The number of supported WIDs, if they are supported at all, can be
limited due to resources. Notifying the user space application the
number of available WIDs allows it to utilize them correctly.
Signed-off-by: Ram Amrani <Ram.Amrani@cavium.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Direct Packet Mode support may be disabled, e.g, due to limited
resources. Notifying the user application prevents wasting cycles
on attempting to send these kind of packets.
Signed-off-by: Ram Amrani <Ram.Amrani@cavium.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Don't make any assumptions on the sg_io_hdr_t::dxfer_direction or the
sg_io_hdr_t::dxferp in order to determine if it is a valid request. The
only way we can check for bad requests is by checking if the length
exceeds 256M.
Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Fixes: 28676d869b (scsi: sg: check for valid direction before starting the
request)
Reported-by: Jason L Tibbitts III <tibbs@math.uh.edu>
Tested-by: Jason L Tibbitts III <tibbs@math.uh.edu>
Suggested-by: Doug Gilbert <dgilbert@interlog.com>
Cc: Doug Gilbert <dgilbert@interlog.com>
Cc: <stable@vger.kernel.org>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Since the PMU register interface is banked per CPU, CPU PMU interrrupts
cannot be handled by a CPU other than the one with the PMU asserting the
interrupt. This means that migrating PMU SPIs, as we do during a CPU
hotplug operation doesn't make any sense and can lead to the IRQ being
disabled entirely if we route a spurious IRQ to the new affinity target.
This has been observed in practice on AMD Seattle, where CPUs on the
non-boot cluster appear to take a spurious PMU IRQ when coming online,
which is routed to CPU0 where it cannot be handled.
This patch passes IRQF_PERCPU for PMU SPIs and forcefully sets their
affinity prior to requesting them, ensuring that they cannot
be migrated during hotplug events. This interacts badly with the DB8500
erratum workaround that ping-pongs the interrupt affinity from the handler,
so we avoid passing IRQF_PERCPU in that case by allowing the IRQ flags
to be overridden in the platdata.
Fixes: 3cf7ee98b8 ("drivers/perf: arm_pmu: move irq request/free into probe")
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
ISP mmu can't support reset operation, it won't get the
expected result when reset, but rest functions work normally.
Add this patch as a WA for this issue.
Signed-off-by: Simon Xue <xxm@rock-chips.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
RK3368 vpu mmu have two irqs, this patch support multi irqs
Signed-off-by: Simon Xue <xxm@rock-chips.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
sa1100 provides its own variant of the clk API rather than using the
generic COMMON_CLK API. This generally works, but it causes some link
errors with drivers using the clk_set_rate, clk_get_parent, clk_set_parent
or clk_round_rate functions when a platform lacks those interfaces.
This adds trivial stub implementations for each of them, based on
the behavior of the COMMON_CLK implementation:
- set_rate() and set_parent() report success without doing anything
- round_rate() returns the clk rate
- get_parent() returns NULL.
This adds the minimal bloat and should do the right thing for
the simple clock hardware in this SoC.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
davinci still has its own clk implementation, but lacks
a clk_get_parent() helper, which can lead to link errors
in randconfig builds.
This adds the usual implementation.
Acked-by: Sekhar Nori <nsekhar@ti.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
In commit 3169663ac5 "ARM: sa11x0/pxa: convert OS timer registers
to IOMEM", the definition of the OSCR macro was changed to be an
__iomem pointer, but the same register is also used by the XIP
code. This patch does the corresponding change here as well.
On PXA, the IRQ register definitions were removed even earlier, in
commit 5d284e353e ("ARM: pxa: avoid accessing interrupt registers
directly"). This patch unfortunately brings some of that back. An
earlier version of my patch moved the code into an external function,
which could not work for CONFIG_XIP_KERNEL+CONFIG_MTD_XIP, so this
restores something close to the original code.
Link: http://lists.infradead.org/pipermail/linux-arm-kernel/2014-March/241716.html
Acked-by: Robert Jarzmik <robert.jarzmik@free.fr>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
A change to the platform data definitions caused a warning in the board code:
arch/arm/mach-davinci/board-da850-evm.c:1221:13: warning: initialization discards ‘const’ qualifier from pointer target type [-Wdiscarded-qualifiers]
arch/arm/mach-davinci/board-da850-evm.c:1231:13: warning: initialization discards ‘const’ qualifier from pointer target type [-Wdiscarded-qualifiers]
This is a bit unfortunate, since we generally like structure definitions to
be const, but as this is legacy code, the easiest way out is still to
remove the 'const' annotation here.
Fixes: 4a5f8ae50b ("[media] davinci: vpif_capture: get subdevs from DT when available")
Fixes: 231ce279e6 ("ARM: davinci: fix const warnings")
Acked-by: Sekhar Nori <nsekhar@ti.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
The ZTE SoC drivers are only useful when building for a ZTE ZX platform.
Fixes: 4c2c2e3971 ("soc: zte: pm_domains: Prepare for supporting ARMv8 zx2967 family")
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Reviewed-by: Baoyou Xie <baoyou.xie@linaro.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Due to a bugfix in wireless tree and the commit mentioned below a merge
was needed which went haywire. So the submitted change resulted in the
function brcmf_sdiod_sgtable_alloc() being called twice during the probe
thus leaking the memory of the first call.
Cc: stable@vger.kernel.org # 4.6.x
Fixes: 4d79289598 ("brcmfmac: switch to new platform data")
Reported-by: Stefan Wahren <stefan.wahren@i2se.com>
Tested-by: Stefan Wahren <stefan.wahren@i2se.com>
Reviewed-by: Hante Meuleman <hante.meuleman@broadcom.com>
Signed-off-by: Arend van Spriel <arend.vanspriel@broadcom.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
The commit to rework the headroom check in start_xmit() now calls
pxskb_expand_head() unconditionally if the header is CoW. Unfortunately,
it does so with the delta between the extant headroom and the header
length, which may be negative if there is already sufficient headroom.
pskb_expand_head() does allow for size being 0, in which case it just
copies, so clamp the header delta to zero.
Opening Chrome (and all my tabs) on a PCIE device was enough to reliably
hit this.
Fixes: 270a6c1f65 ("brcmfmac: rework headroom check in .start_xmit()")
Signed-off-by: Daniel Stone <daniels@collabora.com>
Cc: Arend Van Spriel <arend.vanspriel@broadcom.com>
Cc: James Hughes <james.hughes@raspberrypi.org>
Cc: Hante Meuleman <hante.meuleman@broadcom.com>
Cc: Pieter-Paul Giesberts <pieter-paul.giesberts@broadcom.com>
Cc: Franky Lin <franky.lin@broadcom.com>
Tested-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Drop the unused endpoints. They should only be used when there is an
actual remote-endpoint connected.
Cc: Sekhar Nori <nsekhar@ti.com>
Signed-off-by: Kevin Hilman <khilman@baylibre.com>
Signed-off-by: Sekhar Nori <nsekhar@ti.com>
Drop the unused endpoints. They should only be used when there is
an actual remote-endpoint connected.
Signed-off-by: Kevin Hilman <khilman@baylibre.com>
Signed-off-by: Sekhar Nori <nsekhar@ti.com>
Broadcom BCM53573 SoCs actually have 32 GPIOs, and not 16.
Fixes: 3f37ec79dd ("bcma: support BCM53573 series of wireless SoCs")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
The current code kfree's entry and then dereferences it by accessing
entry->bssid. Avoid the dereference-after-free by moving the kfree
after the access to entry->bssid.
Detected by CoverityScan, CID#1448600 ("Read from pointer after free")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
- Fix wrong irq type for gpio expeander on Armada 388 GP
- Use __pa_symbol instead of virt_to_phys in the mv98dx3236 platform
SMP code
-----BEGIN PGP SIGNATURE-----
iIEEABECAEEWIQQYqXDMF3cvSLY+g9cLBhiOFHI71QUCWW3X4iMcZ3JlZ29yeS5j
bGVtZW50QGZyZWUtZWxlY3Ryb25zLmNvbQAKCRALBhiOFHI71e78AJ4+wNdgOn6C
8PVxz97Ngx2L0O44AgCfXWXbhHye3y4p2wfLfFTBT1prMZ4=
=gydP
-----END PGP SIGNATURE-----
Merge tag 'mvebu-fixes-4.13-1' of git://git.infradead.org/linux-mvebu into fixes
Pull "mvebu fixes for 4.13 (part 1)" from Gregory CLEMENT:
- Fix wrong irq type for gpio expeander on Armada 388 GP
- Use __pa_symbol instead of virt_to_phys in the mv98dx3236 platform
SMP code
* tag 'mvebu-fixes-4.13-1' of git://git.infradead.org/linux-mvebu:
ARM: dts: armada-38x: Fix irq type for pca955
ARM: mvebu: use __pa_symbol in the mv98dx3236 platform SMP code
Add necessary parent clocks for audss (Audio SubSystem, MAUDIO) clock
controller block.
This allows driver to keep EPLL enabled before accessing any MAUDIO
registers thus fixing silent hang. This silent hang appeared with
commit 6edfa11cb3 ("clk: samsung: Add enable/disable operation for
PLL36XX clocks"), e.g. on Odroid U3 usually with last (but unrelated)
messages:
[ 2.382741] input: gpio_keys as /devices/platform/gpio_keys/input/input0
[ 2.405686] usb 1-3: new high-speed USB device number 3 using exynos-ehci
[ 2.419843] max77686-rtc max77686-rtc: setting system clock to 2017-06-21 17:04:13 UTC (1498064653)
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Correct order of sound clock frequencies for ULCB boards
used by r8a7795 and r8a7796 SoCs.
These sounds clock frequencies are used as the ADG clock (output clocks
for audio module) initial setting and sound codec's initial system clock
which needs the maximum clock frequency. Thus, descending order is
required.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1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=bjUb
-----END PGP SIGNATURE-----
Merge tag 'renesas-fixes2-for-v4.13' of https://git.kernel.org/pub/scm/linux/kernel/git/horms/renesas into fixes
Pull "Second Round of Renesas ARM Based SoC Fixes for v4.13" from Simon Horman:
Correct order of sound clock frequencies for ULCB boards
used by r8a7795 and r8a7796 SoCs.
These sounds clock frequencies are used as the ADG clock (output clocks
for audio module) initial setting and sound codec's initial system clock
which needs the maximum clock frequency. Thus, descending order is
required.
* tag 'renesas-fixes2-for-v4.13' of https://git.kernel.org/pub/scm/linux/kernel/git/horms/renesas:
arm64: dts: renesas: ulcb: sound clock-frequency needs descending order
- Fix disable_irq related shared IRQ warnings for omap3 PRM
- Fix omap4 legacy code regression that accidentally removed code that
we still need for PRM interrupts
- Fix dm8168-evm NAND pins and MMC write protect pin direction
- Fix dra71-evm mdio impedance values
-----BEGIN PGP SIGNATURE-----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=ibLi
-----END PGP SIGNATURE-----
Merge tag 'omap-for-v4.13/fixes-merge-window' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap into fixes
Pull "Few fixes for omaps for issues found recently" from Tony Lindgren:
- Fix disable_irq related shared IRQ warnings for omap3 PRM
- Fix omap4 legacy code regression that accidentally removed code that
we still need for PRM interrupts
- Fix dm8168-evm NAND pins and MMC write protect pin direction
- Fix dra71-evm mdio impedance values
* tag 'omap-for-v4.13/fixes-merge-window' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap:
ARM: dts: dra71-evm: mdio: Fix impedance values
ARM: dts: dm816x: Correct the state of the write protect pin
ARM: dts: dm816x: Correct NAND support nodes
ARM: OMAP4: Fix legacy code clean-up regression
ARM: OMAP2+: Fix omap3 prm shared irq
Correct order of sound clock frequencies for Salvator boards
used by r8a7795 and r8a7796 SoCs.
These sounds clock frequencies are used as the ADG clock (output clocks
for audio module) initial setting and sound codec's initial system clock
which needs the maximum clock frequency. Thus, descending order is
required.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQIcBAABAgAGBQJZVhaCAAoJENfPZGlqN0++/zkP/1aHxFA560O3cb4SebIeJKQd
TQFd7YocK1lsqD6DP1ZLQI7ytPfpyAKPctKLRe2NSwr+kf5N8rbYUfZ88pBXe7ez
y9mbFBYcKHvDLrGHtW5hebBGB6HAh2Ll9DMW6Hqdsv8qG1/EqlLwBkA10/VSjaYk
Xfp4XnbyXsWCxbllU04We0uel945rg6tQy6VBszW5QEwMEX4zhaOsmi6z8nCv11E
k8Q6xl2/+FWc/zfivJSzBn59XjQSo+E54yRkmaYcteTY2+8Y787QMvVKkbUAvnHi
DrVplSPdbQPeZInqDiw97fvUNDTIzYatskMF3Lv6HFZ+u8YHTrW6yq2CShWtecUx
7BP8GgwfkLHt8J5PADMpNejzfE5lz+7EPm8+2jhGeoHd0BNjojaFvbyL01pJntwp
nJrSKlLmRB1CDEiubVPUUrY+EWcx7wxK0V8eliWTAn/o1ijJTErbRF91Z/U9BACF
PQUToBZ8MJKbjbrFXmelswkTwORV7au0KwY6Jzm2Vue7TiCZU4BXwyxZ1hmH3dI1
GLOWo5mtlXWcUnHLrPv44x6povNQDCtPGowTyKkqo2vXIAlAkMaG4V5D2tV9ykr7
d0UDWA44fF7QYcgW8hEd1eY0ILnYuX7Zanh0xvb5OBjqLEwsf+FvLel6/eMzg4Pj
tNB+iosh5YumUfCIue0D
=t800
-----END PGP SIGNATURE-----
Merge tag 'renesas-fixes-for-v4.13' of https://git.kernel.org/pub/scm/linux/kernel/git/horms/renesas into fixes
Pull "Renesas ARM Based SoC Fixes for v4.13" from Simon Horman:
Correct order of sound clock frequencies for Salvator boards
used by r8a7795 and r8a7796 SoCs.
These sounds clock frequencies are used as the ADG clock (output clocks
for audio module) initial setting and sound codec's initial system clock
which needs the maximum clock frequency. Thus, descending order is
required.
* tag 'renesas-fixes-for-v4.13' of https://git.kernel.org/pub/scm/linux/kernel/git/horms/renesas:
arm64: renesas: salvator-common: sound clock-frequency needs descending order
mwifiex records information about various channels as it receives scan
information. It does this by appending to a buffer that was sized
to the max number of supported channels on any band, but there are
numerous problems:
(a) scans can return info from more than one band (e.g., both 2.4 and 5
GHz), so the determined "max" is not large enough
(b) some firmware appears to return multiple results for a given
channel, so the max *really* isn't large enough
(c) there is no bounds checking when stashing these stats, so problems
(a) and (b) can easily lead to buffer overflows
Let's patch this by setting a slightly-more-correct max (that accounts
for a combination of both 2.4G and 5G bands) and adding a bounds check
when writing to our statistics buffer.
Due to problem (b), we still might not properly report all known survey
information (e.g., with "iw <dev> survey dump"), since duplicate results
(or otherwise "larger than expected" results) will cause some
truncation. But that's a problem for a future bugfix.
(And because of this known deficiency, only log the excess at the WARN
level, since that isn't visible by default in this driver and would
otherwise be a bit too noisy.)
Fixes: bf35443314 ("mwifiex: channel statistics support for mwifiex")
Cc: <stable@vger.kernel.org>
Cc: Avinash Patil <patila@marvell.com>
Cc: Xinming Hu <huxm@marvell.com>
Signed-off-by: Brian Norris <briannorris@chromium.org>
Reviewed-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Reviewed-by: Ganapathi Bhat <gbhat@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Turns out that just writing CURPOS isn't sufficient to move the cursor
on some platforms. My 830 works just fine, but eg. 945 and PNV don't.
On those platforms we need to arm even the CURPOS update with a
CURBASE write.
Even worse, a write to any of the cursor register apart from CURBASE
will cancel an already pending cursor update. So if we have armed a
CURCNTR/CURBASE update, a subsequent CURPOS write prior to vblank
would cancel that armed update. Thus we're left with a cursor that
doesn't appear to move, or even change shape.
Fix the problem by always performing the CURBASE write after a
CURPOS write. Bspec is somewhat unclear which platforms actually
require this CURBASE write and which don't. So to keep it simple
and to make sure we really fix the problem across all supported
devices, let's just perform the CURBASE write unconditionally.
Cc: Paul Menzel <pmenzel@molgen.mpg.de>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101790
Fixes: 75343a44c9 ("drm/i915: Drop useless posting reads from cursor commit")
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Tested-by: Paul Menzel <paulepanter@users.sourceforge.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20170714155227.6089-1-ville.syrjala@linux.intel.com
(cherry picked from commit 8753d2bc5e)
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Fix the sizeof(ptr) vs. sizeof(*ptr) typo.
Fixes: 2889caa923 ("drm/i915: Eliminate lots of iterations over the execobjects array")
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/20170714151242.517-2-imre.deak@intel.com
(cherry picked from commit edd9003f7f)
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Commit:
a7be6e5a7f ("mm: drop useless local parameters of __register_one_node()")
... removed the last user of parent_node(), so remove the macro.
Reported-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Dou Liyang <douly.fnst@cn.fujitsu.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1501076076-1974-11-git-send-email-douly.fnst@cn.fujitsu.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The hdmi bits simply don't exist, so nerf them. I think audio doesn't
work on gen3 at all, and for the limited color range we should
probably use the colorimetry sdvo paramater instead of the bit in the
port.
But fixing sdvo isn't my goal, I just want to get the backtrace out of
the way, and this takes care of that.
Still, while at it fix the missing read-out of the gen4 audio bit,
maybe that part even works ...
v2: Instead of trying to plug the damage in ->compute_config() make
sure we never set intel_sdvo->is_hdmi, which stops the bad state at
the source. Suggested by Chris Wilson. Also make sure we don't break
this by accident by putting a WARN_ON in place.
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170726193251.25393-1-daniel.vetter@ffwll.ch
lockdep complaints about a locking recursion for the i2c bus lock
because both the sdvo ddc proxy bus and the gmbus nested within use
the same locking class. It's not really a deadlock since we never nest
the other way round, but it's annoying.
Fix it by pulling the gmbus locking into the i2c lock_ops for both
i2c_adapater and making sure that the ddc_proxy_xfer function is
entirely lockless.
Re-layouting the extracted function resulted in some whitespace
cleanups, I figured we might as well keep them.
v2: Review from Chris:
- s/locked/unlocked/ since I got the naming backwards
- Use the vfuncs of the proxied adatper instead of re-rolling copies.
That's more consistent with the other proxying we're doing.
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170726132647.31833-1-daniel.vetter@ffwll.ch
Since we don't need the struct_mutex to acquire the object's pages, call
i915_gem_object_pin_pages() before we bind the object into the GGTT.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170726160038.29487-3-chris@chris-wilson.co.uk
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Actually transferring from shmemfs to the physically contiguous set of
pages should be wholly guarded by its obj->mm.lock!
v2: Remember to free the old pages.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170726160038.29487-2-chris@chris-wilson.co.uk
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reduce acquisition of struct_mutex to the critical regions that must
hold it; for KMS, we need struct_mutex currently only for the purpose of
pinning/unpinning the framebuffer's VMA into the global GTT. This allows
us to avoid taking the struct_mutex when disabling the CRTC (i.e. NULL
framebuffer objects) before a reset. (Not yet achieving the full goal of
avoiding the strut_mutex nesting, but good enough to break the first
half of the reset deadlock.)
v2: Keep pages pinning inside struct_mutex for the moment.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170726160038.29487-1-chris@chris-wilson.co.uk
[danvet: Drop another case of grabbing dev->struct_mutex around
cleanup_planes, which popped up because I had to redo the drm-next
backmerge for entirely different reasons. Acked by Chris on irc.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
When reading the i915_energy_uJ debugfs file, it tries to fetch
MSR_RAPL_POWER_UNIT, which might not be available, like in a vm
environment, causing the exception shown below.
We can easily prevent it by doing a rdmsrl_safe read instead, which will
handle the exception, allowing us to abort the debugfs file read.
This was caught by the new igt@debugfs_test@read_all_entries testcase in
the CI.
unchecked MSR access error: RDMSR from 0x606 at rIP:0xffffffffa0078f66
(i915_energy_uJ+0x36/0xb0 [i915])
Call Trace:
seq_read+0xdc/0x3a0
full_proxy_read+0x4f/0x70
__vfs_read+0x23/0x120
? putname+0x4f/0x60
? rcu_read_lock_sched_held+0x75/0x80
? entry_SYSCALL_64_fastpath+0x5/0xb1
vfs_read+0xa0/0x150
SyS_read+0x44/0xb0
entry_SYSCALL_64_fastpath+0x1c/0xb1
RIP: 0033:0x7f1f5e9f4500
RSP: 002b:00007ffc77e65cf8 EFLAGS: 00000246 ORIG_RAX: 0000000000000000
RAX: ffffffffffffffda RBX: ffffffff8146e003 RCX: 00007f1f5e9f4500
RDX: 0000000000000200 RSI: 00007ffc77e65d10 RDI: 0000000000000006
RBP: ffffc900007abf88 R08: 0000000001eaff20 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000
R13: 0000000000000006 R14: 0000000000000005 R15: 0000000001eb94db
? __this_cpu_preempt_check+0x13/0x20
v2:
- Drop unsigned long long cast and improve calculation (Chris)
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101901
Signed-off-by: Gabriel Krisman Bertazi <krisman@collabora.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/87o9s7zrx3.fsf@dilma.collabora.co.uk
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
The goal here was to minimise doing any thing or any check inside the
kernel that was not strictly required. For a userspace that assumes
complete control over the cache domains, the kernel is usually using
outdated information and may trigger clflushes where none were
required.
However, swapping is a situation where userspace has no knowledge of the
domain transfer, and will leave the object in the CPU cache. The kernel
must flush this out to the backing storage prior to use with the GPU. As
we use an asynchronous task tracked by an implicit fence for this, we
also need to cancel the ASYNC flag on the object so that the object will
wait for the clflush to complete before being executed. This also absolves
userspace of the responsibility imposed by commit 77ae995789 ("drm/i915:
Enable userspace to opt-out of implicit fencing") that its needed to ensure
that the object was out of the CPU cache prior to use on the GPU.
Fixes: 77ae995789 ("drm/i915: Enable userspace to opt-out of implicit fencing")
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101571
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Link: https://patchwork.freedesktop.org/patch/msgid/20170721145037.25105-5-chris@chris-wilson.co.uk
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
I was being overly paranoid in not updating the execobject.offset after
performing the fallback copy where we set reloc.presumed_offset to -1.
The thinking was to ensure that a subsequent NORELOC execbuf would be
forced to process the invalid relocations. However this is overkill so
long as we *only* update the execobject.offset following a successful
update of the relocation value witin the batch. If we have to repeat the
execbuf due to a later interruption, then we may skip the relocations on
the second pass (honouring NORELOC) since the execobject.offset match
the actual offsets (even though reloc.presumed_offset is garbage).
Subsequent calls to execbuf with NORELOC should themselves ensure that
the reloc.presumed_offset have been corrected in case of future
migration.
Reporting back the actual execobject.offset, even when
reloc.presumed_offset is garbage, ensures that reuse of those objects
use the latest information to avoid relocations.
Fixes: 2889caa923 ("drm/i915: Eliminate lots of iterations over the execobjects array")
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101635
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170721145037.25105-4-chris@chris-wilson.co.uk
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
After we detect a i915_vma pin overflow, we call __i915_vma_unpin to
cleanup. However, on an overflow the pin_count bitfield will be zero,
triggering an assertion, even though we the intention is to merely warn
and report the error back to the user (as historically the culprit has
be a leak in the display code).
Fixes: 20dfbde463 ("drm/i915: Wrap vma->pin_count accessors with small inline helpers")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170721145037.25105-2-chris@chris-wilson.co.uk
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
The condition for setting the Loadgen Select bit of
PORT_TX_DW4 register during DDI Vswing Sequence should be
Bit rate <=6 GHz whereas the existing code checks only
Bit Rate < 6GHz. This patch fixes this condition.
While at it also remove the redundant paranthesis.
Fixes: cf54ca8bc5 ("drm/i915/cnl: Implement voltage swing sequence.")
Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Manasi Navare <manasi.d.navare@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1500329122-32662-1-git-send-email-manasi.d.navare@intel.com
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
I wrote this code an year and a half ago and I couldn't exactly
remember the main differences of these two structures when reviewing a
new FBC patch. Add some comments to help explain what's the purpose of
each struct.
For the record, the original commits are:
b183b3f143 ("drm/i915/fbc: introduce struct intel_fbc_reg_params")
aaf78d276b ("drm/i915/fbc: introduce struct intel_fbc_state_cache")
Cc: Praveen Paneri <praveen.paneri@intel.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20170714193822.12121-1-paulo.r.zanoni@intel.com
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
* Don't define it twice.
* Define MSBs first, like the rest of i915_reg.h.
* Add CNL_ prefix to the bit that arrived in CNL.
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170714175228.27019-1-paulo.r.zanoni@intel.com
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
After applying af2788925ae0 ("drm/i915: Squelch reset messages during
selftests") out of sequence, I missed fixing up a call to i915_reset().
Reported-by: kbuild test robot <kbuild-all@01.org>
Fixes: af2788925ae0 ("drm/i915: Squelch reset messages during selftests")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20170725125336.11969-1-chris@chris-wilson.co.uk
Reviewed-by: David Weinehall <david.weinehall@linux.intel.com
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
During our selftests, we try reseting the GPU tens of thousands of
times, flooding the dmesg with our reset spam drowning out any potential
warnings. Add an option to i915_reset()/i915_reset_engine() to specify a
quiet reset for selftesting.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20170721123238.16428-19-chris@chris-wilson.co.uk
Reviewed-by: Michel Thierry <michel.thierry@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Extract the common barrier against rogue hangchecks from disrupting our
direct testing of resets, and in the process expand the lock to include
the per-engine reset shortcuts.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Cc: Michel Thierry <michel.thierry@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170721123238.16428-17-chris@chris-wilson.co.uk
Reviewed-by: Michel Thierry <michel.thierry@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
We require the caller to ensure that the packets they wish to emit into
the CS ring are qword aligned (i.e. have an even number of dwords).
Double check this.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170721161101.1618-1-chris@chris-wilson.co.uk
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
