Fragmentation Needed
errors with the Next-Hop MTU field set to zero. This is explicitly
described as an eventuality that hosts must deal with by the
standard (RFC 1191) since older standards specified that those
bits must be zero.
Linux had a generic (for all of IPv4) implementation of the algorithm
described in the RFC for searching a list of MTU plateaus for a good
value. Commit 46517008e116 ("ipv4: Kill ip_rt_frag_needed().")
removed this as part of the changes to remove the routing cache.
Subsequently any Fragmentation Needed packet with a zero Next-Hop
MTU has been discarded without being passed to the per-protocol
handlers or notifying userspace for raw sockets.
When there is a router which does not implement RFC 1191 on an
MTU limited path then this results in stalled connections since
large packets are discarded and the local protocols are not
notified so they never attempt to lower the pMTU.
One example I have seen is an OpenBSD router terminating IPSec
tunnels. It's worth pointing out that this case is distinct from
the BSD 4.2 bug which incorrectly calculated the Next-Hop MTU
since the commit in question dismissed that as a valid concern.
All of the per-protocols handlers implement the simple approach from
RFC 1191 of immediately falling back to the minimum value. Although
this is sub-optimal it is vastly preferable to connections hanging
indefinitely.
Remove the Next-Hop MTU != 0 check and allow such packets
to follow the normal path.
Fixes: 46517008e116 ("ipv4: Kill ip_rt_frag_needed().") Signed-off-by: Edward Allcutt <edward.allcutt@openmarket.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When in repair-mode and TCP_RECV_QUEUE is set, we end up calling
tcp_push with mss_now being 0. If data is in the send-queue and
tcp_set_skb_tso_segs gets called, we crash because it will divide by
mss_now:
This can easily be reproduced with the following packetdrill-script (the
"magic" with netem, sk_pacing and limit_output_bytes is done to prevent
the kernel from pushing all segments, because hitting the limit without
doing this is not so easy with packetdrill):
// This forces that not all segments of the snd-queue will be pushed
+0 `tc qdisc add dev tun0 root netem delay 10ms`
+0 `sysctl -w net.ipv4.tcp_limit_output_bytes=2`
+0 setsockopt(4, SOL_SOCKET, 47, [2], 4) = 0
// This now will make the write push the remaining segments
+0 setsockopt(4, SOL_SOCKET, 47, [20000], 4) = 0
+0 `sysctl -w net.ipv4.tcp_limit_output_bytes=130000`
// Now we will crash
+0 write(4,...,1000) = 1000
This happens since ec3423257508 (tcp: fix retransmission in repair
mode). Prior to that, the call to tcp_push was prevented by a check for
tp->repair.
The patch fixes it, by adding the new goto-label out_nopush. When exiting
tcp_sendmsg and a push is not required, which is the case for tp->repair,
we go to this label.
When repairing and calling send() with TCP_RECV_QUEUE, the data is
actually put in the receive-queue. So, no push is required because no
data has been added to the send-queue.
Cc: Andrew Vagin <avagin@openvz.org> Cc: Pavel Emelyanov <xemul@parallels.com> Fixes: ec3423257508 (tcp: fix retransmission in repair mode) Signed-off-by: Christoph Paasch <christoph.paasch@uclouvain.be> Acked-by: Andrew Vagin <avagin@openvz.org> Acked-by: Pavel Emelyanov <xemul@parallels.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reported-by: Michel Lespinasse <walken@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Ariel Elior <ariel.elior@qlogic.com> Fixes: d46d132cc021 ("bnx2x: use netdev_alloc_frag()") Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix this problem by freeing percpu stats from dev->destructor() instead
of ndo_uninit()
Reported-by: Madalin-Cristian Bucur <madalin.bucur@freescale.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Tested-by: Madalin-Cristian Bucur <madalin.bucur@freescale.com> Fixes: 5a4ae5f6e7d4 ("vlan: unnecessary to check if vlan_pcpu_stats is NULL") Cc: Li RongQing <roy.qing.li@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
sk_dst_cache has __rcu annotation, so we need a cast to avoid
following sparse error :
include/net/sock.h:1774:19: warning: incorrect type in initializer (different address spaces)
include/net/sock.h:1774:19: expected struct dst_entry [noderef] <asn:4>*__ret
include/net/sock.h:1774:19: got struct dst_entry *dst
Signed-off-by: Eric Dumazet <edumazet@google.com> Reported-by: kbuild test robot <fengguang.wu@intel.com> Fixes: 7f502361531e ("ipv4: irq safe sk_dst_[re]set() and ipv4_sk_update_pmtu() fix") Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We have two different ways to handle changes to sk->sk_dst
First way (used by TCP) assumes socket lock is owned by caller, and use
no extra lock : __sk_dst_set() & __sk_dst_reset()
Another way (used by UDP) uses sk_dst_lock because socket lock is not
always taken. Note that sk_dst_lock is not softirq safe.
These ways are not inter changeable for a given socket type.
ipv4_sk_update_pmtu(), added in linux-3.8, added a race, as it used
the socket lock as synchronization, but users might be UDP sockets.
Instead of converting sk_dst_lock to a softirq safe version, use xchg()
as we did for sk_rx_dst in commit e47eb5dfb296b ("udp: ipv4: do not use
sk_dst_lock from softirq context")
In a follow up patch, we probably can remove sk_dst_lock, as it is
only used in IPv6.
Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Steffen Klassert <steffen.klassert@secunet.com> Fixes: 9cb3a50c5f63e ("ipv4: Invalidate the socket cached route on pmtu events if possible") Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When IP route cache had been removed in linux-3.6, we broke assumption
that dst entries were all freed after rcu grace period. DST_NOCACHE
dst were supposed to be freed from dst_release(). But it appears
we want to keep such dst around, either in UDP sockets or tunnels.
In sk_dst_get() we need to make sure dst refcount is not 0
before incrementing it, or else we might end up freeing a dst
twice.
DST_NOCACHE set on a dst does not mean this dst can not be attached
to a socket or a tunnel.
Then, before actual freeing, we need to observe a rcu grace period
to make sure all other cpus can catch the fact the dst is no longer
usable.
Signed-off-by: Eric Dumazet <edumazet@google.com> Reported-by: Dormando <dormando@rydia.net> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch fixes a kernel BUG_ON in skb_segment. It is hit when
testing two VMs on openvswitch with one VM acting as VXLAN gateway.
During VXLAN packet GSO, skb_segment is called with skb->data
pointing to inner TCP payload. skb_segment calls skb_network_protocol
to retrieve the inner protocol. skb_network_protocol actually expects
skb->data to point to MAC and it calls pskb_may_pull with ETH_HLEN.
This ends up pulling in ETH_HLEN data from header tail. As a result,
pskb_trim logic is skipped and BUG_ON is hit later.
Move skb_push in front of skb_network_protocol so that skb->data
lines up properly.
Signed-off-by: Wei-Chun Chao <weichunc@plumgrid.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Messages from the modem exceeding 256 bytes cause communication
failure.
The WDM protocol is strictly "read on demand", meaning that we only
poll for unread data after receiving a notification from the modem.
Since we have no way to know how much data the modem has to send,
we must make sure that the buffer we provide is "big enough".
Message truncation does not work. Truncated messages are left unread
until the modem has another message to send. Which often won't
happen until the userspace application has given up waiting for the
final part of the last message, and therefore sends another command.
With a proper CDC WDM function there is a descriptor telling us
which buffer size the modem uses. But with this vendor specific
implementation there is no known way to calculate the exact "big
enough" number. It is an unknown property of the modem firmware.
Experience has shown that 256 is too small. The discussion of
this failure ended up concluding that 512 might be too small as
well. So 1024 seems like a reasonable value for now.
Fixes: 41c47d8cfd68 ("net: huawei_cdc_ncm: Introduce the huawei_cdc_ncm driver") Cc: Enrico Mioso <mrkiko.rs@gmail.com> Reported-by: Dan Williams <dcbw@redhat.com> Signed-off-by: Bjørn Mork <bjorn@mork.no> Acked-By: Enrico Mioso <mrkiko.rs@gmail.com> Tested-by: Dan Williams <dcbw@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
skb_cow called in vlan_reorder_header does not free the skb when it failed,
and vlan_reorder_header returns NULL to reset original skb when it is called
in vlan_untag, lead to a memory leak.
Signed-off-by: Li RongQing <roy.qing.li@gmail.com> Acked-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When writing to the sysctl field net.sctp.auth_enable, it can well
be that the user buffer we handed over to proc_dointvec() via
proc_sctp_do_auth() handler contains something other than integers.
In that case, we would set an uninitialized 4-byte value from the
stack to net->sctp.auth_enable that can be leaked back when reading
the sysctl variable, and it can unintentionally turn auth_enable
on/off based on the stack content since auth_enable is interpreted
as a boolean.
Fix it up by making sure proc_dointvec() returned sucessfully.
Fixes: b14878ccb7fa ("net: sctp: cache auth_enable per endpoint") Reported-by: Florian Westphal <fwestpha@redhat.com> Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Acked-by: Vlad Yasevich <vyasevich@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If there is an MSS change (or misbehaving receiver) that causes a SACK
to arrive that covers the end of an skb but is less than one MSS, then
tcp_match_skb_to_sack() was rounding up pkt_len to the full length of
the skb ("Round if necessary..."), then chopping all bytes off the skb
and creating a zero-byte skb in the write queue.
This was visible now because the recently simplified TLP logic in bef1909ee3ed1c ("tcp: fixing TLP's FIN recovery") could find that 0-byte
skb at the end of the write queue, and now that we do not check that
skb's length we could send it as a TLP probe.
Previously we would find the new_len > skb->len check failing, so we
would fall through and set pkt_len = new_len = 4000 and chop off
pkt_len of 4000 from the 4000-byte skb, leaving a 0-byte segment
afterward in the write queue.
With this new commit, we notice that the new new_len >= skb->len check
succeeds, so that we return without trying to fragment.
Fixes: adb92db857ee ("tcp: Make SACK code to split only at mss boundaries") Reported-by: Eric Dumazet <edumazet@google.com> Signed-off-by: Neal Cardwell <ncardwell@google.com> Cc: Eric Dumazet <edumazet@google.com> Cc: Yuchung Cheng <ycheng@google.com> Cc: Ilpo Jarvinen <ilpo.jarvinen@helsinki.fi> Acked-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
sysctl handler proc_sctp_do_hmac_alg(), proc_sctp_do_rto_min() and
proc_sctp_do_rto_max() do not properly reflect some error cases
when writing values via sysctl from internal proc functions such
as proc_dointvec() and proc_dostring().
In all these cases we pass the test for write != 0 and partially
do additional work just to notice that additional sanity checks
fail and we return with hard-coded -EINVAL while proc_do*
functions might also return different errors. So fix this up by
simply testing a successful return of proc_do* right after
calling it.
This also allows to propagate its return value onwards to the user.
While touching this, also fix up some minor style issues.
Fixes: 4f3fdf3bc59c ("sctp: add check rto_min and rto_max in sysctl") Fixes: 3c68198e7511 ("sctp: Make hmac algorithm selection for cookie generation dynamic") Signed-off-by: Daniel Borkmann <dborkman@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The commit "slip: Fix deadlock in write_wakeup" fixes a deadlock caused
by a change made in both slcan and slip. This is a direct port of that
fix.
Signed-off-by: Tyler Hall <tylerwhall@gmail.com> Cc: Oliver Hartkopp <socketcan@hartkopp.net> Cc: Andre Naujoks <nautsch2@gmail.com> Cc: David S. Miller <davem@davemloft.net> Cc: linux-kernel@vger.kernel.org Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Use schedule_work() to avoid potentially taking the spinlock in
interrupt context.
Commit cc9fa74e2a ("slip/slcan: added locking in wakeup function") added
necessary locking to the wakeup function and 367525c8c2/ddcde142be ("can:
slcan: Fix spinlock variant") converted it to spin_lock_bh() because the lock
is also taken in timers.
Disabling softirqs is not sufficient, however, as tty drivers may call
write_wakeup from interrupt context. This driver calls tty->ops->write() with
its spinlock held, which may immediately cause an interrupt on the same CPU and
subsequent spin_bug().
Simply converting to spin_lock_irq/irqsave() prevents this deadlock, but
causes lockdep to point out a possible circular locking dependency
between these locks:
The slip transmit is holding the slip spinlock when calling the tty write.
This grabs the port lock. On an interrupt, the handler grabs the port
lock and calls write_wakeup which grabs the slip lock. This could be a
problem if a serial interrupt occurs on another CPU during the slip
transmit.
To deal with these issues, don't grab the lock in the wakeup function by
deferring the writeout to a workqueue. Also hold the lock during close
when de-assigning the tty pointer to safely disarm the worker and
timers.
This bug is easily reproducible on the first transmit when slip is
used with the standard 8250 serial driver.
Signed-off-by: Tyler Hall <tylerwhall@gmail.com> Cc: Oliver Hartkopp <socketcan@hartkopp.net> Cc: Andre Naujoks <nautsch2@gmail.com> Cc: David S. Miller <davem@davemloft.net> Cc: linux-kernel@vger.kernel.org Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch fixes 3 similar bugs where incoming packets might be routed into
wrong non-wildcard tunnels:
1) Consider the following setup:
ip address add 1.1.1.1/24 dev eth0
ip address add 1.1.1.2/24 dev eth0
ip tunnel add ipip1 remote 2.2.2.2 local 1.1.1.1 mode ipip dev eth0
ip link set ipip1 up
Incoming ipip packets from 2.2.2.2 were routed into ipip1 even if it has dst =
1.1.1.2. Moreover even if there was wildcard tunnel like
ip tunnel add ipip0 remote 2.2.2.2 local any mode ipip dev eth0
but it was created before explicit one (with local 1.1.1.1), incoming ipip
packets with src = 2.2.2.2 and dst = 1.1.1.2 were still routed into ipip1.
Same issue existed with all tunnels that use ip_tunnel_lookup (gre, vti)
2) ip address add 1.1.1.1/24 dev eth0
ip tunnel add ipip1 remote 2.2.146.85 local 1.1.1.1 mode ipip dev eth0
ip link set ipip1 up
Incoming ipip packets with dst = 1.1.1.1 were routed into ipip1, no matter what
src address is. Any remote ip address which has ip_tunnel_hash = 0 raised this
issue, 2.2.146.85 is just an example, there are more than 4 million of them.
And again, wildcard tunnel like
ip tunnel add ipip0 remote any local 1.1.1.1 mode ipip dev eth0
wouldn't be ever matched if it was created before explicit tunnel like above.
Gre & vti tunnels had the same issue.
3) ip address add 1.1.1.1/24 dev eth0
ip tunnel add gre1 remote 2.2.146.84 local 1.1.1.1 key 1 mode gre dev eth0
ip link set gre1 up
Any incoming gre packet with key = 1 were routed into gre1, no matter what
src/dst addresses are. Any remote ip address which has ip_tunnel_hash = 0 raised
the issue, 2.2.146.84 is just an example, there are more than 4 million of them.
Wildcard tunnel like
ip tunnel add gre2 remote any local any key 1 mode gre dev eth0
wouldn't be ever matched if it was created before explicit tunnel like above.
All this stuff happened because while looking for a wildcard tunnel we didn't
check that matched tunnel is a wildcard one. Fixed.
Signed-off-by: Dmitry Popov <ixaphire@qrator.net> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Previous commit c3a0dce35af0 fixed an overrun for the RAR on i218 devices.
This commit also attempted to homogenize the RAR/SHRA access for all parts
accessed by the e1000e driver. This change introduced an error for
assigning MAC addresses to guest OS's for 82579 devices.
Only RAR[0] is accessible to the driver for 82579 parts, and additional
addresses must be placed into the SHRA[L|H] registers. The rar_entry_count
was changed in the previous commit to an inaccurate value that accounted
for all RAR and SHRA registers, not just the ones usable by the driver.
This patch fixes the count to the correct value and adjusts the
e1000_rar_set_pch2lan() function to user the correct index.
Cc: John Greene <jogreene@redhat.com> Signed-off-by: Dave Ertman <davidx.m.ertman@intel.com> Tested-by: Aaron Brown <aaron.f.brown@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com> Cc: "Alexander Y. Fomichev" <aleksandr.fomichev@x5.ru> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
shmem_fault() is the actual culprit in trinity's hole-punch starvation,
and the most significant cause of such problems: since a page faulted is
one that then appears page_mapped(), needing unmap_mapping_range() and
i_mmap_mutex to be unmapped again.
But it is not the only way in which a page can be brought into a hole in
the radix_tree while that hole is being punched; and Vlastimil's testing
implies that if enough other processors are busy filling in the hole,
then shmem_undo_range() can be kept from completing indefinitely.
shmem_file_splice_read() is the main other user of SGP_CACHE, which can
instantiate shmem pagecache pages in the read-only case (without holding
i_mutex, so perhaps concurrently with a hole-punch). Probably it's
silly not to use SGP_READ already (using the ZERO_PAGE for holes): which
ought to be safe, but might bring surprises - not a change to be rushed.
shmem_read_mapping_page_gfp() is an internal interface used by
drivers/gpu/drm GEM (and next by uprobes): it should be okay. And
shmem_file_read_iter() uses the SGP_DIRTY variant of SGP_CACHE, when
called internally by the kernel (perhaps for a stacking filesystem,
which might rely on holes to be reserved): it's unclear whether it could
be provoked to keep hole-punch busy or not.
We could apply the same umbrella as now used in shmem_fault() to
shmem_file_splice_read() and the others; but it looks ugly, and use over
a range raises questions - should it actually be per page? can these get
starved themselves?
The origin of this part of the problem is my v3.1 commit d0823576bf4b
("mm: pincer in truncate_inode_pages_range"), once it was duplicated
into shmem.c. It seemed like a nice idea at the time, to ensure
(barring RCU lookup fuzziness) that there's an instant when the entire
hole is empty; but the indefinitely repeated scans to ensure that make
it vulnerable.
Revert that "enhancement" to hole-punch from shmem_undo_range(), but
retain the unproblematic rescanning when it's truncating; add a couple
of comments there.
Remove the "indices[0] >= end" test: that is now handled satisfactorily
by the inner loop, and mem_cgroup_uncharge_start()/end() are too light
to be worth avoiding here.
But if we do not always loop indefinitely, we do need to handle the case
of swap swizzled back to page before shmem_free_swap() gets it: add a
retry for that case, as suggested by Konstantin Khlebnikov; and for the
case of page swizzled back to swap, as suggested by Johannes Weiner.
Signed-off-by: Hugh Dickins <hughd@google.com> Reported-by: Sasha Levin <sasha.levin@oracle.com> Suggested-by: Vlastimil Babka <vbabka@suse.cz> Cc: Konstantin Khlebnikov <koct9i@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Lukas Czerner <lczerner@redhat.com> Cc: Dave Jones <davej@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit f00cdc6df7d7 ("shmem: fix faulting into a hole while it's
punched") was buggy: Sasha sent a lockdep report to remind us that
grabbing i_mutex in the fault path is a no-no (write syscall may already
hold i_mutex while faulting user buffer).
We tried a completely different approach (see following patch) but that
proved inadequate: good enough for a rational workload, but not good
enough against trinity - which forks off so many mappings of the object
that contention on i_mmap_mutex while hole-puncher holds i_mutex builds
into serious starvation when concurrent faults force the puncher to fall
back to single-page unmap_mapping_range() searches of the i_mmap tree.
So return to the original umbrella approach, but keep away from i_mutex
this time. We really don't want to bloat every shmem inode with a new
mutex or completion, just to protect this unlikely case from trinity.
So extend the original with wait_queue_head on stack at the hole-punch
end, and wait_queue item on the stack at the fault end.
This involves further use of i_lock to guard against the races: lockdep
has been happy so far, and I see fs/inode.c:unlock_new_inode() holds
i_lock around wake_up_bit(), which is comparable to what we do here.
i_lock is more convenient, but we could switch to shmem's info->lock.
This issue has been tagged with CVE-2014-4171, which will require commit f00cdc6df7d7 and this and the following patch to be backported: we
suggest to 3.1+, though in fact the trinity forkbomb effect might go
back as far as 2.6.16, when madvise(,,MADV_REMOVE) came in - or might
not, since much has changed, with i_mmap_mutex a spinlock before 3.0.
Anyone running trinity on 3.0 and earlier? I don't think we need care.
Signed-off-by: Hugh Dickins <hughd@google.com> Reported-by: Sasha Levin <sasha.levin@oracle.com> Tested-by: Sasha Levin <sasha.levin@oracle.com> Cc: Vlastimil Babka <vbabka@suse.cz> Cc: Konstantin Khlebnikov <koct9i@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Lukas Czerner <lczerner@redhat.com> Cc: Dave Jones <davej@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Trinity finds that mmap access to a hole while it's punched from shmem
can prevent the madvise(MADV_REMOVE) or fallocate(FALLOC_FL_PUNCH_HOLE)
from completing, until the reader chooses to stop; with the puncher's
hold on i_mutex locking out all other writers until it can complete.
It appears that the tmpfs fault path is too light in comparison with its
hole-punching path, lacking an i_data_sem to obstruct it; but we don't
want to slow down the common case.
Extend shmem_fallocate()'s existing range notification mechanism, so
shmem_fault() can refrain from faulting pages into the hole while it's
punched, waiting instead on i_mutex (when safe to sleep; or repeatedly
faulting when not).
We should always prefer to use full RTS protection. Using
CTS to self gives a meaningless improvement, but this flow
is much harder for the firmware which is likely to have
issues with it.