Add DT bindings for the peripheral clock controller of the Amlogic T7
SoC family.
Signed-off-by: Jian Hu <jian.hu@amlogic.com>
Reviewed-by: Rob Herring (Arm) <robh@kernel.org>
Link: https://lore.kernel.org/r/20251212022619.3072132-4-jian.hu@amlogic.com
Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Add DT bindings for the SCMI clock controller of the Amlogic T7 SoC family.
Signed-off-by: Jian Hu <jian.hu@amlogic.com>
Acked-by: Rob Herring (Arm) <robh@kernel.org>
Link: https://lore.kernel.org/r/20251212022619.3072132-3-jian.hu@amlogic.com
Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Currently it seems impossible to observe these changes to the file's
attributes. It's useful to be able to do this to see when the file
becomes immutable, for example, so emit IN_ATTRIB via fsnotify_xattr(),
like when changing other inode attributes.
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Link: https://patch.msgid.link/iyvn6qjotpu6cei5jdtsoibfcp6l6rgvn47cwgaucgtucpfy2s@tarta.nabijaczleweli.xyz
Signed-off-by: Jan Kara <jack@suse.cz>
inotify/fanotify do not allow users with no read access to a file to
subscribe to events (e.g. IN_ACCESS/IN_MODIFY), but they do allow the
same user to subscribe for watching events on children when the user
has access to the parent directory (e.g. /dev).
Users with no read access to a file but with read access to its parent
directory can still stat the file and see if it was accessed/modified
via atime/mtime change.
The same is not true for special files (e.g. /dev/null). Users will not
generally observe atime/mtime changes when other users read/write to
special files, only when someone sets atime/mtime via utimensat().
Align fsnotify events with this stat behavior and do not generate
ACCESS/MODIFY events to parent watchers on read/write of special files.
The events are still generated to parent watchers on utimensat(). This
closes some side-channels that could be possibly used for information
exfiltration [1].
[1] https://snee.la/pdf/pubs/file-notification-attacks.pdf
Reported-by: Sudheendra Raghav Neela <sneela@tugraz.at>
CC: stable@vger.kernel.org
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Signed-off-by: Jan Kara <jack@suse.cz>
When pdacf_config() fails, snd_pdacf_probe() returns the error code
directly without freeing the sound card resources allocated by
snd_card_new(), which leads to a memory leak.
Add proper error handling to free the sound card and clear the card
list entry when pdacf_config() fails.
Fixes: 15b99ac172 ("[PATCH] pcmcia: add return value to _config() functions")
Suggested-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Haotian Zhang <vulab@iscas.ac.cn>
Link: https://patch.msgid.link/20251215090433.211-1-vulab@iscas.ac.cn
Signed-off-by: Takashi Iwai <tiwai@suse.de>
TID getting from ieee80211_get_tid() might be out of range of array size
of sta_entry->tids[], so check TID is less than MAX_TID_COUNT. Othwerwise,
UBSAN warn:
UBSAN: array-index-out-of-bounds in drivers/net/wireless/realtek/rtlwifi/rtl8192cu/trx.c:514:30
index 10 is out of range for type 'rtl_tid_data [9]'
Fixes: 8ca4cdef93 ("wifi: rtlwifi: rtl8192cu: Fix TX aggregation")
Signed-off-by: Morning Star <alexbestoso@gmail.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/1764232628-13625-1-git-send-email-pkshih@realtek.com
The indirect IO is necessary for RTL8822CS, but not necessary for other
chips. Otherwiese, it throws errors and becomes unusable.
rtw88_8723cs mmc1:0001:1: WOW Firmware version 11.0.0, H2C version 0
rtw88_8723cs mmc1:0001:1: Firmware version 11.0.0, H2C version 0
rtw88_8723cs mmc1:0001:1: sdio read32 failed (0xf0): -110
rtw88_8723cs mmc1:0001:1: sdio write8 failed (0x1c): -110
rtw88_8723cs mmc1:0001:1: sdio read32 failed (0xf0): -110
By vendor driver, only RTL8822CS and RTL8822ES need indirect IO, but
RTL8822ES isn't supported yet. Therefore, limit it to RTL8822CS only.
Reported-by: Andrey Skvortsov <andrej.skvortzov@gmail.com>
Closes: https://lore.kernel.org/linux-wireless/07a32e2d6c764eb1bd9415b5a921a652@realtek.com/T/#m997b4522f7209ba629561c776bfd1d13ab24c1d4
Fixes: 58de1f91e0 ("wifi: rtw88: sdio: use indirect IO for device registers before power-on")
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Tested-by: Andrey Skvortsov <andrej.skvortzov@gmail.com>
Link: https://patch.msgid.link/1764034729-1251-1-git-send-email-pkshih@realtek.com
When vxpocket_config() fails, vxpocket_probe() returns the error code
directly without freeing the sound card resources allocated by
snd_card_new(), which leads to a memory leak.
Add proper error handling to free the sound card and clear the
allocation bit when vxpocket_config() fails.
Fixes: 15b99ac172 ("[PATCH] pcmcia: add return value to _config() functions")
Signed-off-by: Haotian Zhang <vulab@iscas.ac.cn>
Link: https://patch.msgid.link/20251215042652.695-1-vulab@iscas.ac.cn
Signed-off-by: Takashi Iwai <tiwai@suse.de>
CONFIG_OMAP_USB2 is already enabled as a module in the default defconfig
since commit 8a703a728a ("arm64: defconfig: Enable USB2 PHY Driver").
Remove the duplicate entry to fix the following warning:
arch/arm64/configs/defconfig:1705:warning: override: reassigning to symbol OMAP_USB2
Fixes: 91fe3315cdf9f ("arm64: defconfig: Enable missing AMD/Xilinx drivers")
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/20251015150728.118296-1-prabhakar.mahadev-lad.rj@bp.renesas.com
Signed-off-by: Michal Simek <michal.simek@amd.com>
Over years number of upstream drivers have grown for AMD/Xilinx SOCs
(ZynqMP, Versal, Versal NET) but they are not enabled by default in
defconfig that's why enable all drivers for these SOCs including USB5744
on board USB hub available on Kria ZynqMP based SOMs and Carrier Cards.
Signed-off-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/457c3a128e300241afd20da693d1d80a35d1ece6.1758099050.git.michal.simek@amd.com
Since the following commit, the zynqmp clk driver uses the common
divider_round_rate() when determining the appropriate clock divider for a
requested clock frequency:
1fe15be1fb
This means that all the calculations will be in kHz when determining the
appropriate clock divider for a given cpufreq request. The problem with this
is that the zynqmp.dtsi and zynqmp-clk-ccf.dtsi files have frequency
definitions in Hz, so when dividing requested values in kHz, errors can occur
with the rounding.
For example, the current pss_ref_clk frequency is 33333333 Hz which generates
a cpufreq parent clock frequency of 1199999988 Hz which is the same as the
highest opp-table-cpu frequency in the zynqmp.dtsi.
But if a user requests the value 1199999 kHz as recommended in the available
frequencies:
root@zynqmp:/sys/kernel/tracing# cat /sys/devices/system/cpu/cpufreq/policy0/scaling_available_frequencies
299999 399999 599999 1199999
root@zynqmp:/ # echo 1199999 > /sys/devices/system/cpu/cpufreq/policy0/scaling_setspeed
The calculation will be:
1199999988 / 1199999000 = 1.000001
This will get rounded up to a divider value of 2 giving the following result.
root@zynqmp:/ # cat /sys/devices/system/cpu/cpufreq/policy0/cpuinfo_cur_freq
599999
Also, if a user tries to work around this calculation by using any larger
values, it still will not fix the problem because the driver will use the
largest opp in kHz which leads to the same calculation error.
User requests 1200000
root@zynqmp:/ # echo 1200000 > /sys/devices/system/cpu/cpufreq/policy0/scaling_setspeed
Driver converts any value greater than 1199999 to the largest opp which is
1199999 and then calculates the divider value with the same calculation.
The calculation will still be:
1199999988 / 1199999000 = 1.000001
This will get rounded up to a divider value of 2 giving the following result.
root@zynqmp:/ # cat /sys/devices/system/cpu/cpufreq/policy0/cpuinfo_cur_freq
599999
This means there is no way to configure the zynqmp for the fastest opp using
the current dtsi files.
To fix this issue, this patch updates the zynqmp opp-table-cpu and
pss_ref_clk, so the clock rates are calculated correctly.
root@zynqmp:/sys/kernel/tracing# cat /sys/devices/system/cpu/cpufreq/policy0/scaling_available_frequencies
300000 400000 600000 1200000
root@zynqmp:/ # echo 1200000 > /sys/devices/system/cpu/cpufreq/policy0/scaling_setspeed
root@zynqmp:/ # cat /sys/devices/system/cpu/cpufreq/policy0/cpuinfo_cur_freq
1200000
Signed-off-by: Neal Frager <neal.frager@amd.com>
Acked-by: Jay Buddhabhatti <jay.buddhabhatti@amd.com>
Signed-off-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/20251111070555.1169130-1-neal.frager@amd.com
Currently if a user enqueue a work item using schedule_delayed_work() the
used wq is "system_wq" (per-cpu wq) while queue_delayed_work() use
WORK_CPU_UNBOUND (used when a cpu is not specified). The same applies to
schedule_work() that is using system_wq and queue_work(), that makes use
again of WORK_CPU_UNBOUND.
This lack of consistentcy cannot be addressed without refactoring the API.
This patch continues the effort to refactor worqueue APIs, which has begun
with the change introducing new workqueues and a new alloc_workqueue flag:
commit 128ea9f6cc ("workqueue: Add system_percpu_wq and system_dfl_wq")
commit 930c2ea566 ("workqueue: Add new WQ_PERCPU flag")
system_dfl_wq should be the default workqueue so as not to enforce
locality constraints for random work whenever it's not required.
The old system_unbound_wq will be kept for a few release cycles.
Suggested-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Marco Crivellari <marco.crivellari@suse.com>
Signed-off-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/20251104103942.96647-1-marco.crivellari@suse.com
AMD/Xilinx Versal Window Watchdog Timer has its own power domain, so add
an optional property to describe it.
Signed-off-by: Harini T <harini.t@amd.com>
Acked-by: Conor Dooley <conor.dooley@microchip.com>
Signed-off-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/20250130120233.742879-1-harini.t@amd.com
Unlike zynq, which has a specific compatible for the Cadence spi
controller, zynqmp and versal-net do not have specific compatibles.
In order to "encourage" people to use soc-specific compatibles for new
devices using this IP, add specific compatibles for these devices, with
a fallback to the existing compatible for the r1p6 version of the IP so
that there will be no functional change.
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
Acked-by: Michal Simek <michal.simek@amd.com>
Link: https://lore.kernel.org/r/20251001-cheesy-shucking-c55431bbcae3@spud
Signed-off-by: Michal Simek <michal.simek@amd.com>
We have currently used up 30 out of the 32-bits in the struct ata_device
struct member quirks. Thus, it is only possible to add two more quirks.
Change the struct ata_device struct member quirks from an unsigned int to
an u64.
Doing this core level change now, will make it easier for us now, as we
will not need to also do core level changes once the final two bits are
used as well.
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Add a new libata.force parameter called max_sec.
The parameter can take an arbitrary value using the format:
libata.force=max_sec=<number of 512B sectors>
e.g. libata.force=max_sec=8191
or
libata.force=max_sec=2048
This will allow the user to set an arbitrary maximum command size
(dev->max_sectors) using libata.force.
We cannot remove the existing libata.force parameters "max_sec_128" and
"max_sec_1024", as these are a part of the exising user facing API.
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Currently, no libata.force parameter supports an arbitrary value.
All allowed values, e.g. udma/16, udma/25, udma/33, udma/44, udma/66,
udma/100, udma/133 have hardcoded entries in the force_tbl table.
Add code to allow a libata.force param with the format
libata.force=param=param_value, where param_value can be an arbitrary
value.
This code will be used in a follow up commit.
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Modify the existing libata.force parameters "max_sec_128" and
"max_sec_1024" to use the generic ATA_QUIRK_MAX_SEC quirk rather than
individual quirks.
This also allows us to remove the individual quirks ATA_QUIRK_MAX_SEC_128
and ATA_QUIRK_MAX_SEC_1024.
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Add ata_force_get_fe_for_dev() helper to get the struct ata_force_ent for
a struct ata_device.
Use the helper in ata_force_quirks().
The helper will also be used in follow up commits.
No functional change intended.
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Add a new quirk ATA_QUIRK_MAX_SEC, which has a separate table with device
specific values.
Convert all existing ATA_QUIRK_MAX_SEC_XXX device quirks in
__ata_dev_quirks to the new format.
Quirks ATA_QUIRK_MAX_SEC_128 and ATA_QUIRK_MAX_SEC_1024 cannot be removed
yet, since they are also used by libata.force, which functionally, is a
separate user of the quirks. The quirks will be removed once all users
have been converted to use the new format.
The quirk ATA_QUIRK_MAX_SEC_8191 can be removed since it has no equivalent
libata.force parameter.
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
When a SATA DAS enclosure is connected behind a Thunderbolt PCIe
switch, hot-unplugging the whole enclosure causes pciehp to tear down
the PCI hierarchy before the SCSI layer issues SYNCHRONIZE CACHE and
START STOP UNIT for the disks.
libata still queues these commands and the AHCI driver tries to access
the HBA registers even though the PCI channel is already offline. This
results in a series of timeouts and error recovery attempts, e.g.:
[ 824.778346] pcieport 0000:00:07.0: pciehp: Slot(14): Link Down
[ 891.612720] ata8.00: qc timeout after 5000 msecs (cmd 0xec)
[ 902.876501] ata8.00: qc timeout after 10000 msecs (cmd 0xec)
[ 934.107998] ata8.00: qc timeout after 30000 msecs (cmd 0xec)
[ 936.206431] sd 7:0:0:0: [sda] Synchronize Cache(10) failed:
Result: hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK
...
[ 1006.298356] ata1.00: qc timeout after 5000 msecs (cmd 0xec)
[ 1017.561926] ata1.00: qc timeout after 10000 msecs (cmd 0xec)
[ 1048.791790] ata1.00: qc timeout after 30000 msecs (cmd 0xec)
[ 1050.890035] sd 0:0:0:0: [sdb] Synchronize Cache(10) failed:
Result: hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK
With this patch applied, the same hot-unplug looks like:
[ 59.965496] pcieport 0000:00:07.0: pciehp: Slot(14): Link Down
[ 60.002502] sd 7:0:0:0: [sda] Synchronize Cache(10) failed:
Result: hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK
...
[ 60.103050] sd 0:0:0:0: [sdb] Synchronize Cache(10) failed:
Result: hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK
In this test setup with two disks, the hot-unplug sequence shrinks from
about 226 seconds (~3.8 minutes) between the Link Down event and the
last SYNCHRONIZE CACHE failure to under a second. Without this patch the
total delay grows roughly with the number of disks, because each disk
gets its own SYNCHRONIZE CACHE and qc timeout series.
If the underlying PCI device is already gone, these commands cannot
succeed anyway. Avoid issuing them by introducing
ata_adapter_is_online(), which checks pci_channel_offline() for
PCI-based hosts. It is used from ata_scsi_find_dev() to return NULL,
causing the SCSI layer to fail new commands with DID_BAD_TARGET
immediately, and from ata_qc_issue() to bail out before touching the
HBA registers.
Since such failures would otherwise trigger libata error handling,
ata_adapter_is_online() is also consulted from ata_scsi_port_error_handler().
When the adapter is offline, libata skips ap->ops->error_handler(ap) and
completes error handling using the existing path, rather than running
a full EH sequence against a dead adapter.
With this change, SYNCHRONIZE CACHE and START STOP UNIT commands
issued during hot-unplug fail quickly once the PCI channel is offline,
without qc timeout spam or long libata EH delays.
Suggested-by: Damien Le Moal <dlemoal@kernel.org>
Signed-off-by: Henry Tseng <henrytseng@qnap.com>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Commit 28a3fc2295 ("libata: implement ZBC IN translation") added
ata_scsi_report_zones_complete(). Since the beginning, this function
has disabled IRQs on the local CPU using local_irq_save().
qc->complete_fn is always called with ap->lock held, and the ap->lock
is always taken using spin_lock_irq*().
Thus, this local_irq_save() is superfluous and can be removed.
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
There's three layers of logic in the scheduler that
deal with 'has_blocked' (load) handling of the NOHZ code:
(1) nohz.has_blocked,
(2) rq->has_blocked_load, deal with NOHZ idle balancing,
(3) and cfs_rq_has_blocked(), which is part of the layer
that is passing the SMP load-balancing signal to the
NOHZ layers.
The 'has_blocked' and 'has_blocked_load' names are used
in a mixed fashion, sometimes within the same function.
Standardize on 'has_blocked_load' to make it all easy
to read and easy to grep.
No change in functionality.
Suggested-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Frederic Weisbecker <frederic@kernel.org>
Cc: Shrikanth Hegde <sshegde@linux.ibm.com>
Link: https://patch.msgid.link/aS6yvxyc3JfMxxQW@gmail.com
We have to be careful with vruntime comparisons and subtraction,
due to the possibility of wrapping, so we have macros like:
#define vruntime_gt(field, lse, rse) ({ (s64)((lse)->field - (rse)->field) > 0; })
Which is used like this:
if (vruntime_gt(min_vruntime, se, rse))
se->min_vruntime = rse->min_vruntime;
Replace this with an easier to read pattern that uses the regular
arithmetics operators:
if (vruntime_cmp(se->min_vruntime, ">", rse->min_vruntime))
se->min_vruntime = rse->min_vruntime;
Also replace vruntime subtractions with vruntime_op():
- delta = (s64)(sea->vruntime - seb->vruntime) +
- (s64)(cfs_rqb->zero_vruntime_fi - cfs_rqa->zero_vruntime_fi);
+ delta = vruntime_op(sea->vruntime, "-", seb->vruntime) +
+ vruntime_op(cfs_rqb->zero_vruntime_fi, "-", cfs_rqa->zero_vruntime_fi);
In the vruntime_cmp() and vruntime_op() macros use Use __builtin_strcmp(),
because of __HAVE_ARCH_STRCMP might turn off the compiler optimizations
we rely on here to catch usage bugs.
No change in functionality.
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The ::avg_vruntime field is a misnomer: it says it's an
'average vruntime', but in reality it's the momentary sum
of the weighted vruntimes of all queued tasks, which is
at least a division away from being an average.
This is clear from comments about the math of fair scheduling:
* \Sum (v_i - v0) * w_i := cfs_rq->avg_vruntime
This confusion is increased by the cfs_avg_vruntime() function,
which does perform the division and returns a true average.
The sum of all weighted vruntimes should be named thusly,
so rename the field to ::sum_w_vruntime. (As arguably
::sum_weighted_vruntime would be a bit of a mouthful.)
Understanding the scheduler is hard enough already, without
extra layers of obfuscated naming. ;-)
Also rename related helper functions:
sum_vruntime_add() => sum_w_vruntime_add()
sum_vruntime_sub() => sum_w_vruntime_sub()
sum_vruntime_update() => sum_w_vruntime_update()
With the notable exception of cfs_avg_vruntime(), which
was named accurately.
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://patch.msgid.link/20251201064647.1851919-7-mingo@kernel.org
The ::avg_load field is a long-standing misnomer: it says it's an
'average load', but in reality it's the momentary sum of the load
of all currently runnable tasks. We'd have to also perform a
division by nr_running (or use time-decay) to arrive at any sort
of average value.
This is clear from comments about the math of fair scheduling:
* \Sum w_i := cfs_rq->avg_load
The sum of all weights is ... the sum of all weights, not
the average of all weights.
To make it doubly confusing, there's also an ::avg_load
in the load-balancing struct sg_lb_stats, which *is* a
true average.
The second part of the field's name is a minor misnomer
as well: it says 'load', and it is indeed a load_weight
structure as it shares code with the load-balancer - but
it's only in an SMP load-balancing context where
load = weight, in the fair scheduling context the primary
purpose is the weighting of different nice levels.
So rename the field to ::sum_weight instead, which makes
the terminology of the EEVDF math match up with our
implementation of it:
* \Sum w_i := cfs_rq->sum_weight
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://patch.msgid.link/20251201064647.1851919-6-mingo@kernel.org
Join two identical #ifdef blocks:
#ifdef CONFIG_FAIR_GROUP_SCHED
...
#endif
#ifdef CONFIG_FAIR_GROUP_SCHED
...
#endif
Also mark nested #ifdef blocks in the usual fashion, to make
it more apparent where in a nested hierarchy of #ifdefs we
are at a glance.
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Shrikanth Hegde <sshegde@linux.ibm.com>
Link: https://patch.msgid.link/20251201064647.1851919-2-mingo@kernel.org
Merge series from Cezary Rojewski <cezary.rojewski@intel.com>:
Entire patchset provides no new features and does not alter the code
from functional (user) perspective.
The first two improve IPC-error logging 'mechanism' and align the
catpt-driver with what's done in another Intel's driver: the avs-driver.
In essence, no need to log the error in every function, let the common
handler do so instead.
The last three simplify the code, and fix some spacing issues. All in
all, we get better readability with lower LOC.
The DSP suspend path currently waits unconditionally for a suspend ack
from the firmware. This breaks firmwares that do not implement the
mailbox-based CONFIRMATION handshake, as the DSP never responds and
system suspend fails with -EBUSY.
The driver already uses the WAIT_FW_CONFIRMATION flag to indicate that
the firmware supports the CONFIRMATION handshake at boot. Apply the same
logic during suspend: only send the suspend message and wait for the
suspend ack when the firmware is expected to support it.
Signed-off-by: Iuliana Prodan <iuliana.prodan@nxp.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Reviewed-by: Daniel Baluta <daniel.baluta@nxp.com>
Link: https://lore.kernel.org/r/20251204122825.756106-3-iuliana.prodan@oss.nxp.com
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Rename WAIT_FW_READY to WAIT_FW_CONFIRMATION and FEATURE_DONT_WAIT_FW_READY
to FEATURE_SKIP_FW_CONFIRMATION. This way, the term CONFIRMATION covers:
- waiting for firmware to confirm it is ready to start;
- waiting for any other confirmation from firmware.
Signed-off-by: Iuliana Prodan <iuliana.prodan@nxp.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Reviewed-by: Daniel Baluta <daniel.baluta@nxp.com>
Link: https://lore.kernel.org/r/20251204122825.756106-2-iuliana.prodan@oss.nxp.com
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Firmwares that do not use mailbox communication (e.g., the hello_world
sample) leave priv->tx_ch as NULL. The current suspend logic
unconditionally sends RP_MBOX_SUSPEND_SYSTEM, which is invalid without
an initialized TX channel.
Detect the no_mailboxes case early and skip sending the suspend
message. Instead, proceed directly to the runtime PM suspend path,
which is the correct behavior for firmwares that cannot respond to
mailbox requests.
Signed-off-by: Iuliana Prodan <iuliana.prodan@nxp.com>
Link: https://lore.kernel.org/r/20251204122825.756106-1-iuliana.prodan@oss.nxp.com
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
The old text binding 'gpio-dsp-keystone.txt' was replaced by a DT schema in
commit aff0a1701b
("dt-bindings: gpio: Convert ti,keystone-dsp-gpio to DT schema").
Update the reference to point to the new file.
Signed-off-by: Soham Metha <sohammetha01@gmail.com>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20251203180337.50831-1-sohammetha01@gmail.com
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
The driver_override_show function reads the driver_override string
without holding the device_lock. However, the store function modifies
and frees the string while holding the device_lock. This creates a race
condition where the string can be freed by the store function while
being read by the show function, leading to a use-after-free.
To fix this, replace the rpmsg_string_attr macro with explicit show and
store functions. The new driver_override_store uses the standard
driver_set_override helper. Since the introduction of
driver_set_override, the comments in include/linux/rpmsg.h have stated
that this helper must be used to set or clear driver_override, but the
implementation was not updated until now.
Because driver_set_override modifies and frees the string while holding
the device_lock, the new driver_override_show now correctly holds the
device_lock during the read operation to prevent the race.
Additionally, since rpmsg_string_attr has only ever been used for
driver_override, removing the macro simplifies the code.
Fixes: 39e47767ec ("rpmsg: Add driver_override device attribute for rpmsg_device")
Cc: stable@vger.kernel.org
Signed-off-by: Gui-Dong Han <hanguidong02@gmail.com>
Link: https://lore.kernel.org/r/20251202174948.12693-1-hanguidong02@gmail.com
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Previously, polling an rpmsg endpoint (e.g. /dev/ttyRPMSGx) would
generate EPOLLIN events but no EPOLLOUT events.
Unfortunately, poll support means that we can no longer disable
tx-complete interrupts as there is no way to know whether a poller is
waiting in sendq, so we always need notifications.
Signed-off-by: Patrick Oppenlander <patrick.oppenlander@gmail.com>
Link: https://lore.kernel.org/r/20251022032817.320378-1-patrick.oppenlander@gmail.com
Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Merge series from Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>:
cleanup.h coding style asks to avoid having constructors with redundant
values (= NULL). On purpose, because it is just not making the code
simpler. The constructor should be meaningful not just NULL.
If you do not agree in declaration-in-place-of-use (fair!), then do not
use cleanup.h. If you want to use cleanup.h, then please read cleanup.h
before.
Merge series from Biju <biju.das.au@gmail.com>:
Add support for 24 and 32-bit sample format width for RZ/G2L SoCs. Apart
from this, the patch series includes some code cleanups.
Merge series from Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>:
Please, please stop ending cleanup.h patches for very simple code like:
foo = kzalloc();
kfree(foo);
return;
... *if you do not intend to read cleanup.h*. These changes are making
simple code not necessarily simpler. But worse, if you do not read
cleanup.h then you introduce actually undesired, error-prone and wrong
style of having constructors with redundant values (= NULL).
This is actually worse code.
If you do not agree in declaration-in-place-of-use (fair!), then do not
use cleanup.h. If you want to use cleanup.h, then please read cleanup.h
before.
This is second mixup I see recently around Qualcomm files.
