Even if we don't have extended SCA support, we can have more than 64 CPUs
if we don't enable any HW features that might use the SCA entries.
Now, this works just fine, but we missed a return, which is why we
would actually store the SCA entries. If we have more than 64 CPUs, this
means writing outside of the basic SCA - bad.
Let's fix this. This allows > 64 CPUs when running nested (under vSIE)
without random crashes.
Fixes: a6940674c384 ("KVM: s390: allow 255 VCPUs when sca entries aren't used") Reported-by: Christian Borntraeger <borntraeger@de.ibm.com> Tested-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20180306132758.21034-1-david@redhat.com> Cc: stable@vger.kernel.org Reviewed-by: Cornelia Huck <cohuck@redhat.com> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The vq->vq.num_free hasn't been changed when error happens,
so it shouldn't be changed when handling the error.
Fixes: 780bc7903a32 ("virtio_ring: Support DMA APIs") Cc: Andy Lutomirski <luto@kernel.org> Cc: Michael S. Tsirkin <mst@redhat.com> Cc: stable@vger.kernel.org Signed-off-by: Tiwei Bie <tiwei.bie@intel.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit aa4d86163e4e ("block: loop: switch to VFS ITER_BVEC")
replaced __do_lo_send_write(), which used ITER_KVEC iterators, with
lo_write_bvec() which uses ITER_BVEC iterators. In this change, though,
the WRITE flag was lost:
This flag is necessary for the DAX case because we make decisions based on
whether or not the iterator is a READ or a WRITE in dax_iomap_actor() and
in dax_iomap_rw().
We end up going through this path in configurations where we combine a PMEM
device with 4k sectors, a loopback device and DAX. The consequence of this
missed flag is that what we intend as a write actually turns into a read in
the DAX code, so no data is ever written.
The very simplest test case is to create a loopback device and try and
write a small string to it, then hexdump a few bytes of the device to see
if the write took. Without this patch you read back all zeros, with this
you read back the string you wrote.
For XFS this causes us to fail or panic during the following xfstests:
xfs/074 xfs/078 xfs/216 xfs/217 xfs/250
For ext4 we have a similar issue where writes never happen, but we don't
currently have any xfstests that use loopback and show this issue.
Fix this by restoring the WRITE flag argument to iov_iter_bvec(). This
causes the xfstests to all pass.
Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: stable@vger.kernel.org Fixes: commit aa4d86163e4e ("block: loop: switch to VFS ITER_BVEC") Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Ming Lei <ming.lei@redhat.com> Signed-off-by: Ross Zwisler <ross.zwisler@linux.intel.com> Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Sphinx 1.7 removed sphinx.util.compat.Directive so people
who have upgraded cannot build the documentation. Switch to
docutils.parsers.rst.Directive which has been available since
docutils 0.5 released in 2009.
Bugzilla: https://bugzilla.opensuse.org/show_bug.cgi?id=1083694 Co-developed-by: Takashi Iwai <tiwai@suse.de> Acked-by: Jani Nikula <jani.nikula@intel.com> Cc: stable@vger.kernel.org Signed-off-by: Matthew Wilcox <mawilcox@microsoft.com> Signed-off-by: Jonathan Corbet <corbet@lwn.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This is just a cleanup. It aids handling the special end case in the
next commit.
[akpm@linux-foundation.org: make it work against current -linus, not against -mm]
[akpm@linux-foundation.org: make it work against current -linus, not against -mm some more] Link: http://lkml.kernel.org/r/1ca478d4269125a99bcfb1ca04d7b88ac1aee924.1520011944.git.neelx@redhat.com Signed-off-by: Daniel Vacek <neelx@redhat.com> Cc: Michal Hocko <mhocko@suse.com> Cc: Vlastimil Babka <vbabka@suse.cz> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Pavel Tatashin <pasha.tatashin@oracle.com> Cc: Paul Burton <paul.burton@imgtec.com> Cc: <stable@vger.kernel.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>
Commit b8347c219649 ("x86/debug: Handle warnings before the notifier
chain, to fix KGDB crash") changed the ordering of fixups, and did not
take into account the case of x86 processing non-WARN() and non-BUG()
exceptions. This would lead to output of a false BUG line with no other
information.
In the case of a refcount exception, it would be immediately followed by
the refcount WARN(), producing very strange double-"cut here":
lkdtm: attempting bad refcount_inc() overflow
------------[ cut here ]------------
Kernel BUG at 0000000065f29de5 [verbose debug info unavailable]
------------[ cut here ]------------
refcount_t overflow at lkdtm_REFCOUNT_INC_OVERFLOW+0x6b/0x90 in cat[3065], uid/euid: 0/0
WARNING: CPU: 0 PID: 3065 at kernel/panic.c:657 refcount_error_report+0x9a/0xa4
...
In the prior ordering, exceptions were searched first:
do_trap_no_signal(struct task_struct *tsk, int trapnr, char *str,
...
if (fixup_exception(regs, trapnr))
return 0;
- if (fixup_bug(regs, trapnr))
- return 0;
-
As a result, fixup_bugs()'s is_valid_bugaddr() didn't take into account
needing to search the exception list first, since that had already
happened.
So, instead of searching the exception list twice (once in
is_valid_bugaddr() and then again in fixup_exception()), just add a
simple sanity check to report_bug() that will immediately bail out if a
BUG() (or WARN()) entry is not found.
Link: http://lkml.kernel.org/r/20180301225934.GA34350@beast Fixes: b8347c219649 ("x86/debug: Handle warnings before the notifier chain, to fix KGDB crash") Signed-off-by: Kees Cook <keescook@chromium.org> Cc: Ingo Molnar <mingo@kernel.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Borislav Petkov <bp@alien8.de> Cc: Richard Weinberger <richard.weinberger@gmail.com> Cc: <stable@vger.kernel.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>
If matrix_keypad_stop() is executing and the keypad interrupt is triggered,
disable_row_irqs() may be called by both matrix_keypad_interrupt() and
matrix_keypad_stop() at the same time, causing interrupts to be disabled
twice and the keypad being "stuck" after resuming.
Take lock when setting keypad->stopped to ensure that ISR will not race
with matrix_keypad_stop() disabling interrupts.
The subordinate value indicates the highest bus number which can be
reached downstream though a certain device.
Commit a20c7f36bd3d ("PCI: Do not allocate more buses than available in
parent") ensures that downstream devices cannot assign busnumbers higher
than the upstream device subordinate number, which was indeed illogical.
By default, dw_pcie_setup_rc() inits the Root Complex subordinate to a
value of 0x01.
Due to this combined with above commit, enumeration stops digging deeper
downstream as soon as bus num 0x01 has been assigned, which is always the
case for a bridge device.
This results in all devices behind a bridge bus remaining undetected, as
these would be connected to bus 0x02 or higher.
Fix this by initializing the RC to a subordinate value of 0xff, which is
not altering hardware behaviour in any way, but informs probing function
pci_scan_bridge() later on which reads this value back from register.
The following nasty errors during boot are also fixed by this:
pci_bus 0000:02: busn_res: can not insert [bus 02-ff] under [bus 01] (conflicts with (null) [bus 01])
...
pci_bus 0000:03: [bus 03] partially hidden behind bridge 0000:01 [bus 01]
...
pci_bus 0000:04: [bus 04] partially hidden behind bridge 0000:01 [bus 01]
...
pci_bus 0000:05: [bus 05] partially hidden behind bridge 0000:01 [bus 01]
pci_bus 0000:02: busn_res: [bus 02-ff] end is updated to 05
pci_bus 0000:02: busn_res: can not insert [bus 02-05] under [bus 01] (conflicts with (null) [bus 01])
pci_bus 0000:02: [bus 02-05] partially hidden behind bridge 0000:01 [bus 01]
Fixes: a20c7f36bd3d ("PCI: Do not allocate more buses than available in
parent") Tested-by: Niklas Cassel <niklas.cassel@axis.com> Tested-by: Fabio Estevam <fabio.estevam@nxp.com> Tested-by: Sebastian Reichel <sebastian.reichel@collabora.co.uk> Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com> Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> Signed-off-by: Bjorn Helgaas <bhelgaas@google.com> Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com> Acked-by: Lucas Stach <l.stach@pengutronix.de> Cc: stable@vger.kernel.org # v4.15+ Cc: Binghui Wang <wangbinghui@hisilicon.com> Cc: Jesper Nilsson <jesper.nilsson@axis.com> Cc: Jianguo Sun <sunjianguo1@huawei.com> Cc: Jingoo Han <jingoohan1@gmail.com> Cc: Kishon Vijay Abraham I <kishon@ti.com> Cc: Lucas Stach <l.stach@pengutronix.de> Cc: Mika Westerberg <mika.westerberg@linux.intel.com> Cc: Minghuan Lian <minghuan.Lian@freescale.com> Cc: Mingkai Hu <mingkai.hu@freescale.com> Cc: Murali Karicheri <m-karicheri2@ti.com> Cc: Pratyush Anand <pratyush.anand@gmail.com> Cc: Richard Zhu <hongxing.zhu@nxp.com> Cc: Roy Zang <tie-fei.zang@freescale.com> Cc: Shawn Guo <shawn.guo@linaro.org> Cc: Stanimir Varbanov <svarbanov@mm-sol.com> Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> Cc: Xiaowei Song <songxiaowei@hisilicon.com> Cc: Zhou Wang <wangzhou1@hisilicon.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The allocation of host_data is not null checked, leading to a null
pointer dereference if the allocation fails. Fix this by adding a null
check and return with -ENOMEM.
Fixes: 64b139f97c01 ("MIPS: OCTEON: irq: add CIB and other fixes") Signed-off-by: Colin Ian King <colin.king@canonical.com> Acked-by: David Daney <david.daney@cavium.com> Cc: Ralf Baechle <ralf@linux-mips.org> Cc: "Steven J. Hill" <Steven.Hill@cavium.com> Cc: linux-mips@linux-mips.org Cc: <stable@vger.kernel.org> # 4.0+
Patchwork: https://patchwork.linux-mips.org/patch/18658/ Signed-off-by: James Hogan <jhogan@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currently there is no null check on a failed allocation of board_data,
and hence a null pointer dereference will occurr. Fix this by checking
for the out of memory null pointer.
Fixes: a7473717483e ("MIPS: ath25: add board configuration detection") Signed-off-by: Colin Ian King <colin.king@canonical.com> Cc: Ralf Baechle <ralf@linux-mips.org> Cc: linux-mips@linux-mips.org Cc: <stable@vger.kernel.org> # 3.19+
Patchwork: https://patchwork.linux-mips.org/patch/18657/ Signed-off-by: James Hogan <jhogan@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit a3e6c1eff548 ("MIPS: IRQ: Fix disable_irq on CPU IRQs") fixes an
issue where disable_irq did not actually disable the irq. The bug caused
our IPIs to not be disabled, which actually is the correct behavior.
With the addition of commit a3e6c1eff548 ("MIPS: IRQ: Fix disable_irq on
CPU IRQs"), the IPIs were getting disabled going into suspend, thus
schedule_ipi() was not being called. This caused deadlocks where
schedulable task were not being scheduled and other cpus were waiting
for them to do something.
Add the IRQF_NO_SUSPEND flag so an irq_disable will not be called on the
IPIs during suspend.
Signed-off-by: Justin Chen <justinpopo6@gmail.com> Fixes: a3e6c1eff548 ("MIPS: IRQ: Fix disabled_irq on CPU IRQs") Cc: Florian Fainelli <f.fainelli@gmail.com> Cc: linux-mips@linux-mips.org Cc: stable@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/17385/
[jhogan@kernel.org: checkpatch: wrap long lines and fix commit refs] Signed-off-by: James Hogan <jhogan@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When UVD is in VM mode, there is not uvd handle exchanged,
uvd.handles are always 0. So vcpu_bo always need save,
Otherwise amdgpu driver will fail during suspend/resume.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105021 Signed-off-by: James Zhu <James.Zhu@amd.com> Reviewed-by: Leo Liu <leo.liu@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Max uvd handles should use adev->uvd.max_handles instead of
AMDGPU_MAX_UVD_HANDLES here.
Signed-off-by: James Zhu <James.Zhu@amd.com> Reviewed-by: Leo Liu <leo.liu@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
it is required if a platform supports PCIe root complex
core voltage reduction. After receiving this notification,
SBIOS can apply default PCIe root complex power policy.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Rex Zhu <Rex.Zhu@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Rather than querying it every time we need it.
Also fixes a crash in VM pass through if there is no
root bridge because the cached value fetch already checks
this properly.
v2: fix includes
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=105244 Acked-by: Christian König <christian.koenig@amd.com> Reviewed-by: Rex Zhu<rezhu@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
power containment disabled only on Fiji and compute
power profile. It violates PCIe spec and may cause power
supply failed. Enabling it will fix the issue, even the
fix will drop performance of some compute tests.
Signed-off-by: Eric Huang <JinHuiEric.Huang@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In radeon_device_init, set the need_dma32 flag for Cedar chips
(e.g. FirePro 2270). This fixes, or at least works around, a bug
on PowerPC exposed by last year's commits
Nouveau only exposes support for XBGR2101010. Prior to the atomic
conversion, drm would pass in the wrong format in the framebuffer, but
it was always ignored -- both userspace (xf86-video-nouveau) and the
kernel driver agreed on the layout, so the fact that the format was
wrong didn't matter.
With the atomic conversion, nouveau all of a sudden started caring about
the exact format, and so the previously-working code in
xf86-video-nouveau no longer functioned since the (internally-assigned)
format from the addfb ioctl was wrong.
This change adds infrastructure to allow a drm driver to specify that it
prefers the XBGR format variant for the addfb ioctl, and makes nouveau's
nv50 display driver set it. (Prior gens had no support for 30bpp at all.)
amdgpu's ->runtime_suspend hook calls drm_kms_helper_poll_disable(),
which waits for the output poll worker to finish if it's running.
The output poll worker meanwhile calls pm_runtime_get_sync() in
amdgpu's ->detect hooks, which waits for the ongoing suspend to finish,
causing a deadlock.
Fix by not acquiring a runtime PM ref if the ->detect hooks are called
in the output poll worker's context. This is safe because the poll
worker is only enabled while runtime active and we know that
->runtime_suspend waits for it to finish.
Fixes: d38ceaf99ed0 ("drm/amdgpu: add core driver (v4)") Cc: stable@vger.kernel.org # v4.2+: 27d4ee03078a: workqueue: Allow retrieval of current task's work struct Cc: stable@vger.kernel.org # v4.2+: 25c058ccaf2e: drm: Allow determining if current task is output poll worker Cc: Alex Deucher <alexander.deucher@amd.com> Tested-by: Mike Lothian <mike@fireburn.co.uk> Reviewed-by: Lyude Paul <lyude@redhat.com> Signed-off-by: Lukas Wunner <lukas@wunner.de> Link: https://patchwork.freedesktop.org/patch/msgid/4c9bf72aacae1eef062bd134cd112e0770a7f121.1518338789.git.lukas@wunner.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
radeon's ->runtime_suspend hook calls drm_kms_helper_poll_disable(),
which waits for the output poll worker to finish if it's running.
The output poll worker meanwhile calls pm_runtime_get_sync() in
radeon's ->detect hooks, which waits for the ongoing suspend to finish,
causing a deadlock.
Fix by not acquiring a runtime PM ref if the ->detect hooks are called
in the output poll worker's context. This is safe because the poll
worker is only enabled while runtime active and we know that
->runtime_suspend waits for it to finish.
Stack trace for posterity:
INFO: task kworker/0:3:31847 blocked for more than 120 seconds
Workqueue: events output_poll_execute [drm_kms_helper]
Call Trace:
schedule+0x3c/0x90
rpm_resume+0x1e2/0x690
__pm_runtime_resume+0x3f/0x60
radeon_lvds_detect+0x39/0xf0 [radeon]
output_poll_execute+0xda/0x1e0 [drm_kms_helper]
process_one_work+0x14b/0x440
worker_thread+0x48/0x4a0
nouveau's ->runtime_suspend hook calls drm_kms_helper_poll_disable(),
which waits for the output poll worker to finish if it's running.
The output poll worker meanwhile calls pm_runtime_get_sync() in
nouveau_connector_detect() which waits for the ongoing suspend to finish,
causing a deadlock.
Fix by not acquiring a runtime PM ref if nouveau_connector_detect() is
called in the output poll worker's context. This is safe because
the poll worker is only enabled while runtime active and we know that
->runtime_suspend waits for it to finish.
Other contexts calling nouveau_connector_detect() do require a runtime
PM ref, these comprise:
Introduce a helper to determine if the current task is an output poll
worker.
This allows us to fix a long-standing deadlock in several DRM drivers
wherein the ->runtime_suspend callback waits for the output poll worker
to finish and the worker in turn calls a ->detect callback which waits
for runtime suspend to finish. The ->detect callback is invoked from
multiple call sites and waiting for runtime suspend to finish is the
correct thing to do except if it's executing in the context of the
worker.
v2: Expand kerneldoc to specifically mention deadlock between
output poll worker and autosuspend worker as use case. (Lyude)
Introduce a helper to retrieve the current task's work struct if it is
a workqueue worker.
This allows us to fix a long-standing deadlock in several DRM drivers
wherein the ->runtime_suspend callback waits for a specific worker to
finish and that worker in turn calls a function which waits for runtime
suspend to finish. That function is invoked from multiple call sites
and waiting for runtime suspend to finish is the correct thing to do
except if it's executing in the context of the worker.
Cc: Lai Jiangshan <jiangshanlai@gmail.com> Cc: Dave Airlie <airlied@redhat.com> Cc: Ben Skeggs <bskeggs@redhat.com> Cc: Alex Deucher <alexander.deucher@amd.com> Acked-by: Tejun Heo <tj@kernel.org> Reviewed-by: Lyude Paul <lyude@redhat.com> Signed-off-by: Lukas Wunner <lukas@wunner.de> Link: https://patchwork.freedesktop.org/patch/msgid/2d8f603074131eb87e588d2b803a71765bd3a2fd.1518338788.git.lukas@wunner.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
intel_power_domains_init_hw() calls set_init_power, but when using
runtime power management this call is skipped. This prevents hw readout
from taking place.
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104172 Link: https://patchwork.freedesktop.org/patch/msgid/20180116155324.75120-1-maarten.lankhorst@linux.intel.com Fixes: bc87229f323e ("drm/i915/skl: enable PC9/10 power states during suspend-to-idle") Cc: Nivedita Swaminathan <nivedita.swaminathan@intel.com> Cc: Imre Deak <imre.deak@intel.com> Cc: Patrik Jakobsson <patrik.jakobsson@linux.intel.com> Cc: Jani Nikula <jani.nikula@linux.intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: <stable@vger.kernel.org> # v4.5+ Reviewed-by: Imre Deak <imre.deak@intel.com>
(cherry picked from commit ac25dfed15d470d7f23dd817e965b54aa3f94a1e) Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Looking at the assembly of get_next_timer_interrupt(), address came
from %r8 (ffff95e1f6451188) which is pointing to list_head with single
entry at ffff95e5ff621178.
Avoid that the recently introduced call_rcu() call in the SCSI core
triggers a double call_rcu() call.
Reported-by: Natanael Copa <ncopa@alpinelinux.org> Reported-by: Damien Le Moal <damien.lemoal@wdc.com>
References: https://bugzilla.kernel.org/show_bug.cgi?id=198861 Fixes: 3bd6f43f5cb3 ("scsi: core: Ensure that the SCSI error handler gets woken up") Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com> Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com> Tested-by: Damien Le Moal <damien.lemoal@wdc.com> Cc: Natanael Copa <ncopa@alpinelinux.org> Cc: Damien Le Moal <damien.lemoal@wdc.com> Cc: Alexandre Oliva <oliva@gnu.org> Cc: Pavel Tikhomirov <ptikhomirov@virtuozzo.com> Cc: Hannes Reinecke <hare@suse.com> Cc: Johannes Thumshirn <jthumshirn@suse.de> Cc: <stable@vger.kernel.org> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We're seeing on CI that some contexts don't have the programmed OA
period timer that directs the OA unit on how often to write reports.
The issue is that we're not holding the drm lock from when we edit the
context images down to when we set the exclusive_stream variable. This
leaves a window for the deferred context allocation to call
i915_oa_init_reg_state() that will not program the expected OA timer
value, because we haven't set the exclusive_stream yet.
v2: Drop need_lock from gen8_configure_all_contexts() (Matt)
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Reviewed-by: Matthew Auld <matthew.auld@intel.com> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> Fixes: 701f8231a2f ("drm/i915/perf: prune OA configs")
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=102254
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=103715
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=103755 Link: https://patchwork.freedesktop.org/patch/msgid/20180301110613.1737-1-lionel.g.landwerlin@intel.com Cc: Jani Nikula <jani.nikula@linux.intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com> Cc: intel-gfx@lists.freedesktop.org Cc: <stable@vger.kernel.org> # v4.14+
(cherry picked from commit 41d3fdcd15d5ecf29cc73e8b79c2327ebb54b960) Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The ACK/NACK implementation as found in e.g. the G965 has the falling
clock edge and the release of the data line after the ACK for the received
byte happen at the same time.
This is conformant with the I2C specification, which allows a zero hold
time, see footnote [3]: "A device must internally provide a hold time of
at least 300 ns for the SDA signal (with respect to the V IH(min) of the
SCL signal) to bridge the undefined region of the falling edge of SCL."
Some HDMI-to-VGA converters apparently fail to adhere to this requirement
and latch SDA at the falling clock edge, so instead of an ACK
sometimes a NACK is read and the slave (i.e. the EDID ROM) ends the
transfer.
The bitbanging releases the data line for the ACK only 1/4 bit time after
the falling clock edge, so a slave will see the correct value no matter
if it samples at the rising or the falling clock edge or in the center.
Fallback to bitbanging is already done for the CRT connector.
We no longer use intel_crtc->wm.active for watermarks any more,
which was incorrect. But this uncovered a bug in sanitize_watermarks(),
which meant that we wrote the correct watermarks, but the next
update would still use the wrong hw watermarks for calculating.
This caused all further updates to fail with -EINVAL and the
log would reveal an error like the one below:
[ 10.043902] [drm:ilk_validate_wm_level.part.8 [i915]] Sprite WM0 too large 56 (max 0)
[ 10.043960] [drm:ilk_validate_pipe_wm [i915]] LP0 watermark invalid
[ 10.044030] [drm:intel_crtc_atomic_check [i915]] No valid intermediate pipe watermarks are possible
If we fail to unbind the vma (due to a signal on an active buffer that
needs to be moved for the next execbuf), then we need to clear the
persistent tracking state we setup for this execbuf.
Fixes: c7c6e46f913b ("drm/i915: Convert execbuf to use struct-of-array packing for critical fields")
Testcase: igt/gem_fenced_exec_thrash/no-spare-fences-busy* Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> Cc: <stable@vger.kernel.org> # v4.14+ Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20180219140144.24004-1-chris@chris-wilson.co.uk
(cherry picked from commit ed2f3532321083cf40e4da4e36234880e0136136) Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
stm32_vrefbuf_enable() wrongly checks VRR bit: 0 stands for not ready,
1 for ready. It currently checks the opposite.
This makes enable routine to exit immediately without waiting for ready
flag.
Fixes: 0cdbf481e927 ("regulator: Add support for stm32-vrefbuf") Signed-off-by: Fabrice Gasnier <fabrice.gasnier@st.com> Signed-off-by: Mark Brown <broonie@kernel.org> Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
when sock_create_kern(..., a) returns an error, 'a' might not be a valid
pointer, so it shouldn't be dereferenced to read a->sk->sk_sndbuf and
and a->sk->sk_rcvbuf; not doing that caused the following crash:
This is really being used as an assert that the expected usecnt
is being held and implicitly that the usecnt is valid. Rename it to
assert_uverbs_usecnt and tighten the checks to only accept valid
values of usecnt (eg 0 and < -1 are invalid).
The tigher checkes make the assertion cover more cases and is more
likely to find bugs via syzkaller/etc.
Fixes: 3832125624b7 ("IB/core: Add support for idr types") Signed-off-by: Leon Romanovsky <leon@kernel.org> Signed-off-by: Jason Gunthorpe <jgg@mellanox.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Users of ucma are supposed to provide size of option level,
in most paths it is supposed to be equal to u8 or u16, but
it is not the case for the IB path record, where it can be
multiple of struct ib_path_rec_data.
This patch takes simplest possible approach and prevents providing
values more than possible to allocate.
Reported-by: syzbot+a38b0e9f694c379ca7ce@syzkaller.appspotmail.com Fixes: 7ce86409adcd ("RDMA/ucma: Allow user space to set service type") Signed-off-by: Leon Romanovsky <leonro@mellanox.com> Signed-off-by: Doug Ledford <dledford@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We do want to respect the FLUSH_SYNC argument to nfs_commit_inode() to
ensure that all outstanding COMMIT requests to the inode in question are
complete. Currently we may exit early from both nfs_commit_inode() and
nfs_write_inode() even if there are COMMIT requests in flight, or unstable
writes on the commit list.
In order to get the right semantics w.r.t. sync_inode(), we don't need
to have nfs_commit_inode() reset the inode dirty flags when called from
nfs_wb_page() and/or nfs_wb_all(). We just need to ensure that
nfs_write_inode() leaves them in the right state if there are outstanding
commits, or stable pages.
Reported-by: Scott Mayhew <smayhew@redhat.com> Fixes: dc4fd9ab01ab ("nfs: don't wait on commit in nfs_commit_inode()...") Cc: stable@vger.kernel.org # v4.14+ Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Ensure that we hold a reference to the layout header when processing
the pNFS return-on-close so that the refcount value does not inadvertently
go to zero.
When driver is loaded in Target/Dual mode, it creates QPair to support
MQ and allocates resources for each QPair. This Qpair initialization is
delayed until the FW personality is changed to Dual/Target mode by
issuing chip reset. At the time of chip reset firmware is re-initilized
in correct personality all the QPairs are initialized by sending
MBC_INITIALIZE_MULTIQ (001Fh).
This patch fixes memory leak by adding check to issue
MBC_INITIALIZE_MULTIQ command only while deleting rsp/req queue when the
flag is set for initiator mode, and clean up QPair resources correctly
during the driver unload. This MBX does not need to be issued for
Target/Dual mode because chip reset will reset ISP.
Fixes: d65237c7f0860 ("scsi: qla2xxx: Fix mailbox failure while deleting Queue pairs") Cc: <stable@vger.kernel.org> # 4.10+ Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com> Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Current code manually allocate an fcport structure that is not properly
initialize. Replace kzalloc with qla2x00_alloc_fcport, so that all
fields are initialized. Also set set scan flag to port found
Cc: <stable@vger.kernel.org> Signed-off-by: Quinn Tran <quinn.tran@cavium.com> Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com> Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Original code acquires hardware_lock to add Abort IOCB onto driver
request queue for processing. However, abort_command() will also acquire
hardware lock to look up sp pointer before issuing abort IOCB command
resulting into a deadlock. This patch safely removes the possible
deadlock scenario by removing extra spinlock.
Fixes: 6eb54715b54bb ("qla2xxx: Added interface to send explicit LOGO.") Cc: <stable@vger.kernel.org> # 4.10+ Signed-off-by: Quinn Tran <quinn.tran@cavium.com> Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com> Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Get Port Database MBX cmd is to validate current Login state upon PRLI
completion. Current code looks at the last login state for re-validation
which was incorrect. This patch removed incorrect state check.
Fixes: 15f30a5752287 ("qla2xxx: Use IOCB interface to submit non-critical MBX.") Cc: <stable@vger.kernel.org> # 4.10+ Signed-off-by: Quinn Tran <quinn.tran@cavium.com> Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com> Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Current driver design schedules relogin process via DPC thread every 1
second. In a large fabric, this DPC thread tries to schedule too many
jobs and might get overloaded. As a result of this processing of DPC
thread, it can schedule relogin earlier than 1 second.
Fixes: 726b85487067d ("qla2xxx: Add framework for async fabric discovery") Cc: <stable@vger.kernel.org> # 4.10+ Signed-off-by: Quinn Tran <quinn.tran@cavium.com> Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com> Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When NPort Handle is in use, driver needs to mark the handle as used and
pick another. Instead, the code clears the handle and re-pick the same
handle.
Fixes: 726b85487067d ("qla2xxx: Add framework for async fabric discovery") Cc: <stable@vger.kernel.org> # 4.10+ Signed-off-by: Quinn Tran <quinn.tran@cavium.com> Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix co-existence between Block MQ and Target Mode. Block MQ and
initiator mode requires midlayer queue mapping to check for IRQ to be
affinitized. For target mode, it's not the case.
Fixes: 09620eeb62c41 ("scsi: qla2xxx: Add debug knob for user control workload") Cc: <stable@vger.kernel.org> # 4.12+ Signed-off-by: Quinn Tran <quinn.tran@cavium.com> Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com> Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 5e572cab92f0 ("tpm: Enable CLKRUN protocol for Braswell systems")
added logic in the TPM TIS driver to disable the Low Pin Count CLKRUN
signal during TPM transactions.
Unfortunately this breaks other devices that are attached to the LPC bus
like for example PS/2 mouse and keyboards.
One flaw with the logic is that it assumes that the CLKRUN is always
enabled, and so it unconditionally enables it after a TPM transaction.
But it could be that the CLKRUN# signal was already disabled in the LPC
bus and so after the driver probes, CLKRUN_EN will remain enabled which
may break other devices that are attached to the LPC bus but don't have
support for the CLKRUN protocol.
Fixes: 5e572cab92f0 ("tpm: Enable CLKRUN protocol for Braswell systems") Signed-off-by: Javier Martinez Canillas <javierm@redhat.com> Tested-by: James Ettle <james@ettle.org.uk> Tested-by: Jeffery Miller <jmiller@neverware.com> Reviewed-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> Tested-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The CLKRUN fix caused a few harmless compile-time warnings:
drivers/char/tpm/tpm_tis.c: In function 'tpm_tis_pnp_remove':
drivers/char/tpm/tpm_tis.c:274:23: error: unused variable 'priv' [-Werror=unused-variable]
drivers/char/tpm/tpm_tis.c: In function 'tpm_tis_plat_remove':
drivers/char/tpm/tpm_tis.c:324:23: error: unused variable 'priv' [-Werror=unused-variable]
This removes the variables that have now become unused.
Fixes: 6d0866cbc2d3 ("tpm: Keep CLKRUN enabled throughout the duration of transmit_cmd()") Signed-off-by: Arnd Bergmann <arnd@arndb.de> Reviewed-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> Reviewed-by: James Morris <jmorris@namei.org> Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This flag is only used to warn if CLKRUN_EN wasn't disabled on Braswell
systems, but the only way this can happen is if the code is not correct.
So it's an unnecessary check that just makes the code harder to read.
Suggested-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> Signed-off-by: Javier Martinez Canillas <javierm@redhat.com> Reviewed-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> Tested-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 5e572cab92f0bb5 ("tpm: Enable CLKRUN protocol for Braswell
systems") disabled CLKRUN protocol during TPM transactions and re-enabled
once the transaction is completed. But there were still some corner cases
observed where, reading of TPM header failed for savestate command
while going to suspend, which resulted in suspend failure.
To fix this issue keep the CLKRUN protocol disabled for the entire
duration of a single TPM command and not disabling and re-enabling
again for every TPM transaction. For the other TPM accesses outside
TPM command flow, add a higher level of disabling and re-enabling
the CLKRUN protocol, instead of doing for every TPM transaction.
Fixes: 5e572cab92f0bb5 ("tpm: Enable CLKRUN protocol for Braswell systems") Signed-off-by: Azhar Shaikh <azhar.shaikh@intel.com> Reviewed-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> Tested-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
For some reason, Florian forgot to apply to ip6_route_me_harder
the fix that went in commit 29e09229d9f2 ("netfilter: use
skb_to_full_sk in ip_route_me_harder")
Fixes: ca6fb0651883 ("tcp: attach SYNACK messages to request sockets instead of listener") Signed-off-by: Eric Dumazet <edumazet@google.com> Reported-by: syzbot <syzkaller@googlegroups.com> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
ebt_among is special, it has a dynamic match size and is exempt
from the central size checks.
Therefore it must check that the size of the match structure
provided from userspace is sane by making sure em->match_size
is at least the minimum size of the expected structure.
The module has such a check, but its only done after accessing
a structure that might be out of bounds.
First issue is that INIT_WORK() should be done before mod_timer()
or we risk timer being fired too soon, even with a 1 second timer.
Second issue is that we need to reject too big info->timeout
to avoid overflows in msecs_to_jiffies(info->timeout * 1000), or
risk looping, if result after overflow is 0.
[1]
WARNING: CPU: 1 PID: 5129 at kernel/workqueue.c:1444 __queue_work+0xdf4/0x1230 kernel/workqueue.c:1444
Kernel panic - not syncing: panic_on_warn set ...
The problem is that currently we don't have any check on the
configured port range. A port range == -1 triggers the bug, while
other negative values may require a very long time to complete the
following loop.
This commit addresses the issue swapping the two ends on negative
ranges. The check is performed in nf_nat_l4proto_unique_tuple() since
the nft nat loads the port values from nft registers at runtime.
v1 -> v2: use the correct 'Fixes' tag
v2 -> v3: update commit message, drop unneeded READ_ONCE()
Fixes: 5b1158e909ec ("[NETFILTER]: Add NAT support for nf_conntrack") Reported-by: syzbot+8012e198bd037f4871e5@syzkaller.appspotmail.com Signed-off-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The ledinternal struct can be shared between several different
xt_LED targets, but the related timer is currently initialized only
if the first target requires it. Fix it by unconditionally
initializing the timer struct.
There is a race condition between clusterip_config_entry_put()
and clusterip_config_init(), after we release the spinlock in
clusterip_config_entry_put(), a new proc file with a same IP could
be created immediately since it is already removed from the configs
list, therefore it triggers this warning:
------------[ cut here ]------------
proc_dir_entry 'ipt_CLUSTERIP/172.20.0.170' already registered
WARNING: CPU: 1 PID: 4152 at fs/proc/generic.c:330 proc_register+0x2a4/0x370 fs/proc/generic.c:329
Kernel panic - not syncing: panic_on_warn set ...
As a quick fix, just move the proc_remove() inside the spinlock.
Reported-by: <syzbot+03218bcdba6aa76441a3@syzkaller.appspotmail.com> Fixes: 6c5d5cfbe3c5 ("netfilter: ipt_CLUSTERIP: check duplicate config when initializing") Tested-by: Paolo Abeni <pabeni@redhat.com> Cc: Xin Long <lucien.xin@gmail.com> Cc: Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com> Reviewed-by: Xin Long <lucien.xin@gmail.com> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The rationale for removing the check is only correct for rulesets
generated by ip(6)tables.
In iptables, a jump can only occur to a user-defined chain, i.e.
because we size the stack based on number of user-defined chains we
cannot exceed stack size.
However, the underlying binary format has no such restriction,
and the validation step only ensures that the jump target is a
valid rule start point.
IOW, its possible to build a rule blob that has no user-defined
chains but does contain a jump.
If this happens, no jump stack gets allocated and crash occurs
because no jumpstack was allocated.
Commit d1c4cb447a7e ("ASoC: Intel: Skylake: Fix jack name format
substitution") added Jack name but erroneously added a space as well,
so remove the space in Jack name.
Fixes: d1c4cb447a7e ("ASoC: Intel: Skylake: Fix jack name format substitution") Signed-off-by: Vinod Koul <vinod.koul@intel.com> Signed-off-by: Mark Brown <broonie@kernel.org> Cc: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In configurations without CONFIG_OMAP3 but with secure RAM support,
we now run into a link failure:
arch/arm/mach-omap2/omap-secure.o: In function `omap3_save_secure_ram':
omap-secure.c:(.text+0x130): undefined reference to `save_secure_ram_context'
The omap3_save_secure_ram() function is only called from the OMAP34xx
power management code, so we can simply hide that function in the
appropriate #ifdef.
Fixes: d09220a887f7 ("ARM: OMAP2+: Fix SRAM virt to phys translation for save_secure_ram_context") Acked-by: Tony Lindgren <tony@atomide.com> Tested-by: Dan Murphy <dmurphy@ti.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Gen8 and prior Proliant systems supported the "CRU" interface
to firmware. This interfaces allows linux to "call back" into firmware
to source the cause of an NMI. This feature isn't fully utilized
as the actual source of the NMI isn't printed, the driver only
indicates that the source couldn't be determined when the call
fails.
With the advent of Gen9, iCRU replaces the CRU. The call back
feature is no longer available in firmware. To be compatible and
not attempt to call back into firmware on system not supporting CRU,
the SMBIOS table is consulted to determine if it is safe to
make the call back or not.
This results in about half of the driver code being devoted
to either making CRU calls or determing if it is safe to make
CRU calls. As noted, the driver isn't really using the results of
the CRU calls.
Furthermore, as a consequence of the Spectre security issue, the
BIOS/EFI calls are being wrapped into Spectre-disabling section.
Removing the call back in hpwdt_pretimeout assists in this effort.
As the CRU sourcing of the NMI isn't required for handling the
NMI and there are security concerns with making the call back, remove
the legacy (pre Gen9) NMI sourcing and the DMI code to determine if
the system had the CRU interface.
This corrects:
commit cce78da76601 ("watchdog: hpwdt: Add check for UEFI bits")
The test on HPE SMBIOS extension type 219 record "Misc Features"
bits for UEFI support is incorrect. The definition of the Misc Features
bits in the HPE SMBIOS OEM Extensions specification (and related
firmware) was changed to use a different pair of bits to
represent UEFI supported. Howerver, a corresponding change
to Linux was missed.
Current code/platform work because the iCRU test is working.
But purpose of cce78da766 is to ensure correct functionality
on future systems where iCRU isn't supported.
Guests on new hypersiors might set KVM_ASYNC_PF_DELIVERY_AS_PF_VMEXIT
bit when enabling async_PF, but this bit is reserved on old hypervisors,
which results in a failure upon migration.
To avoid breaking different cases, we are checking for CPUID feature bit
before enabling the feature and nothing else.
Fixes: 52a5c155cf79 ("KVM: async_pf: Let guest support delivery of async_pf from guest mode") Cc: <stable@vger.kernel.org> Reviewed-by: Wanpeng Li <wanpengli@tencent.com> Reviewed-by: David Hildenbrand <david@redhat.com> Signed-off-by: Radim Krčmář <rkrcmar@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
[jwang: port to 4.14] Signed-off-by: Jack Wang <jinpu.wang@profitbricks.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
While working on 16338a9b3ac3 ("bpf, arm64: fix out of bounds access in
tail call") I noticed that ppc64 JIT is partially affected as well. While
the bound checking is correctly performed as unsigned comparison, the
register with the index value however, is never truncated into 32 bit
space, so e.g. a index value of 0x100000000ULL with a map of 1 element
would pass with PPC_CMPLW() whereas we later on continue with the full
64 bit register value. Therefore, as we do in interpreter and other JITs
truncate the value to 32 bit initially in order to fix access.
Fixes: ce0761419fae ("powerpc/bpf: Implement support for tail calls") Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Reviewed-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com> Tested-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The requirements around atomic_add() / atomic64_add() resp. their
JIT implementations differ across architectures. E.g. while x86_64
seems just fine with BPF's xadd on unaligned memory, on arm64 it
triggers via interpreter but also JIT the following crash:
Reason for this is because memory is required to be aligned. In
case of BPF, we always enforce alignment in terms of stack access,
but not when accessing map values or packet data when the underlying
arch (e.g. arm64) has CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS set.
xadd on packet data that is local to us anyway is just wrong, so
forbid this case entirely. The only place where xadd makes sense in
fact are map values; xadd on stack is wrong as well, but it's been
around for much longer. Specifically enforce strict alignment in case
of xadd, so that we handle this case generically and avoid such crashes
in the first place.
I recently noticed a crash on arm64 when feeding a bogus index
into BPF tail call helper. The crash would not occur when the
interpreter is used, but only in case of JIT. Output looks as
follows:
In this case the index used in BPF r3 is the same as in r1
at the time of the call, meaning we fed a pointer as index;
here, it had the value 0xffff808fd7cf0500 which sits in x2.
While I found tail calls to be working in general (also for
hitting the error cases), I noticed the following in the code
emission:
Meaning, the tests passed because commit ddb55992b04d ("arm64:
bpf: implement bpf_tail_call() helper") was using signed compares
instead of unsigned which as a result had the test wrongly passing.
Change this but also the tail call count test both into unsigned
and cap the index as u32. Latter we did as well in 90caccdd8cc0
("bpf: fix bpf_tail_call() x64 JIT") and is needed in addition here,
too. Tested on HiSilicon Hi1616.
Implement a retpoline [0] for the BPF tail call JIT'ing that converts
the indirect jump via jmp %rax that is used to make the long jump into
another JITed BPF image. Since this is subject to speculative execution,
we need to control the transient instruction sequence here as well
when CONFIG_RETPOLINE is set, and direct it into a pause + lfence loop.
The latter aligns also with what gcc / clang emits (e.g. [1]).
JIT dump after patch:
# bpftool p d x i 1
0: (18) r2 = map[id:1]
2: (b7) r3 = 0
3: (85) call bpf_tail_call#12
4: (b7) r0 = 2
5: (95) exit
Commit 9a3efb6b661f ("bpf: fix memory leak in lpm_trie map_free callback function")
fixed a memory leak and removed unnecessary locks in map_free callback function.
Unfortrunately, it introduced a lockdep warning. When lockdep checking is turned on,
running tools/testing/selftests/bpf/test_lpm_map will have:
[ 98.294321] =============================
[ 98.294807] WARNING: suspicious RCU usage
[ 98.295359] 4.16.0-rc2+ #193 Not tainted
[ 98.295907] -----------------------------
[ 98.296486] /home/yhs/work/bpf/kernel/bpf/lpm_trie.c:572 suspicious rcu_dereference_check() usage!
[ 98.297657]
[ 98.297657] other info that might help us debug this:
[ 98.297657]
[ 98.298663]
[ 98.298663] rcu_scheduler_active = 2, debug_locks = 1
[ 98.299536] 2 locks held by kworker/2:1/54:
[ 98.300152] #0: ((wq_completion)"events"){+.+.}, at: [<00000000196bc1f0>] process_one_work+0x157/0x5c0
[ 98.301381] #1: ((work_completion)(&map->work)){+.+.}, at: [<00000000196bc1f0>] process_one_work+0x157/0x5c0
Since actual trie tree removal happens only after no other
accesses to the tree are possible, replacing
rcu_dereference_protected(*slot, lockdep_is_held(&trie->lock))
with
rcu_dereference_protected(*slot, 1)
fixed the issue.
Fixes: 9a3efb6b661f ("bpf: fix memory leak in lpm_trie map_free callback function") Reported-by: Eric Dumazet <edumazet@google.com> Suggested-by: Eric Dumazet <edumazet@google.com> Signed-off-by: Yonghong Song <yhs@fb.com> Reviewed-by: Eric Dumazet <edumazet@google.com> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There is a memory leak happening in lpm_trie map_free callback
function trie_free. The trie structure itself does not get freed.
Also, trie_free function did not do synchronize_rcu before freeing
various data structures. This is incorrect as some rcu_read_lock
region(s) for lookup, update, delete or get_next_key may not complete yet.
The fix is to add synchronize_rcu in the beginning of trie_free.
The useless spin_lock is removed from this function as well.
Fixes: b95a5c4db09b ("bpf: add a longest prefix match trie map implementation") Reported-by: Mathieu Malaterre <malat@debian.org> Reported-by: Alexei Starovoitov <ast@kernel.org> Tested-by: Mathieu Malaterre <malat@debian.org> Signed-off-by: Yonghong Song <yhs@fb.com> Signed-off-by: Alexei Starovoitov <ast@kernel.org> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
syzkaller recently triggered OOM during percpu map allocation;
while there is work in progress by Dennis Zhou to add __GFP_NORETRY
semantics for percpu allocator under pressure, there seems also a
missing bpf_map_precharge_memlock() check in array map allocation.
Given today the actual bpf_map_charge_memlock() happens after the
find_and_alloc_map() in syscall path, the bpf_map_precharge_memlock()
is there to bail out early before we go and do the map setup work
when we find that we hit the limits anyway. Therefore add this for
array map as well.
The entire completions suppress mechanism is currently broken because the
HCA might retry a send operation (due to dropped ack) after the nvme
transaction has completed.
In order to handle this, we signal all send completions and introduce a
separate done handler for async events as they will be handled differently
(as they don't include in-capsule data by definition).
Signed-off-by: Sagi Grimberg <sagi@grimberg.me> Reviewed-by: Max Gurtovoy <maxg@mellanox.com> Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
projects / users / jedix / linux-maple.git / log