If the current code encounters a new type of device on a SoundWire
link, it will abort processing that link and move onto the next
link. However, there is no reason to disallow this setup, it would
appear this was being disallowed to work around issues introduced
by only the first endpoint on each link being checked, which is now
fixed.
The device type shouldn't determine which DAI link it is connected to,
the group ID and aggregation status should.
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20230808132013.889419-11-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The current code checks the first device on a link and assumes
that all the other devices on the link will have the same endpoint
aggregation status and endpoint group ID.
Say for example a system looked like:
SDW0 - Amp 1 (Aggregated, Group 1), Mic 1 (Aggregated, Group 2)
SDW1 - Amp 2 (Aggregated, Group 1), Mic 2 (Aggregated, Group 2)
The current code would create the DAI link for the aggregated amps,
although it is worth noting that the only reason Mic 2 is not added is
the additional check that aborts processing the link when the device
changes. Then when processing the DAI link for the microphones, Mic
2 would not be added, as the check will only be done on the first
device, which would be Amp 2 and thus the wrong group, causing the
whole link to be skipped.
Move the endpoint check to be for each device rather than the first
device on each link.
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20230808132013.889419-10-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The current loops at the top of create_sdw_dailink process the devices
on each link starting from device index adr_index. But adr_index is only
meaningful on the first on these SoundWire links, as it is the index of
the current device on that link. This means devices will be skipped on
later links.
Say for example the system looks like this:
SDW0 - Codec (Not Aggregated), Amp 1 (Aggregated, Group 1)
SDW1 - Amp 2 (Aggregated, Group 1), Amp 3 (Aggregated, Group 1)
The code should create 2 DAI links, one for the CODEC and one for the
aggregated amps. It will create the DAI link for the codec no problem.
When it creates the DAI link for Group 1 however, create_sdw_dailink
will be called with an adr_index of 1, since that is the index of Amp
1 on SDW0. However, as the loop in create_sdw_dailink moves onto SDW1
it will again start from adr_index, skipping Amp 2. Resulting in the amp
DAI link only have amps 1 and 3 in it.
It is reasonable to start at adr_index on the first link, since
earlier devices have by definition already been processed. However,
update the code when processing later links to handle all devices.
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20230808132013.889419-9-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
There are two problems with the current range check on the codec_conf
array.
Firstly, adr_link_next->num_adr refers to the number of devices
on the current SoundWire link, but adr_index refers to the first
SoundWire link involved in the DAI link. This means that subtracting
these two numbers is only meaningful on the first SoundWire link in the
DAI and broken on later links.
Secondly, the intention of the range check is to add the number
of remaining devices on the currently link to the current index
and ensure enough space remains. However, this assumes that all
remaining devices on the SoundWire link will be added to the current
DAI link. Ideally this would not be the case, and devices could be
grouped as the user desired.
Moving the range check into the inner loop both simplifies the code (no
need to add and subtract offsets) and allows future refactoring such
that devices on a single SoundWire link don't have to all be grouped onto
a single DAI link. The check will be processed slightly more often since
it is processed for each device rather each link but this is probe time
and the numbers involved are very small here (4 links, likely no more
than 2-4 devices per link).
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20230808132013.889419-8-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
In create_sdw_dailink, rather than bulk updating the index into the
DAI link components array, at the end of processing a link, do so each
time the code adds a new component. This simplifies things slightly,
as an intermediate variable is no longer needed to track the current
place in the DAI link components array.
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20230808132013.889419-7-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The loops which fill the codec DAI link component structures are split
across create_sdw_dailink and create_codec_dai_name. This causes the
code to be rather confusing, needing to return out the function to allow
the upper loop to iterate. Remove the create_codec_dai_name helper and
pull its code up into create_sdw_dailink, this makes it more obvious
what is happening in the code. This patch makes no functional change
just hoists the code up a level.
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20230808132013.889419-6-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Add a helper function to create a single codec DAI link component
structure. This sets things up for more refactoring of the creating of
the DAI links.
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20230808132013.889419-5-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Move the check for a valid group id into get_dailink_info as
well. This does cause a slight change in behaviour in that the system
will return an error rather than just ignoring the link with an
invalid group id. There are presently no systems with invalid group
ids in mainline and failing seems more appropriate since it will
better highlight the code needs fixing.
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20230808132013.889419-4-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
As get_dailink_info spins through all the links anyway simply check the
link masks there. This saves an extra check and means the code will
fail earlier if the mask is invalid.
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20230808132013.889419-3-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
get_dailink_info already checked if the adr_link pointer was NULL so
there is no need to recheck later in sof_card_dai_links_create.
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20230808132013.889419-2-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Add the missing new lines.
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Link: https://lore.kernel.org/r/20230808132013.889419-1-ckeepax@opensource.cirrus.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The module_rpmsg_driver() will set "THIS_MODULE" to driver.owner when
register a rpmsg_driver driver, so it is redundant initialization to set
driver.owner in the statement. Remove it for clean code.
Signed-off-by: Li Zetao <lizetao1@huawei.com>
Link: https://lore.kernel.org/r/20230808021728.2978035-1-lizetao1@huawei.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Use GFP_KERNEL instead of GFP_ATOMIC while reading a binary policy in
sens_read() and cat_read(), similar to surrounding code.
Signed-off-by: Christian Göttsche <cgzones@googlemail.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
close(2) is a special case which guarantees a shallow kernel stack,
making delegation to task_work machinery unnecessary. Said delegation is
problematic as it involves atomic ops and interrupt masking trips, none
of which are cheap on x86-64. Forcing close(2) to do it looks like an
oversight in the original work.
Moreover presence of CONFIG_RSEQ adds an additional overhead as fput()
-> task_work_add(..., TWA_RESUME) -> set_notify_resume() makes the
thread returning to userspace land in resume_user_mode_work(), where
rseq_handle_notify_resume takes a SMAP round-trip if rseq is enabled for
the thread (and it is by default with contemporary glibc).
Sample result when benchmarking open1_processes -t 1 from will-it-scale
(that's an open + close loop) + tmpfs on /tmp, running on the Sapphire
Rapid CPU (ops/s):
stock+RSEQ: 1329857
stock-RSEQ: 1421667 (+7%)
patched: 1523521 (+14.5% / +7%) (with / without rseq)
Patched result is the same regardless of rseq as the codepath is avoided.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Andi reported (see link below) a regression when printing the
'duration_time' tool event, where it gets printed as "not counted" for
most of the CPUs, fix it by skipping zero counts for tool events.
Reported-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Andi Kleen <ak@linux.intel.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Athira Rajeev <atrajeev@linux.vnet.ibm.com>
Cc: Claire Jensen <cjense@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/all/ZMlrzcVrVi1lTDmn@tassilo/
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Use a simple logical shift and increment to calculate the number of slots
taken by the DMA segment boundary.
At least GCC-13 is not able to optimize the expression, producing this
horrible assembly code on x86:
cmpq $-1, %rcx
je .L364
addq $2048, %rcx
shrq $11, %rcx
movq %rcx, %r13
.L331:
// rest of the function here...
// after function epilogue and return:
.L364:
movabsq $9007199254740992, %r13
jmp .L331
After the optimization, the code looks more reasonable:
shrq $11, %r11
leaq 1(%r11), %rbx
Signed-off-by: Petr Tesarik <petr.tesarik.ext@huawei.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Move the comment down in front of the loop that actually sets the list
member of struct io_tlb_slot to zero.
Fixes: 26a7e09478 ("swiotlb: refactor swiotlb_tbl_map_single")
Signed-off-by: Petr Tesarik <petr.tesarik.ext@huawei.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
After commit f22f9aaf6c ("selinux: remove the runtime disable
functionality"), the comment on selinux_hooks[] is out-of-date,
remove the last paragraph about runtime disable functionality.
Signed-off-by: Xiu Jianfeng <xiujianfeng@huawei.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
microcode_mutex is only used by reload_store(). It has a comment saying
"to synchronize with each other". Other user of this mutex have been
removed in the commits
181b6f40e9 ("x86/microcode: Rip out the OLD_INTERFACE").
b6f86689d5 ("x86/microcode: Rip out the subsys interface gunk")
The sysfs interface does not need additional synchronisation vs itself
because it is provided as kernfs_ops::mutex which is acquired in
kernfs_fop_write_iter().
Remove the superfluous microcode_mutex.
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Reviewed-by: Sohil Mehta <sohil.mehta@intel.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20230804075853.JF_n6GXC@linutronix.de
- Fix a freeze consistency check in gfs2_trans_add_meta().
- Don't use filemap_splice_read as it can cause deadlocks on gfs2.
-----BEGIN PGP SIGNATURE-----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=pPXf
-----END PGP SIGNATURE-----
Merge tag 'gfs2-v6.4-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2
Pull gfs2 fixes from Andreas Gruenbacher:
- Fix a freeze consistency check in gfs2_trans_add_meta()
- Don't use filemap_splice_read as it can cause deadlocks on gfs2
* tag 'gfs2-v6.4-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2:
gfs2: Don't use filemap_splice_read
gfs2: Fix freeze consistency check in gfs2_trans_add_meta
core_reg_fixup() complicates sharing the get-reg-list test with
other architectures. Rather than work at keeping it, with plenty
of #ifdeffery, just delete it, as it's unlikely to test a kernel
based on anything older than v5.2 with the get-reg-list test,
which is a test meant to check for regressions in new kernels.
(And, an older version of the test can still be used for older
kernels if necessary.)
Signed-off-by: Andrew Jones <ajones@ventanamicro.com>
Signed-off-by: Haibo Xu <haibo1.xu@intel.com>
Signed-off-by: Anup Patel <anup@brainfault.org>
Rename vcpu_config to vcpu_reg_list to be more specific and add
it to kvm_util.h. While it may not get used outside get-reg-list
tests, exporting it doesn't hurt, as long as it has a unique enough
name. This is a step in the direction of sharing most of the get-
reg-list test code between architectures.
Signed-off-by: Andrew Jones <ajones@ventanamicro.com>
Signed-off-by: Haibo Xu <haibo1.xu@intel.com>
Signed-off-by: Anup Patel <anup@brainfault.org>
print_reg() and its helpers only use the vcpu_config pointer for
config_name(). So just pass the config name in instead, which is used
as a prefix in asserts. print_reg() can now be compiled independently
of config_name().
Signed-off-by: Andrew Jones <ajones@ventanamicro.com>
Signed-off-by: Haibo Xu <haibo1.xu@intel.com>
Signed-off-by: Anup Patel <anup@brainfault.org>
The check doesn't prove much anyway, as the reg lists could be
messed up too. Just drop the check to simplify making print_reg
more independent.
Signed-off-by: Andrew Jones <ajones@ventanamicro.com>
Signed-off-by: Haibo Xu <haibo1.xu@intel.com>
Signed-off-by: Anup Patel <anup@brainfault.org>
The original author of aarch64/get-reg-list.c (me) was wearing
tunnel vision goggles when implementing str_with_index(). There's
no reason to have such a special case string function. Instead,
take inspiration from glib and implement strdup_printf. The
implementation builds on vasprintf() which requires _GNU_SOURCE,
but we require _GNU_SOURCE in most files already.
Signed-off-by: Andrew Jones <ajones@ventanamicro.com>
Signed-off-by: Haibo Xu <haibo1.xu@intel.com>
Signed-off-by: Anup Patel <anup@brainfault.org>
Commit f3fb16291f ("soc: starfive: Move the power-domain driver to the
genpd dir") moves:
drivers/{soc/starfive/jh71xx_pmu.c => genpd/starfive/jh71xx-pmu.c}
However, the update to the file entry in MAINTAINERS only addresses the
move in directories, but misses the renaming from jh71xx_pmu.c to
jh71xx-pmu.c. Hence, ./scripts/get_maintainer.pl --self-test=patterns
complains about a broken reference.
Adjust the file entry in STARFIVE JH71XX PMU CONTROLLER DRIVER.
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@gmail.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Buffers mapped with pgprot_writecombined() are not correctly
flushed. This triggers issues on VPU access using random
memory content such as MMU translation faults, invalid context
descriptors being fetched and can lead to VPU FW crashes.
Fixes: 647371a660 ("accel/ivpu: Add GEM buffer object management")
Cc: stable@vger.kernel.org # 6.3+
Signed-off-by: Karol Wachowski <karol.wachowski@linux.intel.com>
Reviewed-by: Stanislaw Gruszka <stanislaw.gruszka@linux.intel.com>
Signed-off-by: Stanislaw Gruszka <stanislaw.gruszka@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230802063735.3005291-1-stanislaw.gruszka@linux.intel.com
The bd71815 regulator driver includes the legacy header
<linux/gpio.h> for no reason, it is already using the proper
<linux/gpio/consumer.h> include. Drop the include.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20230808-descriptors-regulator-v1-11-939b5e84dd18@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
The bd71828 includes the legacy header <linux/gpio.h> for no
reason, drop the include. The documentation mentions GPIO but there
is no usage of the GPIO namespace.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20230808-descriptors-regulator-v1-10-939b5e84dd18@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
The max20086 regulator driver includes the legacy header
<linux/gpio.h> for no reason, it is already using the proper
<linux/gpio/consumer.h> include. Drop the include.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20230808-descriptors-regulator-v1-8-939b5e84dd18@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
The mcp16502 regulator driver includes the legacy header
<linux/gpio.h> for no reason, it is already using the proper
<linux/gpio/consumer.h> include. Drop the include.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20230808-descriptors-regulator-v1-7-939b5e84dd18@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
The RPI panel regulator driver includes the legacy header
<linux/gpio.h> for no reason, this is a driver and <linux/gpio/driver.h>
is already included. Drop the include.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20230808-descriptors-regulator-v1-3-939b5e84dd18@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Document the System Control registers regions found on all Amlogic
SoC families and it's clock, power, pinctrl and phy subnodes.
The regions has various independent registers tied to other
hardware devices, thus the syscon compatible.
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Neil Armstrong <neil.armstrong@linaro.org>
Link: https://lore.kernel.org/r/20230706-b4-amlogic-bindings-convert-take2-v3-3-f63de6f12dcc@linaro.org
Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
A switch from OSI to PC mode is only possible if all CPUs other than the
calling one are OFF, either through a call to CPU_OFF or not yet booted.
Currently OSI mode is enabled before power domains are created. In cases
where CPUidle states are not using hierarchical CPU topology the bail out
path tries to switch back to PC mode which gets denied by firmware since
other CPUs are online at this point and creates inconsistent state as
firmware is in OSI mode and Linux in PC mode.
This change moves enabling OSI mode after power domains are created,
this would makes sure that hierarchical CPU topology is used before
switching firmware to OSI mode.
Cc: stable@vger.kernel.org
Fixes: 70c179b498 ("cpuidle: psci: Allow PM domain to be initialized even if no OSI mode")
Signed-off-by: Maulik Shah <quic_mkshah@quicinc.com>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Genpd parent and child domain topology created using dt_idle_pd_init_topology()
needs to be removed during error cases.
Add new helper function dt_idle_pd_remove_topology() for same.
Cc: stable@vger.kernel.org
Reviewed-by: Ulf Hanssson <ulf.hansson@linaro.org>
Signed-off-by: Maulik Shah <quic_mkshah@quicinc.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>