| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Current code enables ASPM by default, it allows MHI to enter M2 state.
In case of one MSI vector, system hang is observed if ath11k does MHI
register reading in this state. The issue was reported on Dell XPS 13
9310 but is seen also on XPS 15 and XPS 17 laptops.
The workaround here is to prevent MHI from entering M2 state, this can
be done by disabling ASPM if only one MSI vector is used. When using 32
vectors ASPM is enabled as before.
Tested-on: QCA6390 hw2.0 PCI WLAN.HST.1.0.1-01740-QCAHSTSWPLZ_V2_TO_X86-1
Signed-off-by: Carl Huang <cjhuang@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Signed-off-by: Baochen Qiang <bqiang@codeaurora.org>
Link: https://lore.kernel.org/r/20211026041722.5271-1-bqiang@codeaurora.org
|
|
On some platforms it's not possible to allocate 32 MSI vectors for various
reasons, be it kernel configuration, VT-d disabled, buggy BIOS etc. So
ath11k was not able to use QCA6390 PCI devices on those platforms. Add
support for one MSI vector to solve that.
In case of one MSI vector, interrupt migration needs to be disabled. This
is because when interrupt migration happens, the msi_data may change.
However, msi_data is already programmed to rings during initial phase and
ath11k has no way to know that msi_data is changed during run time and
reprogram again.
In case of one MSI vector, MHI subsystem should not use IRQF_NO_SUSPEND
as QCA6390 doesn't set this flag too. Ath11k doesn't need to leave
IRQ enabled in suspend state.
Tested-on: QCA6390 hw2.0 PCI WLAN.HST.1.0.1-01740-QCAHSTSWPLZ_V2_TO_X86-1
Signed-off-by: Carl Huang <cjhuang@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Signed-off-by: Baochen Qiang <bqiang@codeaurora.org>
Link: https://lore.kernel.org/r/20211026041714.5219-1-bqiang@codeaurora.org
|
|
This is to prepare for one MSI vector support. IRQ enable and disable
of CE and DP are done only in case of multiple MSI vectors.
Tested-on: QCA6390 hw2.0 PCI WLAN.HST.1.0.1-01740-QCAHSTSWPLZ_V2_TO_X86-1
Signed-off-by: Carl Huang <cjhuang@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Signed-off-by: Baochen Qiang <bqiang@codeaurora.org>
Link: https://lore.kernel.org/r/20211026041705.5167-1-bqiang@codeaurora.org
|
|
Like ATH11K_PCI_IRQ_CE0_OFFSET, define ATH11K_PCI_IRQ_DP_OFFSET for
DP to save the IRQ instead of base_vector from MSI config.
Tested-on: QCA6390 hw2.0 PCI WLAN.HST.1.0.1-01740-QCAHSTSWPLZ_V2_TO_X86-1
Signed-off-by: Carl Huang <cjhuang@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Signed-off-by: Baochen Qiang <bqiang@codeaurora.org>
Link: https://lore.kernel.org/r/20211026041655.5112-1-bqiang@codeaurora.org
|
|
This change adds two flags to indicate whether IRQ handler for CE
and DP can be called. This is because in one MSI vector case,
interrupt is not disabled in hif_stop and hif_irq_disable. Otherwise,
MHI interrupt is disabled too.
Tested-on: QCA6390 hw2.0 PCI WLAN.HST.1.0.1-01740-QCAHSTSWPLZ_V2_TO_X86-1
Signed-off-by: Carl Huang <cjhuang@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Signed-off-by: Baochen Qiang <bqiang@codeaurora.org>
Link: https://lore.kernel.org/r/20211026041646.5060-1-bqiang@codeaurora.org
|
|
The reservation mode of interrupts in kernel assigns a dummy vector
when the interrupt is allocated and assigns a real vector when the
request_irq is called. The reservation mode helps to ease vector
pressure when devices with a large amount of queues/interrupts
are initialized, but only a minimal subset of those queues/interrupts
is actually used.
So on reservation mode, the msi_data may change after request_irq
is called, so ath11k reads msi_data again after request_irq is called,
and then the correct msi_data is programmed into QCA6390 hardware
components. Without this change, spurious interrupt occurs in case of
one MSI vector. When VT-d in BIOS is enabled and ath11k can get 32 MSI
vectors, ath11k always get the same msi_data before and after request_irq,
that's why this change is only required when one MSI vector is to be
supported.
Tested-on: QCA6390 hw2.0 PCI WLAN.HST.1.0.1-01740-QCAHSTSWPLZ_V2_TO_X86-1
Signed-off-by: Carl Huang <cjhuang@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20211026041636.5008-1-bqiang@codeaurora.org
|
|
The ARCH_FEATURES function ID is a 32-bit SMC call, which returns
a 32-bit result per the SMCCC spec. Current code is doing a 64-bit
comparison against -1 (SMCCC_RET_NOT_SUPPORTED) to detect that the
feature is unimplemented. That check doesn't work in a Hyper-V VM,
where the upper 32-bits are zero as allowed by the spec.
Cast the result as an 'int' so the comparison works. The change also
makes the code consistent with other similar checks in this file.
Fixes: 821b67fa4639 ("firmware: smccc: Add ARCH_SOC_ID support")
Signed-off-by: Michael Kelley <mikelley@microsoft.com>
Reviewed-by: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
|
|
Arm SCMI fixes for v5.16
Couple of fixes for sparse warnings(type error assignment in voltage and
sensor protocols), add proper propagation of error from scmi_pm_domain_probe
handling agent discovery response in base protocol correctly and a fix
to avoid null pointer de-reference in the error path.
* tag 'scmi-fixes-5.16' of git://git.kernel.org/pub/scm/linux/kernel/git/sudeep.holla/linux:
firmware: arm_scmi: Fix type error assignment in voltage protocol
firmware: arm_scmi: Fix type error in sensor protocol
firmware: arm_scmi: pm: Propagate return value to caller
firmware: arm_scmi: Fix base agent discover response
firmware: arm_scmi: Fix null de-reference on error path
Link: https://lore.kernel.org/r/20211118121656.4014764-1-sudeep.holla@arm.com
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
|
|
Fix possible NULL pointer dereference in OP-TEE driver
* tag 'optee-fix-for-v5.16' of git://git.linaro.org/people/jens.wiklander/linux-tee:
optee: fix kfree NULL pointer
Link: https://lore.kernel.org/r/20211117125747.GA2896197@jade
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
|
|
Don't use _HIZ macros but also provide a val. This is more consistent
with the other macros and, thus, easier to read. Also shorter.
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Link: https://lore.kernel.org/r/20211119110442.4946-1-wsa+renesas@sang-engineering.com
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
|
|
Instead of writing fixed values with undocumented bits which happen to
be set on some SoCs, better switch to read-modify-write operations
changing only bits which are documented. This is way more future-proof
as we don't know yet how these bits may be on upcoming SoCs.
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Tested-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Reviewed-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Link: https://lore.kernel.org/r/20211117093710.14430-1-wsa+renesas@sang-engineering.com
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
|
|
No need to open code regmap_update_bits().
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Tested-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Link: https://lore.kernel.org/r/20211117102902.20062-1-wsa+renesas@sang-engineering.com
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
|
|
Some clocks shouldn't be automatically disabled in clk_disable_unused(),
otherwise kernel hangs. Mark those clocks with:
- CLK_IS_CRITICAL flag, when there won't be any consumers for that
clock, but system can't function when it's gated
- CLK_IGNORE_UNUSED flag, when consumer driver will be probably added
later
That makes it possible to run the kernel without passing the
"clk_ignore_unused" param.
Next clocks were modified:
- "gout_dpu_cmu_dpu_pclk": CLK_IGNORE_UNUSED
Will be enabled later in DSIM driver (Display Serial Interface
Master).
- "gout_gpio_peri_pclk": CLK_IGNORE_UNUSED
"gout_gpio_cmgp_pclk": CLK_IGNORE_UNUSED
"gout_gpio_hsi_pclk": CLK_IGNORE_UNUSED
Should be probably enabled in corresponding GPIO driver later, or
made CLK_IS_CRITICAL. "gout_gpio_peri_clk" is actually used by LEDs
on Exynos850-based dev board, so kernel hangs if this clock is not
running. Other clocks were marked as "ignore unused" to prevent
similar issues for other use cases or boards that might be added
later.
- "gout_cci_aclk": CLK_IS_CRITICAL
CCI (Cache Coherent Interconnect): obviously is critical.
- "gout_gic_clk": CLK_IS_CRITICAL
GIC (Generic Interrupt Controller): obviously is critical.
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20211121232741.6967-7-semen.protsenko@linaro.org
|
|
This patch silences the following clang warning:
| drivers/memory/renesas-rpc-if.c:253:14: warning: cast to smaller integer
| type 'enum rpcif_type' from 'const void *' [-Wvoid-pointer-to-enum-cast]
| rpc->type = (enum rpcif_type)of_device_get_match_data(dev);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Fixes: b04cc0d912eb8 ("memory: renesas-rpc-if: Add support for RZ/G2L")
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/20211121180155.9062-1-prabhakar.mahadev-lad.rj@bp.renesas.com
[krzysztof: drop enum rpcif_type cast]
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
|
|
When guc timestamp ping worker runs it takes the spinlock and calls
with_intel_runtime_pm. Since with_intel_runtime_pm may sleep, move the
spinlock inside __update_guc_busyness_stats.
Signed-off-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20211120014201.26480-1-umesh.nerlige.ramappa@intel.com
|
|
CMU_CMGP clock domain provides clocks for CMGP IP-core (Common GPIO).
CMGP module encapsulates next blocks:
- 8 GPIO lines
- 1 GPADC
- 2 USI blocks, each can be configured to provide one of
UART/SPI/HSI2C serial interfaces
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Link: https://lore.kernel.org/r/20211121232741.6967-5-semen.protsenko@linaro.org
|
|
CMU_APM clock domain provides clocks for APM IP-core (Active Power
Management). According to Exynos850 TRM, CMU_APM generates I3C, Mailbox,
Speedy, Timer, WDT, RTC and PMU clocks for BLK_ALIVE.
This patch adds next clocks:
- bus clocks in CMU_TOP needed for CMU_APM
- all internal CMU_APM clocks
- leaf clocks for I3C, Speedy and RTC IP-cores
- bus clocks for CMU_CMGP and CMU_CHUB
CMU_APM doesn't belong to Power Domains, but platform driver is used for
its registration to keep its bus clock always running. Otherwise rtc-s3c
driver disables that clock and system freezes.
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Acked-by: Chanwoo Choi <cw00.choi@samsung.com>
Link: https://lore.kernel.org/r/20211121232741.6967-3-semen.protsenko@linaro.org
|
|
Function sunxi_rsb_hw_exit() is sometimes called with pm runtime
disabled, so in such cases pm_runtime_resume() will fail with -EACCES.
Instead of doing whole dance of enabling pm runtime and thus clock just
to disable it again immediately, just check if disabling clock is
needed. That way calling pm_runtime_resume() is not needed at all.
Fixes: 4a0dbc12e618 ("bus: sunxi-rsb: Implement runtime power management")
Signed-off-by: Jernej Skrabec <jernej.skrabec@gmail.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://lore.kernel.org/r/20211121083537.612473-1-jernej.skrabec@gmail.com
|
|
While it is useful to build all of the CCU drivers at once, only 1-3 of
them will be loaded at a time, or possibly none of them if the kernel is
booted on a non-sunxi platform. These CCU drivers are relatively large;
32-bit drivers have 30-50k of data each, while the 64-bit ones are
50-75k due to the increased pointer overhead. About half of that data
comes from relocations. Let's allow the user to build these drivers as
modules so only the necessary data is loaded.
As a first step, convert the CCUs that are already platform drivers.
When the drivers are built as modules, normally the file name becomes
the module name. However, the current file names are inconsistent with
the <platform>-<peripheral> name used everywhere else: the devicetree
bindings, the platform driver names, and the Kconfig symbols. Use
Makfile logic to rename the modules so they follow the usual pattern.
Signed-off-by: Samuel Holland <samuel@sholland.org>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://lore.kernel.org/r/20211119033338.25486-3-samuel@sholland.org
|
|
For the individual CCU drivers to be built as modules, the ops structs,
helper functions, and callback registration functions must be exported.
These symbols are intended for use only by the adjacent CCU drivers, so
export them into the SUNXI_CCU namespace.
of_sunxi_ccu_probe is not exported because it is only used by built-in
OF clock providers.
Signed-off-by: Samuel Holland <samuel@sholland.org>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Link: https://lore.kernel.org/r/20211119033338.25486-2-samuel@sholland.org
|
|
The macro p_to_snd_cx18_card() is just a variant for container_of(),
not actually used inside the code. So, drop it.
This address a clang warning.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
The macro p_to_snd_ivtv_card() is just a variant for container_of(),
not actually used inside the code. So, drop it.
This address a clang warning.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
This function is defined twice, one at cx25821-i2c and the other
inside cx25821-core. It turns that only the first
is actually used inside the code.
This causes a clang warning.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
Add the following Telit LE910S1 composition:
0x9200: tty
Signed-off-by: Daniele Palmas <dnlplm@gmail.com>
Link: https://lore.kernel.org/r/20211119140319.10448-1-dnlplm@gmail.com
Cc: stable@vger.kernel.org
Signed-off-by: Johan Hovold <johan@kernel.org>
|
|
The G2 decoder block produces NV12 4x4 tiled format (NV12_4L4).
Enable the G2 post-processor block, in order to produce regular NV12.
The logic in hantro_postproc.c is leveraged to take care of allocating
the extra buffers and configure the post-processor, which is
significantly simpler than the one on the G1.
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
The struct is not used outside this file, so it can be static.
Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Reviewed-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
VeriSilicon Hantro G2 core supports VP9 codec.
[hverkuil: add kerneldoc line for HANTRO_MODE_VP9_DEC]
Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Reviewed-by: Benjamin Gaignard <benjamin.gaignard@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
VeriSilicon Hantro G2 core supports other codecs besides hevc.
Factor out some common code in preparation for vp9 support.
Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Reviewed-by: Benjamin Gaignard <benjamin.gaignard@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
Add more consistency in the way registers are named.
Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Reviewed-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
The Rockchip VDEC supports VP9 profile 0 up to 4096x2304@30fps. Add
a backend for this new format.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Adrian Ratiu <adrian.ratiu@collabora.com>
Co-developed-by: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
Provide code common to vp9 drivers in one central location.
Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
Add the VP9 stateless decoder controls plus the documentation that goes
with it.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Co-developed-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Adrian Ratiu <adrian.ratiu@collabora.com>
Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Co-developed-by: Daniel Almeida <daniel.almeida@collabora.com>
Signed-off-by: Daniel Almeida <daniel.almeida@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
The G2 core decoder engine produces NV12_4L4 format,
which is a simple NV12 4x4 tiled format. The driver currently
hides this format by always enabling the post-processor engine,
and therefore offering NV12 directly.
This is done without using the logic in hantro_postproc.c
and therefore makes it difficult to add VP9 cleanly.
Since fixing this is not easy, add a small quirk to force
NV12 if HEVC was configured, but otherwise declare NV12_4L4
as the pixel format in imx8mq_vpu_g2_variant.dec_fmts.
This will be used by the VP9 decoder which will be added soon.
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
Add a 'postprocessed' boolean property to struct hantro_fmt
to signal that a format is produced by the post-processor.
This will allow to introduce the G2 post-processor in a simple way.
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
Turns out the post-processor block on the G2 core is substantially
different from the one on the G1 core. Introduce hantro_postproc_ops
with .enable and .disable methods, which will allow to support
the G2 post-processor cleanly.
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Reviewed-by: Benjamin Gaignard <benjamin.gaignard@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
When the post-processor hardware block is enabled, the driver
allocates an internal queue of buffers for the decoder enginer,
and uses the vb2 queue for the post-processor engine.
For instance, on a G1 core, the decoder engine produces NV12 buffers
and the post-processor engine can produce YUY2 buffers. The decoder
engine expects motion vectors to be appended to the NV12 buffers,
but this is only required for CODECs that need motion vectors,
such as H.264.
Fix the post-processor logic accordingly.
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
|
|
Dan reports that smatch has found idxd_wq_quiesce() is being called inside
the idxd->dev_lock. idxd_wq_quiesce() calls wait_for_completion() and
therefore it can sleep. Move the call outside of the spinlock as it does
not need device lock.
Fixes: 5b0c68c473a1 ("dmaengine: idxd: support reporting of halt interrupt")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/163716858508.1721911.15051495873516709923.stgit@djiang5-desk3.ch.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
The print function dev_err() is redundant because
platform_get_irq() already prints an error.
Signed-off-by: Xu Wang <vulab@iscas.ac.cn>
Link: https://lore.kernel.org/r/20211116013306.784-1-vulab@iscas.ac.cn
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
The commit in the Fixes: tag has changed the logic of the code and now it
is likely that the probe will return an early success (0), even if not
completely executed.
This should lead to a crash or similar issue later on when the code
accesses to some never allocated resources.
Change the '!err' into a 'err' when checking if
'dma_set_mask_and_coherent()' has failed or not.
While at it, simplify the code and remove the "can't success code" related
to 32 DMA mask.
As stated in [1], 'dma_set_mask_and_coherent(DMA_BIT_MASK(64))' can't fail
if 'dev->dma_mask' is non-NULL. And if it is NULL, it would fail for the
same reason when tried with DMA_BIT_MASK(32).
[1]: https://lkml.org/lkml/2021/6/7/398
Fixes: ecb8c88bd31c ("dmaengine: dw-edma-pcie: switch from 'pci_' to 'dma_' API")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Link: https://lore.kernel.org/r/935fbb40ae930c5fe87482a41dcb73abf2257973.1636492127.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
This new CDMA binding for device_prep_dma_memcpy_sg was partially borrowed
from xlnx kernel tree, an expanded with extended address space support when
linking descriptor segments and checking for incorrect zero transfer size.
Signed-off-by: Adrian Larumbe <adrianml@alumnos.upm.es>
Link: https://lore.kernel.org/r/20211101180825.241048-4-adrianml@alumnos.upm.es
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
This is the old DMA_SG interface that was removed in commit
c678fa66341c ("dmaengine: remove DMA_SG as it is dead code in kernel"). It
has been renamed to DMA_MEMCPY_SG to better match the MEMSET and MEMSET_SG
naming convention.
It should only be used for mem2mem copies, either main system memory or
CPU-addressable device memory (like video memory on a PCI graphics card).
Bringing back this interface was prompted by the need to use the Xilinx
CDMA device for mem2mem SG transfers.
Signed-off-by: Adrian Larumbe <adrianml@alumnos.upm.es>
Link: https://lore.kernel.org/r/20211101180825.241048-3-adrianml@alumnos.upm.es
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
Coverity complains of an uninitialized variable:
5. uninit_use_in_call: Using uninitialized value config.dst_per when calling axi_chan_config_write. [show details]
6. uninit_use_in_call: Using uninitialized value config.hs_sel_src when calling axi_chan_config_write. [show details]
CID 121164 (#1-3 of 3): Uninitialized scalar variable (UNINIT)
7. uninit_use_in_call: Using uninitialized value config.src_per when calling axi_chan_config_write. [show details]
418 axi_chan_config_write(chan, &config);
Fix this by initializing the structure to 0 which should at least be benign in axi_chan_config_write(). Also fix
what looks like a cut-n-paste error when initializing config.hs_sel_dst.
Fixes: 824351668a413 ("dmaengine: dw-axi-dmac: support DMAX_NUM_CHANNELS > 8")
Cc: Eugeniy Paltsev <Eugeniy.Paltsev@synopsys.com>
Cc: Vinod Koul <vkoul@kernel.org>
Cc: dmaengine@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Tim Gardner <tim.gardner@canonical.com>
Link: https://lore.kernel.org/r/20211025181656.31658-1-tim.gardner@canonical.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
"Interrupt handle revoked" is an event that happens when the driver is
running on a guest kernel and the VM is migrated to a new machine.
The device will trigger an interrupt that signals to the guest driver
that the interrupt handles need to be replaced.
The misc irq thread function calls a helper function to handle the
event. The function uses the WQ percpu_ref to quiesce the kernel
submissions. It then replaces the interrupt handles by requesting
interrupt handle command for each I/O MSIX vector. Once the handle is
updated, the driver will unblock the submission path to allow new
submissions.
The submitter will attempt to acquire a percpu_ref before submission. When
the request fails, it will wait on the wq_resurrect 'completion'.
The driver does anticipate the possibility of descriptors being submitted
before the WQ percpu_ref is killed. If a descriptor has already been
submitted, it will return with incorrect interrupt handle status. The
descriptor will be re-submitted with the new interrupt handle on the
completion path. For descriptors with incorrect interrupt handles,
completion interrupt won't be triggered.
At the completion of the interrupt handle refresh, the handling function
will call idxd_int_handle_refresh_drain() to issue drain descriptors to
each of the wq with associated interrupt handle. The drain descriptor will have
interrupt request set but without completion record. This will ensure all
descriptors with incorrect interrupt completion handle get drained and
a completion interrupt is triggered for the guest driver to process them.
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Co-Developed-by: Sanjay Kumar <sanjay.k.kumar@intel.com>
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/163528420189.3925689.18212568593220415551.stgit@djiang5-desk3.ch.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
Handle a descriptor that has been marked with invalid interrupt handle
error in status. Create a work item that will resubmit the descriptor. This
typically happens when the driver has handled the revoke interrupt handle
event and has a new interrupt handle.
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/163528419601.3925689.4166517602890523193.stgit@djiang5-desk3.ch.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
Add a locked version of idxd_quiesce() call so that the quiesce can be
called with a lock in situations where the lock is not held by the caller.
In the driver probe/remove path, the lock is already held, so the raw
version can be called w/o locking.
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/163528418980.3925689.5841907054957931211.stgit@djiang5-desk3.ch.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
The helper is called at the completion of the interrupt handle refresh
event. It issues drain descriptors to each of the wq with associated
interrupt handle. The drain descriptor will have interrupt request set but
without completion record. This will ensure all descriptors with incorrect
interrupt completion handle get drained and a completion interrupt is
triggered for the guest driver to process them.
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/163528418315.3925689.7944718440052849626.stgit@djiang5-desk3.ch.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
In preparation of supporting interrupt handle revoke event, move the
interrupt handle assignment to right before the descriptor to be submitted.
This allows the interrupt handle revoke logic to assign the latest
interrupt handle on submission.
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/163528417767.3925689.7730411152122952808.stgit@djiang5-desk3.ch.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
Attach int_handle to irq_entry. This removes the separate management of int
handles and reduces the confusion of interating through int handles that is
off by 1 count.
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/163528417065.3925689.11505755433684476288.stgit@djiang5-desk3.ch.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
Refactor the completion function to allow skipping of descriptor freeing on
the submission failure path. This completely removes descriptor freeing
from the submit failure path and leave the responsibility to the caller.
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/163528416222.3925689.12859769271667814762.stgit@djiang5-desk3.ch.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
|
Use 'for_each_set_bit()' instead of hand wrinting it. It is much less
version.
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Acked-by: Peter Ujfalusi <peter.ujfalusi@gmail.com>
Link: https://lore.kernel.org/r/47a7415d3aff8dfb66780bd6f80b085db4503bf7.1637263609.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Vinod Koul <vkoul@kernel.org>
|
Carl Huang
Michael Kelley
Arnd Bergmann
Wolfram Sang
Sam Protsenko
Lad Prabhakar
Umesh Nerlige Ramappa
Jernej Skrabec
Samuel Holland
Mauro Carvalho Chehab
Daniele Palmas
Andrzej Pietrasiewicz
Ezequiel Garcia
Dave Jiang
Xu Wang
Christophe JAILLET
Adrian Larumbe
Tim Gardner