1. MEI_INTEROP_TIMEOUT is in seconds not in jiffies
so we use mei_secs_to_jiffies macro
While cold boot is fast this is relevant in resume
2. wait_event_interruptible_timeout can return with
-ERESTARTSYS so do not override it with -ETIMEDOUT
3.Adjust error message
ME HW ready bit is down after hw reset was asserted or on error.
Only on error we need to enter the reset flow, additional reset
need to be prevented when reset was triggered during
initialization , power up/down or a reset is already in progress
If there are UNUSABLE regions in the machine memory map, dom0 will
attempt to map them 1:1 which is not permitted by Xen and the kernel
will crash.
There isn't anything interesting in the UNUSABLE region that the dom0
kernel needs access to so we can avoid making the 1:1 mapping and
treat it as RAM.
We only do this for dom0, as that is where tboot case shows up.
A PV domU could have an UNUSABLE region in its pseudo-physical map
and would need to be handled in another patch.
This fixes a boot failure on hosts with tboot.
tboot marks a region in the e820 map as unusable and the dom0 kernel
would attempt to map this region and Xen does not permit unusable
regions to be mapped by guests.
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
[v1: Altered the patch and description with domU's with UNUSABLE regions] Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This is the updated version of df54d6fa5427 ("x86 get_unmapped_area():
use proper mmap base for bottom-up direction") that only randomizes the
mmap base address once.
Signed-off-by: Radu Caragea <sinaelgl@gmail.com> Reported-and-tested-by: Jeff Shorey <shoreyjeff@gmail.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Michel Lespinasse <walken@google.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Rik van Riel <riel@redhat.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Adrian Sendroiu <molecula2788@gmail.com> Cc: Kamal Mostafa <kamal@canonical.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The commit isn't necessarily wrong, but because it recalculates the
random mmap_base every time, it seems to confuse user memory allocators
that expect contiguous mmap allocations even when the mmap address isn't
specified.
In particular, the MATLAB Java runtime seems to be unhappy. See
https://bugzilla.kernel.org/show_bug.cgi?id=60774
So we'll want to apply the random offset only once, and Radu has a patch
for that. Revert this older commit in order to apply the other one.
Reported-by: Jeff Shorey <shoreyjeff@gmail.com> Cc: Radu Caragea <sinaelgl@gmail.com> Cc: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There is a nasty bug in the SCSI SG_IO ioctl that in some circumstances
leads to one process writing data into the address space of some other
random unrelated process if the ioctl is interrupted by a signal.
What happens is the following:
- A process issues an SG_IO ioctl with direction DXFER_FROM_DEV (ie the
underlying SCSI command will transfer data from the SCSI device to
the buffer provided in the ioctl)
- Before the command finishes, a signal is sent to the process waiting
in the ioctl. This will end up waking up the sg_ioctl() code:
result = wait_event_interruptible(sfp->read_wait,
(srp_done(sfp, srp) || sdp->detached));
but neither srp_done() nor sdp->detached is true, so we end up just
setting srp->orphan and returning to userspace:
srp->orphan = 1;
write_unlock_irq(&sfp->rq_list_lock);
return result; /* -ERESTARTSYS because signal hit process */
At this point the original process is done with the ioctl and
blithely goes ahead handling the signal, reissuing the ioctl, etc.
- Eventually, the SCSI command issued by the first ioctl finishes and
ends up in sg_rq_end_io(). At the end of that function, we run through:
write_lock_irqsave(&sfp->rq_list_lock, iflags);
if (unlikely(srp->orphan)) {
if (sfp->keep_orphan)
srp->sg_io_owned = 0;
else
done = 0;
}
srp->done = done;
write_unlock_irqrestore(&sfp->rq_list_lock, iflags);
if (likely(done)) {
/* Now wake up any sg_read() that is waiting for this
* packet.
*/
wake_up_interruptible(&sfp->read_wait);
kill_fasync(&sfp->async_qp, SIGPOLL, POLL_IN);
kref_put(&sfp->f_ref, sg_remove_sfp);
} else {
INIT_WORK(&srp->ew.work, sg_rq_end_io_usercontext);
schedule_work(&srp->ew.work);
}
Since srp->orphan *is* set, we set done to 0 (assuming the
userspace app has not set keep_orphan via an SG_SET_KEEP_ORPHAN
ioctl), and therefore we end up scheduling sg_rq_end_io_usercontext()
to run in a workqueue.
- In workqueue context we go through sg_rq_end_io_usercontext() ->
sg_finish_rem_req() -> blk_rq_unmap_user() -> ... ->
bio_uncopy_user() -> __bio_copy_iov() -> copy_to_user().
The key point here is that we are doing copy_to_user() on a
workqueue -- that is, we're on a kernel thread with current->mm
equal to whatever random previous user process was scheduled before
this kernel thread. So we end up copying whatever data the SCSI
command returned to the virtual address of the buffer passed into
the original ioctl, but it's quite likely we do this copying into a
different address space!
As suggested by James Bottomley <James.Bottomley@hansenpartnership.com>,
add a check for current->mm (which is NULL if we're on a kernel thread
without a real userspace address space) in bio_uncopy_user(), and skip
the copy if we're on a kernel thread.
There's no reason that I can think of for any caller of bio_uncopy_user()
to want to do copying on a kernel thread with a random active userspace
address space.
Huge thanks to Costa Sapuntzakis <costa@purestorage.com> for the
original pointer to this bug in the sg code.
Signed-off-by: Roland Dreier <roland@purestorage.com> Tested-by: David Milburn <dmilburn@redhat.com> Cc: Jens Axboe <axboe@kernel.dk> Signed-off-by: James Bottomley <JBottomley@Parallels.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We want ppc64 to be able to select between optimised assembly
checksum routines in big endian and the generic lib/checksum.c
routines in little endian.
The lpfc driver is forcing CONFIG_GENERIC_CSUM on which means
we are unable to make the decision to enable it in the arch
Kconfig. If the option exists it is always forced on.
This got introduced in 3.10 via commit 6a7252fdb0c3 ([SCSI] lpfc:
fix up Kconfig dependencies). I spoke to Randy about it and
the original issue was with CRC_T10DIF not being defined.
As such, remove the select of CONFIG_GENERIC_CSUM.
Signed-off-by: Anton Blanchard <anton@samba.org> Signed-off-by: James Bottomley <JBottomley@Parallels.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Apparently, the ref_count for some scsi_device drops down to zero,
triggering device removal through execute_in_process_context(), while
the lldd error recovery thread iterates through a scsi device list.
Unfortunately, execute_in_process_context() decides to immediately
execute that device removal function, instead of scheduling asynchronous
execution, since it detects process context and thinks it is safe to do
so. But almost all calls to shost_for_each_device() in our lldd are
inside spin_lock_irq, even in thread context. Obviously, schedule()
inside spin_lock_irq sections is a bad idea.
Change the lldd to use the proper iterator function,
__shost_for_each_device(), in combination with required locking.
Occurences that need to be changed include all calls in zfcp_erp.c,
since those might be executed in zfcp error recovery thread context
with a lock held.
Other occurences of shost_for_each_device() in zfcp_fsf.c do not
need to be changed (no process context, no surrounding locking).
The problem was introduced in Linux 2.6.37 by commit b62a8d9b45b971a67a0f8413338c230e3117dff5
"[SCSI] zfcp: Use SCSI device data zfcp_scsi_dev instead of zfcp_unit".
Reported-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Martin Peschke <mpeschke@linux.vnet.ibm.com> Signed-off-by: Steffen Maier <maier@linux.vnet.ibm.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch adds wait_event_interruptible_lock_irq_timeout(), which is a
straight-forward descendant of wait_event_interruptible_timeout() and
wait_event_interruptible_lock_irq().
The zfcp driver used to call wait_event_interruptible_timeout()
in combination with some intricate and error-prone locking. Using
wait_event_interruptible_lock_irq_timeout() as a replacement
nicely cleans up that locking.
This rework removes a situation that resulted in a locking imbalance
in zfcp_qdio_sbal_get():
BUG: workqueue leaked lock or atomic: events/1/0xffffff00/10
last function: zfcp_fc_wka_port_offline+0x0/0xa0 [zfcp]
It was introduced by commit c2af7545aaff3495d9bf9a7608c52f0af86fb194
"[SCSI] zfcp: Do not wait for SBALs on stopped queue", which had a new
code path related to ZFCP_STATUS_ADAPTER_QDIOUP that took an early exit
without a required lock being held. The problem occured when a
special, non-SCSI I/O request was being submitted in process context,
when the adapter's queues had been torn down. In this case the bug
surfaced when the Fibre Channel port connection for a well-known address
was closed during a concurrent adapter shut-down procedure, which is a
rare constellation.
This patch also fixes these warnings from the sparse tool (make C=1):
drivers/s390/scsi/zfcp_qdio.c:224:12: warning: context imbalance in
'zfcp_qdio_sbal_check' - wrong count at exit
drivers/s390/scsi/zfcp_qdio.c:244:5: warning: context imbalance in
'zfcp_qdio_sbal_get' - unexpected unlock
Last but not least, we get rid of that crappy lock-unlock-lock
sequence at the beginning of the critical section.
It is okay to call zfcp_erp_adapter_reopen() with req_q_lock held.
As Arjan pointed out, we mustn't do anything related to PCI
configuration until the device is properly enabled with
pci_enable_device().
Reported-by: Arjan van de Ven <arjan@linux.intel.com> Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This is because we nulify ctx->vif in iwlagn_mac_remove_interface()
before calling some other functions that teardown interface. To fix
just check ctx->vif on iwl_chswitch_done(). We should not call
ieee80211_chswitch_done() as channel switch works were already canceled
by mac80211 in ieee80211_do_stop() -> ieee80211_mgd_stop().
Fixing support for the Silicon Image 3826 port multiplier, by applying
to it the same quirks applied to the Silicon Image 3726. Specifically
fixes the repeated timeout/reset process which previously afflicted
the 3726, as described from line 290. Slightly based on notes from:
We want the data stored in "addr" and "qual", but the extra ampersands
mean we are copying stack data instead.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: John W. Linville <linville@tuxdriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 94ae9843 (usb: phy: rename all phy drivers to phy-$name-usb.c)
renamed drivers/usb/phy/otg_fsm.h to drivers/usb/phy/phy-fsm-usb.h
but changed drivers/usb/phy/phy-fsm-usb.c to include not existing
"phy-otg-fsm.h" instead of new "phy-fsm-usb.h". This breaks building:
...
drivers/usb/phy/phy-fsm-usb.c:32:25: fatal error: phy-otg-fsm.h: No such file or directory
compilation terminated.
make[3]: *** [drivers/usb/phy/phy-fsm-usb.o] Error 1
This commit also missed to modify drivers/usb/phy/phy-fsl-usb.h
to include new "phy-fsm-usb.h" instead of "otg_fsm.h" resulting
in another build breakage:
...
In file included from drivers/usb/phy/phy-fsl-usb.c:46:0:
drivers/usb/phy/phy-fsl-usb.h:18:21: fatal error: otg_fsm.h: No such file or directory
compilation terminated.
make[3]: *** [drivers/usb/phy/phy-fsl-usb.o] Error 1
Being a low-level component, various drivers (e.g. olpc-battery) assume
that it is ok to communicate with the OLPC Embedded Controller during
probe. Therefore the OLPC EC driver must be initialised before other
drivers try to use it. This was the case until it was recently moved
out of arch/x86 and restructured around commits ac2504151f5a ("Platform:
OLPC: turn EC driver into a platform_driver") and 85f90cf6ca56 ("x86:
OLPC: switch over to using new EC driver on x86").
Use arch_initcall so that olpc-ec is readied earlier, matching the
previous behaviour.
Fixes a regression introduced in Linux-3.6 where various drivers such as
olpc-battery and olpc-xo1-sci failed to load due to an inability to
communicate with the EC. The user-visible effect was a lack of battery
monitoring, missing ebook/lid switch input devices, etc.
Signed-off-by: Daniel Drake <dsd@laptop.org> Cc: Andres Salomon <dilinger@queued.net> Cc: Paul Fox <pgf@laptop.org> Cc: Thomas Gleixner <tglx@linutronix.de> 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>
Fix the issue with improper counting number of flying bio requests for
BIO_EOPNOTSUPP error detection case.
The sb_nbio must be incremented exactly the same number of times as
complete() function was called (or will be called) because
nilfs_segbuf_wait() will call wail_for_completion() for the number of
times set to sb_nbio:
do {
wait_for_completion(&segbuf->sb_bio_event);
} while (--segbuf->sb_nbio > 0);
Two functions complete() and wait_for_completion() must be called the
same number of times for the same sb_bio_event. Otherwise,
wait_for_completion() will hang or leak.
Remove double call of bio_put() in nilfs_end_bio_write() for the case of
BIO_EOPNOTSUPP error detection. The issue was found by Dan Carpenter
and he suggests first version of the fix too.
Already existing property flags are filled wrong for properties created from
initial FDT. This could cause problems if this DYNAMIC device-tree functions
are used later, i.e. properties are attached/detached/replaced. Simply dumping
flags from the running system show, that some initial static (not allocated via
kzmalloc()) nodes are marked as dynamic.
I putted some debug extensions to property_proc_show(..) :
..
+ if (OF_IS_DYNAMIC(pp))
+ pr_err("DEBUG: xxx : OF_IS_DYNAMIC\n");
+ if (OF_IS_DETACHED(pp))
+ pr_err("DEBUG: xxx : OF_IS_DETACHED\n");
when you operate on the nodes (e.g.: ~$ cat /proc/device-tree/*some_node*) you
will see that those flags are filled wrong, basically in most cases it will dump
a DYNAMIC or DETACHED status, which is in not true.
(BTW. this OF_IS_DETACHED is a own define for debug purposes which which just
make a test_bit(OF_DETACHED, &x->_flags)
If nodes are dynamic kernel is allowed to kfree() them. But it will crash
attempting to do so on the nodes from FDT -- they are not allocated via
kzmalloc().
Signed-off-by: Wladislav Wiebe <wladislav.kw@gmail.com> Acked-by: Alexander Sverdlin <alexander.sverdlin@nsn.com> Signed-off-by: Rob Herring <rob.herring@calxeda.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
After any "soft gfx reset" we must manually invalidate the TLBs
associated with each ring. Empirically, it seems that a
suspend/resume or D3-D0 cycle count as a "soft reset". The symptom is
that the hardware would fail to note the new address for its status
page, and so it would continue to write the shadow registers and
breadcrumbs into the old physical address (now used by something
completely different, scary). Whereas the driver would read the new
status page and never see any progress, it would appear that the GPU
hung immediately upon resume.
Based on a patch by naresh kumar kachhi <naresh.kumar.kacchi@intel.com>
Reported-by: Thiago Macieira <thiago@kde.org>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=64725 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Tested-by: Thiago Macieira <thiago@kde.org> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This bug (introduced in 3.10) in WREG32_OR made
commit d3418eacad403033e95e49dc14afa37c2112c134
"drm/radeon/evergreen: setup HDMI before enabling it"
cause a regression. Sometimes audio over HDMI wasn't working, sometimes
display was corrupted.
This fixes:
https://bugzilla.kernel.org/show_bug.cgi?id=60687
https://bugzilla.kernel.org/show_bug.cgi?id=60709
https://bugs.freedesktop.org/show_bug.cgi?id=67767
Commit dcd7b8bd63cb81c5b973bf86510ca3c80bbbd162 ("staging: comedi: put
module _after_ detach" by myself) reversed a couple of calls in
`comedi_device_attach()` when recovering from an error returned by the
low-level driver's 'attach' handler. Unfortunately, that introduced a
NULL pointer dereference bug as `dev->driver` is NULL after the call to
`comedi_device_detach()`. We still have a pointer to the low-level
comedi driver structure in the `driv` variable, so use that instead.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit f6f91b0d9fd9 ("ARM: allow kuser helpers to be removed from the
vector page") introduced some help text for the CONFIG_KUSER_HELPERS
option which is rather contradictory.
Let's fix that, and improve it a little.
Signed-off-by: Nicolas Pitre <nico@linaro.org> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This is a port of c95eb3184ea1 ("ARM: 7809/1: perf: fix event validation
for software group leaders") to arm64, which fixes a panic in the arm64
perf backend found as a result of Vince's fuzzing tool.
This is a port of d9f966357b14 ("ARM: 7810/1: perf: Fix array out of
bounds access in armpmu_map_hw_event()") to arm64, which fixes an oops
in the arm64 perf backend found as a result of Vince's fuzzing tool.
When a event is being bound to a VCPU there is a window between the
EVTCHNOP_bind_vpcu call and the adjustment of the local per-cpu masks
where an event may be lost. The hypervisor upcalls the new VCPU but
the kernel thinks that event is still bound to the old VCPU and
ignores it.
There is even a problem when the event is being bound to the same VCPU
as there is a small window beween the clear_bit() and set_bit() calls
in bind_evtchn_to_cpu(). When scanning for pending events, the kernel
may read the bit when it is momentarily clear and ignore the event.
Avoid this by masking the event during the whole bind operation.
Signed-off-by: David Vrabel <david.vrabel@citrix.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Reviewed-by: Jan Beulich <jbeulich@suse.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The sizeof() argument in init_evtchn_cpu_bindings() is incorrect
resulting in only the first 64 (or 32 in 32-bit guests) ports having
their bindings being initialized to VCPU 0.
In most cases this does not cause a problem as request_irq() will set
the irq affinity which will set the correct local per-cpu mask.
However, if the request_irq() is called on a VCPU other than 0, there
is a window between the unmasking of the event and the affinity being
set were an event may be lost because it is not locally unmasked on
any VCPU. If request_irq() is called on VCPU 0 then local irqs are
disabled during the window and the race does not occur.
Fix this by initializing all NR_EVENT_CHANNEL bits in the local
per-cpu masks.
Signed-off-by: David Vrabel <david.vrabel@citrix.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
OpenFirmware wasn't quite following the protocol described in boot.txt
and the kernel has detected this through use of the sentinel value
in boot_params. OFW does zero out almost all of the stuff that it should
do, but not the sentinel.
This causes the kernel to clear olpc_ofw_header, which breaks x86 OLPC
support.
OpenFirmware has now been fixed. However, it would be nice if we could
maintain Linux compatibility with old firmware versions. To do that, we just
have to avoid zeroing out olpc_ofw_header.
OFW does not write to any other parts of the header that are being zapped
by the sentinel-detection code, and all users of olpc_ofw_header are
somewhat protected through checking for the OLPC_OFW_SIG magic value
before using it. So this should not cause any problems for anyone.
Signed-off-by: Daniel Drake <dsd@laptop.org> Link: http://lkml.kernel.org/r/20130809221420.618E6FAB03@dev.laptop.org Acked-by: Yinghai Lu <yinghai@kernel.org> Signed-off-by: H. Peter Anvin <hpa@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This should actually be returning an ERR_PTR on error instead of NULL.
That was how it was designed and all the callers expect it.
[AV: actually, that's what "VFS: Make clone_mnt()/copy_tree()/collect_mounts()
return errors" missed - originally collect_mounts() was expected to return
NULL on failure]
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Patch fixes zd1201 not to use stack as URB transfer_buffer. URB buffers need
to be DMA-able, which stack is not.
Patch is only compile tested.
Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi> Signed-off-by: John W. Linville <linville@tuxdriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
An older PVHVM guest (v3.0 based) crashed during vCPU hot-plug with:
kernel BUG at drivers/xen/events.c:1328!
RCU has detected that a CPU has not entered a quiescent state within the
grace period. It needs to send the CPU a reschedule IPI if it is not
offline. rcu_implicit_offline_qs() does this check:
/*
* If the CPU is offline, it is in a quiescent state. We can
* trust its state not to change because interrupts are disabled.
*/
if (cpu_is_offline(rdp->cpu)) {
rdp->offline_fqs++;
return 1;
}
Else the CPU is online. Send it a reschedule IPI.
The CPU is in the middle of being hot-plugged and has been marked online
(!cpu_is_offline()). See start_secondary():
start_secondary() then waits for the CPU bringing up the hot-plugged CPU to
mark it as active:
/*
* Wait until the cpu which brought this one up marked it
* online before enabling interrupts. If we don't do that then
* we can end up waking up the softirq thread before this cpu
* reached the active state, which makes the scheduler unhappy
* and schedule the softirq thread on the wrong cpu. This is
* only observable with forced threaded interrupts, but in
* theory it could also happen w/o them. It's just way harder
* to achieve.
*/
while (!cpumask_test_cpu(smp_processor_id(), cpu_active_mask))
cpu_relax();
/* enable local interrupts */
local_irq_enable();
The CPU being hot-plugged will be marked active after it has been fully
initialized by the CPU managing the hot-plug. In the Xen PVHVM case
xen_smp_intr_init() is called to set up the hot-plugged vCPU's
XEN_RESCHEDULE_VECTOR.
The hot-plugging CPU is marked online, not marked active and does not have
its IPI vectors set up. rcu_implicit_offline_qs() sees the hot-plugging
cpu is !cpu_is_offline() and tries to send it a reschedule IPI:
This will lead to:
clocksource_watchdog() uses cpu_online_mask to pick the next CPU to handle
a watchdog timer:
/*
* Cycle through CPUs to check if the CPUs stay synchronized
* to each other.
*/
next_cpu = cpumask_next(raw_smp_processor_id(), cpu_online_mask);
if (next_cpu >= nr_cpu_ids)
next_cpu = cpumask_first(cpu_online_mask);
watchdog_timer.expires += WATCHDOG_INTERVAL;
add_timer_on(&watchdog_timer, next_cpu);
This resulted in an attempt to send an IPI to a hot-plugging CPU that
had not initialized its reschedule vector. One option would be to make
the RCU code check to not check for CPU offline but for CPU active.
As becoming active is done after a CPU is online (in older kernels).
But Srivatsa pointed out that "the cpu_active vs cpu_online ordering has been
completely reworked - in the online path, cpu_active is set *before* cpu_online,
and also, in the cpu offline path, the cpu_active bit is reset in the CPU_DYING
notification instead of CPU_DOWN_PREPARE." Drilling in this the bring-up
path: "[brought up CPU].. send out a CPU_STARTING notification, and in response
to that, the scheduler sets the CPU in the cpu_active_mask. Again, this mask
is better left to the scheduler alone, since it has the intelligence to use it
judiciously."
The conclusion was that:
"
1. At the IPI sender side:
It is incorrect to send an IPI to an offline CPU (cpu not present in
the cpu_online_mask). There are numerous places where we check this
and warn/complain.
2. At the IPI receiver side:
It is incorrect to let the world know of our presence (by setting
ourselves in global bitmasks) until our initialization steps are complete
to such an extent that we can handle the consequences (such as
receiving interrupts without crashing the sender etc.)
" (from Srivatsa)
As the native code enables the interrupts at some point we need to be
able to service them. In other words a CPU must have valid IPI vectors
if it has been marked online.
It doesn't need to handle the IPI (interrupts may be disabled) but needs
to have valid IPI vectors because another CPU may find it in cpu_online_mask
and attempt to send it an IPI.
This patch will change the order of the Xen vCPU bring-up functions so that
Xen vectors have been set up before start_secondary() is called.
It also will not continue to bring up a Xen vCPU if xen_smp_intr_init() fails
to initialize it.
Orabug 13823853
Signed-off-by Chuck Anderson <chuck.anderson@oracle.com> Acked-by: Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Signed-off-by: Jonghwan Choi <jhbird.choi@samsung.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
It was finally narrowed down to unloading a module that was being traced.
It was actually more than that. When functions are being traced, there's
a table of all functions that have a ref count of the number of active
tracers attached to that function. When a function trace callback is
registered to a function, the function's record ref count is incremented.
When it is unregistered, the function's record ref count is decremented.
If an inconsistency is detected (ref count goes below zero) the above
warning is shown and the function tracing is permanently disabled until
reboot.
The ftrace callback ops holds a hash of functions that it filters on
(and/or filters off). If the hash is empty, the default means to filter
all functions (for the filter_hash) or to disable no functions (for the
notrace_hash).
When a module is unloaded, it frees the function records that represent
the module functions. These records exist on their own pages, that is
function records for one module will not exist on the same page as
function records for other modules or even the core kernel.
Now when a module unloads, the records that represents its functions are
freed. When the module is loaded again, the records are recreated with
a default ref count of zero (unless there's a callback that traces all
functions, then they will also be traced, and the ref count will be
incremented).
The problem is that if an ftrace callback hash includes functions of the
module being unloaded, those hash entries will not be removed. If the
module is reloaded in the same location, the hash entries still point
to the functions of the module but the module's ref counts do not reflect
that.
With the help of Steve and Joern, we found a reproducer:
Using uinput module and uinput_release function.
cd /sys/kernel/debug/tracing
modprobe uinput
echo uinput_release > set_ftrace_filter
echo function > current_tracer
rmmod uinput
modprobe uinput
# check /proc/modules to see if loaded in same addr, otherwise try again
echo nop > current_tracer
[BOOM]
The above loads the uinput module, which creates a table of functions that
can be traced within the module.
We add uinput_release to the filter_hash to trace just that function.
Enable function tracincg, which increments the ref count of the record
associated to uinput_release.
Remove uinput, which frees the records including the one that represents
uinput_release.
Load the uinput module again (and make sure it's at the same address).
This recreates the function records all with a ref count of zero,
including uinput_release.
Disable function tracing, which will decrement the ref count for uinput_release
which is now zero because of the module removal and reload, and we have
a mismatch (below zero ref count).
The solution is to check all currently tracing ftrace callbacks to see if any
are tracing any of the module's functions when a module is loaded (it already does
that with callbacks that trace all functions). If a callback happens to have
a module function being traced, it increments that records ref count and starts
tracing that function.
There may be a strange side effect with this, where tracing module functions
on unload and then reloading a new module may have that new module's functions
being traced. This may be something that confuses the user, but it's not
a big deal. Another approach is to disable all callback hashes on module unload,
but this leaves some ftrace callbacks that may not be registered, but can
still have hashes tracing the module's function where ftrace doesn't know about
it. That situation can cause the same bug. This solution solves that case too.
Another benefit of this solution, is it is possible to trace a module's
function on unload and load.
Link: http://lkml.kernel.org/r/20130705142629.GA325@redhat.com Reported-by: Jörn Engel <joern@logfs.org> Reported-by: Dave Jones <davej@redhat.com> Reported-by: Steve Hodgson <steve@purestorage.com> Tested-by: Steve Hodgson <steve@purestorage.com> Signed-off-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Uprobes suffer the same problem that kprobes have. There's a race between
writing to the "enable" file and removing the probe. The probe checks for
it being in use and if it is not, goes about deleting the probe and the
event that represents it. But the problem with that is, after it checks
if it is in use it can be enabled, and the deletion of the event (access
to the probe) will fail, as it is in use. But the uprobe will still be
deleted. This is a problem as the event can reference the uprobe that
was deleted.
The fix is to remove the event first, and check to make sure the event
removal succeeds. Then it is safe to remove the probe.
When the event exists, either ftrace or perf can enable the probe and
prevent the event from being removed.
When a probe is being removed, it cleans up the event files that correspond
to the probe. But there is a race between writing to one of these files
and deleting the probe. This is especially true for the "enable" file.
CPU 0 CPU 1
----- -----
fd = open("enable",O_WRONLY);
probes_open()
release_all_trace_probes()
unregister_trace_probe()
if (trace_probe_is_enabled(tp))
return -EBUSY
A test program was written that used two threads to simulate the
above scenario adding a nanosleep() interval to change the timings
and after several thousand runs, it was able to trigger this bug
and crash:
The unregister_trace_probe() must be done first, and if it fails it must
fail the removal of the kprobe.
Several changes have already been made by Oleg Nesterov and Masami Hiramatsu
to allow moving the unregister_probe_event() before the removal of
the probe and exit the function if it fails. This prevents the tp
structure from being used after it is freed.
Change trace_remove_event_call(call) to return the error if this
call is active. This is what the callers assume but can't verify
outside of the tracing locks. Both trace_kprobe.c/trace_uprobe.c
need the additional changes, unregister_trace_probe() should abort
if trace_remove_event_call() fails.
The caller is going to free this call/file so we must ensure that
nobody can use them after trace_remove_event_call() succeeds.
debugfs should be fine after the previous changes and event_remove()
does TRACE_REG_UNREGISTER, but still there are 2 reasons why we need
the additional checks:
- There could be a perf_event(s) attached to this tp_event, so the
patch checks ->perf_refcount.
- TRACE_REG_UNREGISTER can be suppressed by FTRACE_EVENT_FL_SOFT_MODE,
so we simply check FTRACE_EVENT_FL_ENABLED protected by event_mutex.
Change remove_event_file_dir() to clear ->i_private for every
file we are going to remove.
We need to check file->dir != NULL because event_create_dir()
can fail. debugfs_remove_recursive(NULL) is fine but the patch
moves it under the same check anyway for readability.
spin_lock(d_lock) and "d_inode != NULL" check are not needed
afaics, but I do not understand this code enough.
tracing_open_generic_file() and tracing_release_generic_file()
can go away, ftrace_enable_fops and ftrace_event_filter_fops()
use tracing_open_generic() but only to check tracing_disabled.
This fixes all races with event_remove() or instance_delete().
f_op->read/write/whatever can never use the freed file/call,
all event/* files were changed to check and use ->i_private
under event_mutex.
Note: this doesn't not fix other problems, event_remove() can
destroy the active ftrace_event_call, we need more changes but
those changes are completely orthogonal.
Preparation for the next patch. Extract the common code from
remove_event_from_tracers() and __trace_remove_event_dirs()
into the new helper, remove_event_file_dir().
The patch looks more complicated than it actually is, it also
moves remove_subsystem() up to avoid the forward declaration.
event_id_read() is racy, ftrace_event_call can be already freed
by trace_remove_event_call() callers.
Change event_create_dir() to pass "data = call->event.type", this
is all event_id_read() needs. ftrace_event_id_fops no longer needs
tracing_open_generic().
We add the new helper, event_file_data(), to read ->i_private, it
will have more users.
Note: currently ACCESS_ONCE() and "id != 0" check are not needed,
but we are going to change event_remove/rmdir to clear ->i_private.
If a ftrace ops is registered with the SAVE_REGS flag set, and there's
already a ops registered to one of its functions but without the
SAVE_REGS flag, there's a small race window where the SAVE_REGS ops gets
added to the list of callbacks to call for that function before the
callback trampoline gets set to save the regs.
The problem is, the function is not currently saving regs, which opens
a small race window where the ops that is expecting regs to be passed
to it, wont. This can cause a crash if the callback were to reference
the regs, as the SAVE_REGS guarantees that regs will be set.
To fix this, we add a check in the loop case where it checks if the ops
has the SAVE_REGS flag set, and if so, it will ignore it if regs is
not set.
Signed-off-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
tracing_open() and tracing_snapshot_open() are racy, the memory
inode->i_private points to can be already freed.
Convert these last users of "inode->i_private == trace_cpu" to
use "i_private = trace_array" and rely on tracing_get_cpu().
v2: incorporate the fix from Steven, tracing_release() must not
blindly dereference file->private_data unless we know that
the file was opened for reading.
But tc (and tr) can be already freed before trace_array_get() is called.
And it doesn't matter whether this file is per-cpu or it was created by
init_tracer_debugfs(), free_percpu() or kfree() are equally bad.
Note that even 1. is not safe, the freed memory can be unmapped. But even
if it was safe trace_array_get() can wrongly succeed if we also race with
the next new_instance_create() which can re-allocate the same tr, or tc
was overwritten and ->tr points to the valid tr. In this case 3. uses the
freed/reused memory.
Add the new trivial helper, trace_create_cpu_file() which simply calls
trace_create_file() and encodes "cpu" in "struct inode". Another helper,
tracing_get_cpu() will be used to read cpu_nr-or-RING_BUFFER_ALL_CPUS.
The patch abuses ->i_cdev to encode the number, it is never used unless
the file is S_ISCHR(). But we could use something else, say, i_bytes or
even ->d_fsdata. In any case this hack is hidden inside these 2 helpers,
it would be trivial to change them if needed.
This patch only changes tracing_init_debugfs_percpu() to use the new
trace_create_cpu_file(), the next patches will change file_operations.
Note: tracing_get_cpu(inode) is always safe but you can't trust the
result unless trace_array_get() was called, without trace_types_lock
which acts as a barrier it can wrongly return RING_BUFFER_ALL_CPUS.
Link: http://lkml.kernel.org/r/20130723152554.GA23710@redhat.com Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Oleg Nesterov <oleg@redhat.com> Signed-off-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Wait for disabling all running kprobe handlers when a kprobe
event is disabled, since the caller, trace_remove_event_call()
supposes that a removing event is disabled completely by
disabling the event.
With this change, ftrace can ensure that there is no running
event handlers after disabling it.
As the firmware API has changed significantly and we don't
have support code for the old APIs, bump the version to be
able to release the version 7 API firmware. Unfortunately
this means that the driver in 3.9 and 3.10 can't work, but
that's still better than crashing the device/driver there.
Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The MCAST queue should be enabled after DTIM only.
According to fw API, the MCAST must not be attached to any
station, but should appear in the mcast_qid of the AP's
mac context only.
In multicast, there is no retries nor RTS since there is no
specific recipient that can ACK or send CTS. This means
that we must not use the rate scale table for multicast
frames.
This true for any frame that doesn't have a valid
ieee80211_sta pointer.
The AP mode needs to use the MCAST fifo for the MCAST
frames sent after the DTIM. This fifo needs to be
configured with the same parameters as the VOICE FIFO.
A separate SCD queue is mapped to this fifo - the cab_queue
(cab stands for Content After Beacon). This queue isn't
connected to any station, but rather to the MAC context.
This queue should (and is already) be set as the MCAST
queue - this is part of the of MAC context command.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com> Reviewed-by: Ilan Peer <ilan.peer@intel.com> Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Without the new LLCP_CONNECTING state, non blocking sockets will be
woken up with a POLLHUP right after calling connect() because their
state is stuck at LLCP_CLOSED.
That prevents userspace from implementing any proper non blocking
socket based NFC p2p client.
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Due to a bug with RTC IMR, we cannot consider at91sam9x5 RTC compatible
with the previous one. Modify DT compatibility string, even if the driver
is not yet modified to take it into account.
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Vineet Gupta [Tue, 20 Aug 2013 08:08:10 +0000 (13:38 +0530)]
ARC: gdbserver breakage in Big-Endian configuration #1
[Based on mainline commit 502a0c775c7f0a: "ARC: pt_regs update #5"]
gdbserver needs @stop_pc, served by ptrace, but fetched from pt_regs
differently, based on in_brkpt_traps(), which in turn relies on
additional machine state in pt_regs->event bitfield.
unsigned long orig_r8:16, event:16;
For big endian config, this macro was returning false, despite being in
breakpoint Trap exception, causing wrong @stop_pc to be returned to gdb.
Issue #1: In BE, @event above is at offset 2 in word, while a STW insn
at offset 0 was used to update it. Resort to using ST insn
which updates the half-word at right location.
Issue #2: The union involving bitfields causes all the members to be
laid out at offset 0. So with fix #1 above, ASM was now
updating at offset 2, "C" code was still referencing at
offset 0. Fixed by wrapping bitfield in a struct.
In theory, under a given ACPI namespace node there should be only
one child device object with _ADR whose value matches a given bus
address exactly. In practice, however, there are systems in which
multiple child device objects under a given parent have _ADR matching
exactly the same address. In those cases we use _STA to determine
which of the multiple matching devices is enabled, since some systems
are known to indicate which ACPI device object to associate with the
given physical (usually PCI) device this way.
Unfortunately, as it turns out, there are systems in which many
device objects under the same parent have _ADR matching exactly the
same bus address and none of them has _STA, in which case they all
should be regarded as enabled according to the spec. Still, if
those device objects are supposed to represent bridges (e.g. this
is the case for device objects corresponding to PCIe ports), we can
try harder and skip the ones that have no child device objects in the
ACPI namespace. With luck, we can avoid using device objects that we
are not expected to use this way.
Although this only works for bridges whose children also have ACPI
namespace representation, it is sufficient to address graphics
adapter detection issues on some systems, so rework the code finding
a matching device ACPI handle for a given bus address to implement
this idea.
Introduce a new function, acpi_find_child(), taking three arguments:
the ACPI handle of the device's parent, a bus address suitable for
the device's bus type and a bool indicating if the device is a
bridge and make it work as outlined above. Reimplement the function
currently used for this purpose, acpi_get_child(), as a call to
acpi_find_child() with the last argument set to 'false' and make
the PCI subsystem use acpi_find_child() with the bridge information
passed as the last argument to it. [Lan Tianyu notices that it is
not sufficient to use pci_is_bridge() for that, because the device's
subordinate pointer hasn't been set yet at this point, so use
hdr_type instead.]
This change fixes a regression introduced inadvertently by commit 33f767d (ACPI: Rework acpi_get_child() to be more efficient) which
overlooked the fact that for acpi_walk_namespace() "post-order" means
"after all children have been visited" rather than "on the way back",
so for device objects without children and for namespace walks of
depth 1, as in the acpi_get_child() case, the "post-order" callbacks
ordering is actually the same as the ordering of "pre-order" ones.
Since that commit changed the namespace walk in acpi_get_child() to
terminate after finding the first matching object instead of going
through all of them and returning the last one, it effectively
changed the result returned by that function in some rare cases and
that led to problems (the switch from a "pre-order" to a "post-order"
callback was supposed to prevent that from happening, but it was
ineffective).
As it turns out, the systems where the change made by commit 33f767d actually matters are those where there are multiple ACPI
device objects representing the same PCIe port (which effectively
is a bridge). Moreover, only one of them, and the one we are
expected to use, has child device objects in the ACPI namespace,
so the regression can be addressed as described above.
References: https://bugzilla.kernel.org/show_bug.cgi?id=60561 Reported-by: Peter Wu <lekensteyn@gmail.com> Tested-by: Vladimir Lalov <mail@vlalov.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Acked-by: Bjorn Helgaas <bhelgaas@google.com> Cc: Peter Wu <lekensteyn@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Once do_acpi_find_child() has found the first matching handle, it
makes the acpi_get_child() loop stop and return that handle. On some
platforms, though, there are multiple devices with the same value of
"_ADR" in the same namespace scope, and if one of them is enabled,
the others will be disabled. For example:
If DEV0 and DEV1 are disabled and DEV2 is enabled, the handle of DEV2
should be returned, but actually the function always returns the
handle of DEV0.
To address that issue, make do_acpi_find_child() evaluate _STA to
check the device status. If a matching device object exists, but is
disabled, acpi_get_child() will continue to walk the namespace in the
hope of finding an enabled one. If one is found, its handle will be
returned, but otherwise the function will return the handle of the
disabled object found before (in case it is enabled going forward).
[rjw: Changelog] Signed-off-by: Jeff Wu <zlinuxkernel@gmail.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Cc: Peter Wu <lekensteyn@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
TX status notification can get lost, or the frames could
get stuck on the queue, so don't wait for the callback
from the driver forever and instead time out after half
a second.
Any uaccess between guest_enter and guest_exit could trigger a page fault,
the page fault handler would handle it as a guest fault and translate a
user address as guest address.
Signed-off-by: Dominik Dingel <dingel@linux.vnet.ibm.com> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Johannes Berg <johannes.berg@intel.com> Cc: Andrei Otcheretianski <andrei.otcheretianski@intel.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When jbd2_journal_dirty_metadata() returns error,
__ext4_handle_dirty_metadata() stops the handle. However callers of this
function do not count with that fact and still happily used now freed
handle. This use after free can result in various issues but very likely
we oops soon.
The motivation of adding __ext4_journal_stop() into
__ext4_handle_dirty_metadata() in commit 9ea7a0df seems to be only to
improve error reporting. So replace __ext4_journal_stop() with
ext4_journal_abort_handle() which was there before that commit and add
WARN_ON_ONCE() to dump stack to provide useful information.
Reported-by: Sage Weil <sage@inktank.com> Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In file included from fs/exec.c:61:0:
arch/s390/include/asm/tlb.h:35:23: error: expected identifier or '(' before 'unsigned'
arch/s390/include/asm/tlb.h:36:1: warning: no semicolon at end of struct or union [enabled by default]
arch/s390/include/asm/tlb.h: In function 'tlb_gather_mmu':
arch/s390/include/asm/tlb.h:57:5: error: 'struct mmu_gather' has no member named 'end'
Broken due to commit 2b047252d0 ("Fix TLB gather virtual address range
invalidation corner cases").
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Guenter Roeck <linux@roeck-us.net>
[ Oh well. We had build testing for ppc amd um, but no s390 - Linus ] Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
As pointed out by Andreas Schwab, pointers passed to ARAnyM NatFeat calls
should be physical addresses, not virtual addresses.
Fortunately on Atari, physical and virtual kernel addresses are the same,
as long as normal kernel memory is concerned, so this usually worked fine
without conversion.
But for modules, pointers to literal strings are located in vmalloc()ed
memory. Depending on the version of ARAnyM, this causes the nf_get_id()
call to just fail, or worse, crash ARAnyM itself with e.g.
Gotcha! Illegal memory access. Atari PC = $968c
This is a big issue for distro kernels, who want to have all drivers as
loadable modules in an initrd.
Add a wrapper for nf_get_id() that copies the literal to the stack to
work around this issue.
As diagnosed by Andreas, fs/btrfs/volumes.c:__btrfs_map_block()
calls
do_div(stripe_nr, stripe_len);
with stripe_len u64, while do_div() assumes the divisor is a 32-bit number.
Due to the lack of truncation in the m68k-specific implementation of
do_div(), the division is performed using the upper 32-bit word of
stripe_len, which is zero.
This was introduced by commit 53b381b3abeb86f12787a6c40fee9b2f71edc23b
("Btrfs: RAID5 and RAID6"), which changed the divisor from
map->stripe_len (struct map_lookup.stripe_len is int) to a 64-bit temporary.
It is possible to construct an event group with a software event as a
group leader and then subsequently add a hardware event to the group.
This results in the event group being validated by adding all members
of the group to a fake PMU and attempting to allocate each event on
their respective PMU.
Unfortunately, for software events wthout a corresponding arm_pmu, this
results in a kernel crash attempting to dereference the ->get_event_idx
function pointer.
This patch fixes the problem by checking explicitly for software events
and ignoring those in event validation (since they can always be
scheduled). We will probably want to revisit this for 3.12, since the
validation checks don't appear to work correctly when dealing with
multiple hardware PMUs anyway.
Reported-by: Vince Weaver <vincent.weaver@maine.edu> Tested-by: Vince Weaver <vincent.weaver@maine.edu> Tested-by: Mark Rutland <mark.rutland@arm.com> Signed-off-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
"Since v3.7.2 on two independent machines a very specific Git
repository fails in 9/10 cases on git-fsck due to an SHA1/memory
failures. This only occurs on a very specific repository and can be
reproduced stably on two independent laptops. Git mailing list ran
out of ideas and for me this looks like some very exotic kernel issue"
and bisected the failure to the backport of commit 53a59fc67f97 ("mm:
limit mmu_gather batching to fix soft lockups on !CONFIG_PREEMPT").
That commit itself is not actually buggy, but what it does is to make it
much more likely to hit the partial TLB invalidation case, since it
introduces a new case in tlb_next_batch() that previously only ever
happened when running out of memory.
The real bug is that the TLB gather virtual memory range setup is subtly
buggered. It was introduced in commit 597e1c3580b7 ("mm/mmu_gather:
enable tlb flush range in generic mmu_gather"), and the range handling
was already fixed at least once in commit e6c495a96ce0 ("mm: fix the TLB
range flushed when __tlb_remove_page() runs out of slots"), but that fix
was not complete.
The problem with the TLB gather virtual address range is that it isn't
set up by the initial tlb_gather_mmu() initialization (which didn't get
the TLB range information), but it is set up ad-hoc later by the
functions that actually flush the TLB. And so any such case that forgot
to update the TLB range entries would potentially miss TLB invalidates.
Rather than try to figure out exactly which particular ad-hoc range
setup was missing (I personally suspect it's the hugetlb case in
zap_huge_pmd(), which didn't have the same logic as zap_pte_range()
did), this patch just gets rid of the problem at the source: make the
TLB range information available to tlb_gather_mmu(), and initialize it
when initializing all the other tlb gather fields.
This makes the patch larger, but conceptually much simpler. And the end
result is much more understandable; even if you want to play games with
partial ranges when invalidating the TLB contents in chunks, now the
range information is always there, and anybody who doesn't want to
bother with it won't introduce subtle bugs.
Ben verified that this fixes his problem.
Reported-bisected-and-tested-by: Ben Tebulin <tebulin@googlemail.com> Build-testing-by: Stephen Rothwell <sfr@canb.auug.org.au> Build-testing-by: Richard Weinberger <richard.weinberger@gmail.com> Reviewed-by: Michal Hocko <mhocko@suse.cz> Acked-by: Peter Zijlstra <peterz@infradead.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch fixes a kernel panic that can occur when disconnecting a
wireless USB->serial device. When the serial device disconnects, the
device cleanup procedure ends up calling usb_hcd_disable_endpoint on the
serial device's endpoints. The wusbcore uses the ABORT_RPIPE command to
abort all transfers on the given endpoint but it does not properly give
back the URBs when the transfer results return from the HWA. This patch
prevents the transfer result processing code from bailing out when it sees
a WA_XFER_STATUS_ABORTED result code so that these urbs are flushed
properly by usb_hcd_disable_endpoint. It also updates wa_urb_dequeue to
handle the case where the endpoint has already been cleaned up when
usb_kill_urb is called which is where the panic originally occurred.
Signed-off-by: Thomas Pugliese <thomas.pugliese@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
pm_qos_update_request_timeout() updates a qos and then schedules
a delayed work item to bring the qos back down to the default
after the timeout. When the work item runs, pm_qos_work_fn() will
call pm_qos_update_request() and deadlock because it tries to
cancel itself via cancel_delayed_work_sync(). Future callers of
that qos will also hang waiting to cancel the work that is
canceling itself. Let's extract the little bit of code that does
the real work of pm_qos_update_request() and call it from the
work function so that we don't deadlock.
Before ed1ac6e (PM: don't use [delayed_]work_pending()) this didn't
happen because the work function wouldn't try to cancel itself.
[backport to 3.10 - gregkh]
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org> Reviewed-by: Tejun Heo <tj@kernel.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This fixes an issue where the bulk-in urb used for incoming data transfer
is not resubmitted if the packet recieved contains an error status. This
results in the driver locking until the port is closed and re-opened.
Tested on a custom board with a Cinterion GSM module.
Signed-off-by: Matt Burtch <matt@grid-net.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Since commits 4005ad4390bf (EHCI: implement new semantics for
URB_ISO_ASAP) and c75c5ab575af (ALSA: USB: adjust for changed 3.8 USB
API) became widely distributed, people have been experiencing problems
with audio transfers. The slightest underrun causes complete failure,
requiring the audio stream to be restarted.
It turns out that the current isochronous API doesn't handle underruns
in the best way. The ALSA developers would much rather have transfers
that are submitted too late be accepted and complete in the normal
fashion, rather than being refused outright.
This patch implements the requested approach. When an isochronous URB
submission is so late that all its scheduled slots have already
expired, a debugging message will be printed in the log and the URB
will be accepted as usual. Assuming it was submitted by a completion
handler (which is normally the case), it will complete shortly
thereafter with all the usb_iso_packet_descriptor status fields marked
-EXDEV.
Make sure to fail properly if the device is not accepted during attach
in order to avoid null-pointer derefs (of missing interface private
data) at disconnect or release.
Signed-off-by: Johan Hovold <jhovold@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The parallel-port code of the drivers used a stack allocated
control-request buffer for asynchronous (and possibly deferred) control
requests. This not only violates the no-DMA-from-stack requirement but
could also lead to corrupt control requests being submitted.
Signed-off-by: Johan Hovold <jhovold@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix bug in device-type detection on big-endian machines originally
introduced by commit 0eafe4de ("USB: serial: mos7840: add support for
MCS7810 devices") which always matched on little-endian product ids.
Reported-by: kbuild test robot <fengguang.wu@intel.com> Signed-off-by: Johan Hovold <jhovold@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix endianess bugs in firmware handling introduced by commits cb7a7c6a
("ti_usb_3410_5052: add Multi-Tech modem support") and 05a3d905
("ti_usb_3410_5052: support alternate firmware") which made the driver
use the wrong firmware for certain devices on big-endian machines.
Signed-off-by: Johan Hovold <jhovold@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This is similar to the race Linus had reported, but in this case
it's an older bug: nl80211_prepare_wdev_dump() uses the wiphy
index in cb->args[0] as it is and thus parses the message over
and over again instead of just once because 0 is the first valid
wiphy index. Similar code in nl80211_testmode_dump() correctly
offsets the wiphy_index by 1, do that here as well.
Reported-by: Ben Hutchings <ben@decadent.org.uk> Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We've added a fake mute control (setting the amp volume to zero) for
CX5051 at commit [3868137e: ALSA: hda - Add a fake mute feature], but
this feature was overlooked in the generic parser implementation. Now
the driver lacks of mute controls on these codecs.
The fix is just to check both AC_AMPCAP_MUTE and AC_AMPCAP_MIN_MUTE
bits in each place checking the amp capabilities.
projects / linux-stable / log