This makes the io-uring tracepoints consistent. Where it makes sense
the tracepoints start with the following four fields:
- context (ring)
- request
- user_data
- opcode.
Signed-off-by: Stefan Roesch <shr@fb.com>
Link: https://lore.kernel.org/r/20220214180430.70572-3-shr@fb.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This introduces the __fill_cqe function. This is necessary
to correctly issue the io_uring_complete tracepoint.
Signed-off-by: Stefan Roesch <shr@fb.com>
Link: https://lore.kernel.org/r/20220214180430.70572-2-shr@fb.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
wqe->lock is abused, it now protects acct->work_list, hash stuff,
nr_workers, wqe->free_list and so on. Lets first get the work_list out
of the wqe-lock mess by introduce a specific lock for work list. This
is the first step to solve the huge contension between work insertion
and work consumption.
good thing:
- split locking for bound and unbound work list
- reduce contension between work_list visit and (worker's)free_list.
For the hash stuff, since there won't be a work with same file in both
bound and unbound work list, thus they won't visit same hash entry. it
works well to use the new lock to protect hash stuff.
Results:
set max_unbound_worker = 4, test with echo-server:
nice -n -15 ./io_uring_echo_server -p 8081 -f -n 1000 -l 16
(-n connection, -l workload)
before this patch:
Samples: 2M of event 'cycles:ppp', Event count (approx.): 1239982111074
Overhead Command Shared Object Symbol
28.59% iou-wrk-10021 [kernel.vmlinux] [k] native_queued_spin_lock_slowpath
8.89% io_uring_echo_s [kernel.vmlinux] [k] native_queued_spin_lock_slowpath
6.20% iou-wrk-10021 [kernel.vmlinux] [k] _raw_spin_lock
2.45% io_uring_echo_s [kernel.vmlinux] [k] io_prep_async_work
2.36% iou-wrk-10021 [kernel.vmlinux] [k] _raw_spin_lock_irqsave
2.29% iou-wrk-10021 [kernel.vmlinux] [k] io_worker_handle_work
1.29% io_uring_echo_s [kernel.vmlinux] [k] io_wqe_enqueue
1.06% iou-wrk-10021 [kernel.vmlinux] [k] io_wqe_worker
1.06% io_uring_echo_s [kernel.vmlinux] [k] _raw_spin_lock
1.03% iou-wrk-10021 [kernel.vmlinux] [k] __schedule
0.99% iou-wrk-10021 [kernel.vmlinux] [k] tcp_sendmsg_locked
with this patch:
Samples: 1M of event 'cycles:ppp', Event count (approx.): 708446691943
Overhead Command Shared Object Symbol
16.86% iou-wrk-10893 [kernel.vmlinux] [k] native_queued_spin_lock_slowpat
9.10% iou-wrk-10893 [kernel.vmlinux] [k] _raw_spin_lock
4.53% io_uring_echo_s [kernel.vmlinux] [k] native_queued_spin_lock_slowpat
2.87% iou-wrk-10893 [kernel.vmlinux] [k] io_worker_handle_work
2.57% iou-wrk-10893 [kernel.vmlinux] [k] _raw_spin_lock_irqsave
2.56% io_uring_echo_s [kernel.vmlinux] [k] io_prep_async_work
1.82% io_uring_echo_s [kernel.vmlinux] [k] _raw_spin_lock
1.33% iou-wrk-10893 [kernel.vmlinux] [k] io_wqe_worker
1.26% io_uring_echo_s [kernel.vmlinux] [k] try_to_wake_up
spin_lock failure from 25.59% + 8.89% = 34.48% to 16.86% + 4.53% = 21.39%
TPS is similar, while cpu usage is from almost 400% to 350%
Signed-off-by: Hao Xu <haoxu@linux.alibaba.com>
Link: https://lore.kernel.org/r/20220206095241.121485-2-haoxu@linux.alibaba.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Clang warns:
fs/io_uring.c:9396:9: warning: variable 'ret' is uninitialized when used here [-Wuninitialized]
return ret;
^~~
fs/io_uring.c:9373:13: note: initialize the variable 'ret' to silence this warning
int fd, ret;
^
= 0
1 warning generated.
Just return 0 directly and reduce the scope of ret to the if statement,
as that is the only place that it is used, which is how the function was
before the fixes commit.
Fixes: 1a75fac9a0f9 ("io_uring: avoid ring quiesce while registering/unregistering eventfd")
Link: https://github.com/ClangBuiltLinux/linux/issues/1579
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Link: https://lore.kernel.org/r/20220207162410.1013466-1-nathan@kernel.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
None of the opcodes in io_uring_register use ring quiesce anymore. Hence
io_register_op_must_quiesce always returns false and io_ctx_quiesce is
never called.
Signed-off-by: Usama Arif <usama.arif@bytedance.com>
Link: https://lore.kernel.org/r/20220204145117.1186568-6-usama.arif@bytedance.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
IORING_SETUP_R_DISABLED prevents submitting requests and so there will be
no requests until IORING_REGISTER_ENABLE_RINGS is called. And
IORING_REGISTER_RESTRICTIONS works only before
IORING_REGISTER_ENABLE_RINGS is called. Hence ring quiesce is not needed
for these opcodes.
Signed-off-by: Usama Arif <usama.arif@bytedance.com>
Link: https://lore.kernel.org/r/20220204145117.1186568-5-usama.arif@bytedance.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This is done using the RCU data structure (io_ev_fd). eventfd_async is
moved from io_ring_ctx to io_ev_fd which is RCU protected hence avoiding
ring quiesce which is much more expensive than an RCU lock. The place
where eventfd_async is read is already under rcu_read_lock so there is no
extra RCU read-side critical section needed.
Signed-off-by: Usama Arif <usama.arif@bytedance.com>
Link: https://lore.kernel.org/r/20220204145117.1186568-4-usama.arif@bytedance.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This is done by creating a new RCU data structure (io_ev_fd) as part of
io_ring_ctx that holds the eventfd_ctx.
The function io_eventfd_signal is executed under rcu_read_lock with a
single rcu_dereference to io_ev_fd so that if another thread unregisters
the eventfd while io_eventfd_signal is still being executed, the
eventfd_signal for which io_eventfd_signal was called completes
successfully.
The process of registering/unregistering eventfd is already done under
uring_lock so multiple threads won't enter a race condition while
registering/unregistering eventfd.
With the above approach ring quiesce can be avoided which is much more
expensive then using RCU lock. On the system tested, io_uring_register
with IORING_REGISTER_EVENTFD takes less than 1ms with RCU lock, compared
to 15ms before with ring quiesce.
Signed-off-by: Usama Arif <usama.arif@bytedance.com>
Link: https://lore.kernel.org/r/20220204145117.1186568-3-usama.arif@bytedance.com
[axboe: long line fixups]
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The information on whether eventfd is registered is not very useful and
would result in the tracepoint being enclosed in an rcu_readlock in a
later patch that tries to avoid ring quiesce for registering eventfd.
Suggested-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Usama Arif <usama.arif@bytedance.com>
Link: https://lore.kernel.org/r/20220204145117.1186568-2-usama.arif@bytedance.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Rather more fixes here than I'm comfortable with, we've had several
vendors noticing issues late in the release cycle all of which are valid
and reasonable fixes but it adds up to a much larger change set than I'd
like. Several of the AMD fixes look like cleanups from the subject
lines but are actually fixing user visible problems as well.
If you were to merge this for 5.18 rather than 5.17 it wouldn't be the
end of the world, stable will probably backport everything anyway.
-----BEGIN PGP SIGNATURE-----
iQEzBAABCgAdFiEEreZoqmdXGLWf4p/qJNaLcl1Uh9AFAmIp6mAACgkQJNaLcl1U
h9C9uAf/ZoIHlCa7mWD7VixBSl1DEsoV52KXUgzbzZBdZCTml45HxpsjVFCU3rZm
V7Cqsy3+Vbdz+U0KFnSm2KM0YfYAKsJaKXdvP2TKAEyLqJkCQxgYaN5SnlWFWjEy
H6yEzdOdRRn0tFR/DUbwxU2fALrXhAizK3FgX7jJraxE0dOcYXeKAe7nyjMq8cU3
K4ySo7dXyyamRSUrqkQLXBPXFktLjVZuhCfP1NaMJFr+sozj9EdN3VIgRZaJ1oet
qizjiibam+AxBlQMe8xLeteFPw2iy9qTQvzV2vLF7URMyQyHcS+oq8//AHtCbyG5
JD9C2NimWkKqyRuPh2gq1o37tuIIFg==
=o3xQ
-----END PGP SIGNATURE-----
Merge tag 'asoc-fix-v5.17-rc7' of https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound into for-linus
ASoC: Potential fixes for v5.17
Rather more fixes here than I'm comfortable with, we've had several
vendors noticing issues late in the release cycle all of which are valid
and reasonable fixes but it adds up to a much larger change set than I'd
like. Several of the AMD fixes look like cleanups from the subject
lines but are actually fixing user visible problems as well.
If you were to merge this for 5.18 rather than 5.17 it wouldn't be the
end of the world, stable will probably backport everything anyway.
Commit e2ae38cf3d ("vhost: fix hung thread due to erroneous iotlb
entries") tries to reject the IOTLB message whose size is zero. But
the size is not necessarily meaningful, one example is the batching
hint, so the commit breaks that.
Fixing this be reject zero size message only if the message is used to
update/invalidate the IOTLB.
Fixes: e2ae38cf3d ("vhost: fix hung thread due to erroneous iotlb entries")
Reported-by: Eli Cohen <elic@nvidia.com>
Cc: Anirudh Rayabharam <mail@anirudhrb.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Link: https://lore.kernel.org/r/20220310075211.4801-1-jasowang@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Eli Cohen <elic@nvidia.com>
Remove dev_err() messages after platform_get_irq*() failures.
platform_get_irq() already prints an error.
Generated by: scripts/coccinelle/api/platform_get_irq.cocci
Signed-off-by: Yihao Han <hanyihao@vivo.com>
Link: https://lore.kernel.org/r/20220310094806.13734-1-hanyihao@vivo.com
Signed-off-by: Mark Brown <broonie@kernel.org>
There is a spelling mistake in a dev_info message. Fix it.
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Reviewed-by: Codrin Ciubotariu <codrin.ciubotariu@microchip.com>
Link: https://lore.kernel.org/r/20220309164116.178685-1-colin.i.king@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Fix to return a negative error code from the error handling case instead
of 0, as done elsewhere in this function.
Fixes: 7e7292dba2 ("ASoC: fsl: add imx-es8328 machine driver")
Signed-off-by: Wang Wensheng <wangwensheng4@huawei.com>
Link: https://lore.kernel.org/r/20220310091902.129299-1-wangwensheng4@huawei.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The print function dev_err() is redundant because
platform_get_irq() already prints an error.
Eliminate the follow coccicheck warning:
./sound/soc/atmel/mchp-pdmc.c:991:2-9: line 991 is redundant because
platform_get_irq() already prints an error.
Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com>
Reviewed-by: Codrin Ciubotariu <codrin.ciubotariu@microchip.com>
Link: https://lore.kernel.org/r/20220310082756.1183-1-jiapeng.chong@linux.alibaba.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The TX clock source may be changed in next case, need to
disable it when stop, otherwise the TX may not work after
changing the clock source, error log is:
aplay: pcm_write:2058: write error: Input/output error
Fixes: a2388a498a ("ASoC: fsl: Add S/PDIF CPU DAI driver")
Signed-off-by: Shengjiu Wang <shengjiu.wang@nxp.com>
Reviewed-by: Fabio Estevam <festevam@gmail.com>
Link: https://lore.kernel.org/r/1646879863-27711-1-git-send-email-shengjiu.wang@nxp.com
Signed-off-by: Mark Brown <broonie@kernel.org>
One fix for type conversion issues when working out maximum
scatter/gather segment sizes which caused problems for some systems
which where the limits overflow due to the type conversion.
-----BEGIN PGP SIGNATURE-----
iQEzBAABCgAdFiEEreZoqmdXGLWf4p/qJNaLcl1Uh9AFAmIp6voACgkQJNaLcl1U
h9C9egf/ZApR13YUalLMx/aziI3NZhrv8vTfGVaD5s2szX6POVB8oxdxbSO6jAgU
3i0o7sYB1RvG7ncoHR+2FaE5ZI/DQg0LFNeAJe4pqRSavHIlgKf5htx3+B9tzyZa
SHPz1CBEM1rES3vLvWaM+1sxhildwTJeSO2HmTJusFcTYOu/dsAEJoEScj4/I6jU
UE3LJU7G8Zon4Xrc1pEmQOnyYxJAc1/JXXTyLayXYCfItTOfIEapforJDBu9/oBF
BwOUF41d0G8NxWP9cXXi7eti6G5rVmAeVMBP9wg5ibqG4w9PKTXa/LBcVZK9/cQn
QttHxgPDtlDZbY9dZc+U07WaVHPZnA==
=xDqO
-----END PGP SIGNATURE-----
Merge tag 'spi-fix-v5.17-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi
Pull spi fix from Mark Brown:
"One fix for type conversion issues when working out maximum
scatter/gather segment sizes.
It caused problems for some systems where the limits overflow
due to the type conversion"
* tag 'spi-fix-v5.17-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi:
spi: Fix invalid sgs value
The kernel test robot discovered that building without
HARDEN_BRANCH_PREDICTOR issues a warning due to a missing
argument to pr_info().
Add the missing argument.
Reported-by: kernel test robot <lkp@intel.com>
Fixes: 9dd78194a3 ("ARM: report Spectre v2 status through sysfs")
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
- fix a probe failure for Tegra241 GPIO controller in gpio-tegra186
- revert changes that caused a regression in the sysfs user-space interface
- correct the debounce time conversion in GPIO ACPI
- statify a struct in gpio-sim and fix a typo
- update registers in correct order (hardware quirk) in gpio-ts4900
-----BEGIN PGP SIGNATURE-----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=GQmw
-----END PGP SIGNATURE-----
Merge tag 'gpio-fixes-for-v5.17' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux
Pull gpio fixes from Bartosz Golaszewski:
- fix a probe failure for Tegra241 GPIO controller in gpio-tegra186
- revert changes that caused a regression in the sysfs user-space
interface
- correct the debounce time conversion in GPIO ACPI
- statify a struct in gpio-sim and fix a typo
- update registers in correct order (hardware quirk) in gpio-ts4900
* tag 'gpio-fixes-for-v5.17' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux:
gpio: sim: fix a typo
gpio: ts4900: Do not set DAT and OE together
gpio: sim: Declare gpio_sim_hog_config_item_ops static
gpiolib: acpi: Convert ACPI value of debounce to microseconds
gpio: Revert regression in sysfs-gpio (gpiolib.c)
gpio: tegra186: Add IRQ per bank for Tegra241
Merge series from Ranjani Sridharan <ranjani.sridharan@linux.intel.com>:
In preparation for adding support for the new IPC version that has been
introduced in the SOF firmware, this patch set includes some clean ups
and necessary modifications to commonly used functions that will be
re-used across different IPC-specific code.
Merge series from Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>:
This patchset adds a number of improvements for ES8336-based Intel
platforms, which are not well supported at all in Linux. Since
Christmas 2021, we've seen dozens of reports of broken audio [1].
The fundamental problem is that those platforms were built for Windows
but using an I2S codec - instead of the HDaudio traditional
solution. As a result, we are missing all the usual information needed
to configure the audio card (which I2S, what configuration, DMICs or
not, etc). The situation is similar to Baytrail with all possible
permutations enabled.
Some of the information can be discovered by checking the contents of
the 'NHLT' ACPI table. This helps discover at run-time which SSP to
use, and the number of microphones present. This NHLT-based solution
helps remove quirks that were added earlier.
Unfortunately, there are still a number of platform properties that
are not described by ACPI, just as GPIOs used for speakers, jack
detection inversion, etc. For some case, quirks are still provided in
the machine drivers.
Additional work will likely be needed, e.g. to detect which MCLK needs
to be used, refine the UCM settings, add the ES8326 codec driver, but
this is a first-step towards an 'out of the box' experience on Intel
platforms.
This patchset touches the sound/hda/intel-nhlt parts but should IMHO
be merged in the ASoC tree.
I would like to acknowledge the help of Nikolai Kostrigin, Mauro
Carvalho Chehab, Huajun Li, David Yang (@yangxiaohua2009) and other
GitHub testers.
[1] https://github.com/thesofproject/linux/issues?q=is%3Aissue+is%3Aopen+label%3A%22codec+ES8336%22
It was added in commit 766121ba5d ("arm64/mte: Add userspace interface
for enabling asymmetric mode").
Signed-off-by: Evgenii Stepanov <eugenis@google.com>
Reviewed-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20220309215943.87831-1-eugenis@google.com
Signed-off-by: Will Deacon <will@kernel.org>
AIC2 changes the IRQ fwspec to add a cell. Always use the second-to-last
cell for the MSI handling, so it will work for both AIC1 and AIC2 devices.
Signed-off-by: Hector Martin <marcan@marcan.st>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20220309192123.152028-2-marcan@marcan.st
struct dp_link_dpll is a pointless wrapper around struct dpll.
Just store the desired link rate into struct dpll::dot and
we're done.
v2: Document the full divider as a proper decimal number on chv
Nuke bogus eDP 1.4 comments for chv while at it
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220307233940.4161-9-ville.syrjala@linux.intel.com
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Remove the comment specifying the exact formulat for calculating
the DPLL frequency from the *_find_best_dpll() functions. Each
platform variant has its own way to calculate these and we have
the code already to do that. These comments are entirely redundant
and often even wrong so just get rid of them.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220307233940.4161-3-ville.syrjala@linux.intel.com
Acked-by: Jani Nikula <jani.nikula@intel.com>
Unify vlv/chv with earlier platforms so that the sturct dpll::dot
represents the /5 clock frequency (ie. DP symbol rate or HDMI
TMDS rate) rather than the *5 fast clock (/2 of the bitrate).
Makes life a little less confusing to get the same number back
in .dot which we fed into the DPLL algorithm.
v2: Actually just include the 5x in the final P divider
Do the same change to the hand rolled gvt code
v3: Missed a few *5 in *_find_best_dpll()
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220309214301.22899-1-ville.syrjala@linux.intel.com
When using bigjoiner it's useful to know the offset of each
individual pipe in the whole set of joined pipes. Let's include
that information in our PIPESRC rectangle. With this we can make
the plane clipping code blissfully unaware of bigjoiner usage, as
all we have to do is remove the pipe's offset from the final plane
destination coordinates.
v2: Use intel_bigjoiner_num_pipes()
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220223131315.18016-14-ville.syrjala@linux.intel.com
Reviewed-by: Manasi Navare <manasi.d.navare@intel.com>
Just noticed this when applying Andy's patch. s/childred/children/
Fixes: cb8c474e79 ("gpio: sim: new testing module")
Signed-off-by: Bartosz Golaszewski <brgl@bgdev.pl>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
This works around an issue with the hardware where both OE and
DAT are exposed in the same register. If both are updated
simultaneously, the harware makes no guarantees that OE or DAT
will actually change in any given order and may result in a
glitch of a few ns on a GPIO pin when changing direction and value
in a single write.
Setting direction to input now only affects OE bit. Setting
direction to output updates DAT first, then OE.
Fixes: 9c6686322d ("gpio: add Technologic I2C-FPGA gpio support")
Signed-off-by: Mark Featherston <mark@embeddedTS.com>
Signed-off-by: Kris Bahnsen <kris@embeddedTS.com>
Signed-off-by: Bartosz Golaszewski <brgl@bgdev.pl>
An emparassing typo: allocating a pointer instead of the object
pointed to. No harm done, as the pointer is large enough for
what we are using the object for, but still...
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20220310050238.4478-1-guozhengkui@vivo.com
Add support for ABE CAN Debugger devices using the gs_usb driver.
Link: https://lore.kernel.org/all/20220309124132.291861-22-mkl@pengutronix.de
Signed-off-by: Ben Evans <benny.j.evans92@gmail.com>
Signed-off-by: Peter Fink <pfink@christ-es.de>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
For example CANext FD needs to distinguish between bittiming constants
valid for arbitration phase and data phase to reach maximum
performance at higher speeds.
Link: https://lore.kernel.org/all/20220309124132.291861-20-mkl@pengutronix.de
Signed-off-by: Peter Fink <pfink@christ-es.de>
Signed-off-by: Christoph Möhring <cmoehring@christ-es.de>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
The CANtact Pro from LinkLayer Labs is based on the LPC54616 µC, which
is affected by the NXP LPC USB transfer erratum. However, the current
firmware (version 2) doesn't set the
GS_CAN_FEATURE_REQ_USB_QUIRK_LPC546XX bit.
This patch sets the feature GS_CAN_FEATURE_REQ_USB_QUIRK_LPC546XX to
workaround this issue.
For the GS_USB_BREQ_DATA_BITTIMING USB control message the CANtact Pro
firmware uses a request value, which is already used by the
candleLight firmware for a different purpose
(GS_USB_BREQ_GET_USER_ID).
This patch set the feature GS_CAN_FEATURE_QUIRK_BREQ_CANTACT_PRO to
workaround this issue.
Link: https://lore.kernel.org/all/20220309124132.291861-19-mkl@pengutronix.de
Cc: Ryan Edwards <ryan.edwards@gmail.com>
Signed-off-by: Peter Fink <pfink@christ-es.de>
Signed-off-by: Christoph Möhring <cmoehring@christ-es.de>
[mkl: improve check for CANtact Pro and add GS_CAN_FEATURE_QUIRK_BREQ_CANTACT_PRO quirk]
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
For the GS_USB_BREQ_DATA_BITTIMING USB control message the CANtact Pro
firmware uses a request value, which is already used by the
candleLight firmware for a different
purpose (GS_USB_BREQ_GET_USER_ID).
This patch adds a quirk to use the CANtact Pro's value for the
GS_USB_BREQ_DATA_BITTIMING USB control message instead of the official
one.
Link: https://lore.kernel.org/all/20220309124132.291861-18-mkl@pengutronix.de
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Introduce a workaround for a NXP chip errata on LPC546xx
controllers (Errata sheet LPC546xx / USB.15).
According to the document corruption can occur when the following
conditions are met:
* A TX (IN) transfer happens after a RX (OUT) transfer.
* The RX (OUT) transfer length is 4 + N * 16 (N >= 0) bytes.
Even though the struct gs_host_frame has a size of 76 bytes for a FD
frame, which does not apply to the above rule, corruption could be
seen.
Adding a dummy byte to break the second condition also on transfer
lengths with 4 + N * 8 bytes reliably circumvents USB transfer data
corruption.
The firmware can now request this quirk by setting
GS_CAN_FEATURE_REQ_USB_QUIRK_LPC546XX.
Link: https://lore.kernel.org/all/20220309124132.291861-17-mkl@pengutronix.de
Signed-off-by: Peter Fink <pfink@christ-es.de>
Signed-off-by: Christoph Möhring <cmoehring@christ-es.de>
Signed-off-by: Alexander Schartner <aschartner@christ-es.de>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
CANtact Pro from Linklayer is the first gs_usb compatible device
supporting CAN-FD with a different HW and re-written candlelight FW.
Support for CAN-FD is indicated by the device setting the
GS_CAN_FEATURE_FD flag. CAN-FD support is requested by the driver with
the GS_CAN_MODE_FD flag. The CAN-FD specific data bit timing
parameters are set with the GS_USB_BREQ_DATA_BITTIMING control
message.
This patch is based on the Eric Evenchick's gs_usb_fd driver (which
itself is a fork of gs_usb). The gs_usb_fd code base was reintegrated
into the gs_usb driver, and reworked to not break the existing
classical-CAN only hardware.
Link: https://lore.kernel.org/all/20220309124132.291861-16-mkl@pengutronix.de
Link: https://github.com/linklayer/gs_usb_fd/issues/2
Co-developed-by: Eric Evenchick <eric@evenchick.com>
Signed-off-by: Eric Evenchick <eric@evenchick.com>
Signed-off-by: Peter Fink <pfink@christ-es.de>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Modify struct gs_host_frame to make use of a union and
DECLARE_FLEX_ARRAY to be able to store different data (lengths), which
will be added in later commits.
Store the gs_host_frame length in TX direction (host -> device) in
struct gs_can::hf_size_tx and RX direction (device -> host) in struct
gs_usb::hf_size_rx so it must be calculated only once.
Link: https://lore.kernel.org/all/20220309124132.291861-15-mkl@pengutronix.de
Signed-off-by: Peter Fink <pfink@christ-es.de>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Some STM32G3 chips support up to 3 CAN-FD channels, increase number of
supported channels in this driver to 3 accordingly.
Link: https://lore.kernel.org/all/20220309124132.291861-14-mkl@pengutronix.de
Suggested-by: Ryan Edwards <ryan.edwards@gmail.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Introduce the variable udev in the gs_usb_probe() function to hold a
pointer to the struct usb_device. This avoids recalculating the value
several times in this function.
Link: https://lore.kernel.org/all/20220309124132.291861-13-mkl@pengutronix.de
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
