Add a new struct to save the current position in the rsb masters_list
while sending the rsb names to other nodes. The rsb names are sent in
multiple chunks, and for each new chunk, the new "dlm_dir_dump" struct
saves the last position in the masters_list. The new struct is also
used to save more information to sanity check the recovery process.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
Move the rsb root_list from the lockspace to a stack variable since
it is now only used by the ls_recover() function.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
Add a new "masters_list" for master rsb structs, with a new
rwlock. The new list is created and used during the recovery
process to send the master rsb names to new nodes. With this
change, the current "root_list" can be used without locking.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
Move dlm_create_root_list() and dlm_release_root_list() to
recover.c and declare them static because they are only used
there.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
Replace GFP_NOFS with GFP_ATOMIC. Also stop using idr_preload which
uses a non-bh spin_lock. This is further preparation for softirq
message processing.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
When zapping TDP MMU SPTEs under read-lock, processes zapped SPTEs *after*
flushing TLBs and after replacing the special REMOVED_SPTE with '0'.
When zapping an SPTE that points to a page table, processing SPTEs after
flushing TLBs minimizes contention on the child SPTEs (e.g. vCPUs won't
hit write-protection faults via stale, read-only child SPTEs), and
processing after replacing REMOVED_SPTE with '0' minimizes the amount of
time vCPUs will be blocked by the REMOVED_SPTE.
Processing SPTEs after setting the SPTE to '0', i.e. in parallel with the
SPTE potentially being replacing with a new SPTE, is safe because KVM does
not depend on completing the processing before a new SPTE is installed, and
the processing is done on a subset of the page tables that is disconnected
from the root, and thus unreachable by other tasks (after the TLB flush).
KVM already relies on similar logic, as kvm_mmu_zap_all_fast() can result
in KVM processing all SPTEs in a given root after vCPUs create mappings in
a new root.
In VMs with a large (400+) number of vCPUs, it can take KVM multiple
seconds to process a 1GiB region mapped with 4KiB entries, e.g. when
disabling dirty logging in a VM backed by 1GiB HugeTLB. During those
seconds, if a vCPU accesses the 1GiB region being zapped it will be
stalled until KVM finishes processing the SPTE and replaces the
REMOVED_SPTE with 0.
Re-ordering the processing does speed up the atomic-zaps somewhat, but
the main benefit is avoiding blocking vCPU threads.
Before:
$ ./dirty_log_perf_test -s anonymous_hugetlb_1gb -v 416 -b 1G -e
...
Disabling dirty logging time: 509.765146313s
$ ./funclatency -m tdp_mmu_zap_spte_atomic
msec : count distribution
0 -> 1 : 0 | |
2 -> 3 : 0 | |
4 -> 7 : 0 | |
8 -> 15 : 0 | |
16 -> 31 : 0 | |
32 -> 63 : 0 | |
64 -> 127 : 0 | |
128 -> 255 : 8 |** |
256 -> 511 : 68 |****************** |
512 -> 1023 : 129 |********************************** |
1024 -> 2047 : 151 |****************************************|
2048 -> 4095 : 60 |*************** |
After:
$ ./dirty_log_perf_test -s anonymous_hugetlb_1gb -v 416 -b 1G -e
...
Disabling dirty logging time: 336.516838548s
$ ./funclatency -m tdp_mmu_zap_spte_atomic
msec : count distribution
0 -> 1 : 0 | |
2 -> 3 : 0 | |
4 -> 7 : 0 | |
8 -> 15 : 0 | |
16 -> 31 : 0 | |
32 -> 63 : 0 | |
64 -> 127 : 0 | |
128 -> 255 : 12 |** |
256 -> 511 : 166 |****************************************|
512 -> 1023 : 101 |************************ |
1024 -> 2047 : 137 |********************************* |
Note, KVM's processing of collapsible SPTEs is still extremely slow and
can be improved. For example, a significant amount of time is spent
calling kvm_set_pfn_{accessed,dirty}() for every last-level SPTE, even
when processing SPTEs that all map the same folio. But avoiding blocking
vCPUs and contending SPTEs is valuable regardless of how fast KVM can
process collapsible SPTEs.
Link: https://lore.kernel.org/all/20240320005024.3216282-1-seanjc@google.com
Cc: Vipin Sharma <vipinsh@google.com>
Suggested-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: David Matlack <dmatlack@google.com>
Reviewed-by: Vipin Sharma <vipinsh@google.com>
Link: https://lore.kernel.org/r/20240307194059.1357377-1-dmatlack@google.com
[sean: massage changelog]
Signed-off-by: Sean Christopherson <seanjc@google.com>
Commit 228804a35c tried to add a refcount check to
gfs2_scan_glock_lru() to make sure that glocks that are still referenced
cannot be freed. It failed to account for the bias state_change() adds
to the refcount for held glocks, so held glocks are no longer removed
from the glock cache, which can lead to out-of-memory problems. Fix
that. (The inodes those glocks are associated with do get shrunk and do
get pushed out of memory.)
In addition, use the same eligibility check in gfs2_scan_glock_lru() and
gfs2_dispose_glock_lru().
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
This reverts commit 62862485a4.
Commit 62862485a4 tried to fix issues introduced by commit
228804a35c ("gfs2: Make glock lru list scanning safer"), but like that
commit, it failed to account for the bias state_change() adds to the
glock reference count for locked glocks. Revert commit 62862485a4 so
that we can fix commit 228804a35c properly.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Commit 3e11e53041 tries to suppress dlm_lock() lock conversion errors
that occur when the lockspace has already been released.
It does that by setting and checking the SDF_SKIP_DLM_UNLOCK flag. This
conflicts with the intended meaning of the SDF_SKIP_DLM_UNLOCK flag, so
check whether the lockspace is still allocated instead.
(Given the current DLM API, checking for this kind of error after the
fact seems easier that than to make sure that the lockspace is still
allocated before calling dlm_lock(). Changing the DLM API so that users
maintain the lockspace references themselves would be an option.)
Fixes: 3e11e53041 ("GFS2: ignore unlock failures after withdraw")
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
The GLF_LOCK flag is protected by the gl->gl_lockref.lock spin lock
which is held when entering run_queue(), so we can use test_bit() and
set_bit() here.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
In gfs2_dispose_glock_lru(), we want to skip glocks which are in the
process of transitioning state (as indicated by the set GLF_LOCK flag),
but we we don't need to set that flag for requesting a state transition.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Function gfs2_glock_queue_put() puts a glock reference by enqueuing
glock work instead of putting the reference directly. This ensures that
the operation won't sleep, but it is costly and really only necessary
when putting the final glock reference. Replace it with a new
gfs2_glock_put_async() function that only queues glock work when putting
the last glock reference.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
In function signal_our_withdraw(), we are calling gfs2_glock_queue_put()
in a context in which we are actually allowed to sleep, so replace that
with a simple call to gfs2_glock_put().
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
In gfs2_jindex_free(), set sdp->sd_jdesc to NULL under the log flush
lock to provide exclusion against gfs2_log_flush().
In gfs2_log_flush(), check if sdp->sd_jdesc is non-NULL before
dereferencing it. Otherwise, we could run into a NULL pointer
dereference when outstanding glock work races with an unmount
(glock_work_func -> run_queue -> do_xmote -> inode_go_sync ->
gfs2_log_flush).
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Commit fffe9bee14 ("gfs2: Delay withdraw from atomic context")
switched from gfs2_withdraw() to gfs2_withdraw_delayed() in
gfs2_ail_error(), but failed to then check if a delayed withdraw had
occurred. Fix that by adding the missing check in __gfs2_ail_flush(),
where the spin locks are already dropped and a withdraw is possible.
Fixes: fffe9bee14 ("gfs2: Delay withdraw from atomic context")
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
When calling gfs2_glock_get(), consistently pass the CREATE or NO_CREATE
flag for the create argument. This is the only place that passes 0
instead.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Follow up to commit e7beb8b6de ("gfs2: Rename SDF_DEACTIVATING to
SDF_KILL") and change the description of the SDF_KILL flag from
"Deactivating" to "Killing".
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
The type argument of gfs2_meta_check_ii() is always set to "magic
number", so remove that argument and hardcode the string in
gfs2_meta_check_ii(). Change the string to "bad magic number" to
emphasize that the problem is the incorrect magic number.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Get rid of attempts to create multi-line syslog entries; this only makes
the messages harder to read.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
gfs2_consist_inode() logs an error message with the source file and line
number. When we jump before calling it, the line number becomes less
useful as it no longer relates to the source of the error. To aid
troubleshooting, replace the gotos with the gfs2_consist_inode() calls
so that the error messages are more informative.
Signed-off-by: Andrew Price <anprice@redhat.com>
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
nouveau:
- regression fix for GSP display enable.
-----BEGIN PGP SIGNATURE-----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=pPLP
-----END PGP SIGNATURE-----
Merge tag 'drm-fixes-2024-04-09' of https://gitlab.freedesktop.org/drm/kernel
Pull drm nouveau fix from Dave Airlie:
"A previous fix to nouveau devinit on the GSP paths fixed the Turing
but broke Ampere, I did some more digging and found the proper fix.
Sending it early as I want to make sure it makes the next 6.8 stable
kernels to fix the regression.
Regular fixes will be at end of week as usual.
nouveau:
- regression fix for GSP display enable"
* tag 'drm-fixes-2024-04-09' of https://gitlab.freedesktop.org/drm/kernel:
nouveau: fix devinit paths to only handle display on GSP.
In current configuration, wm8904 codec on Dahlia carrier board provides
distorted audio output. This happens due to reference clock is fixed to
25MHz and no FLL is enabled. During playback following parameters are set:
44100Hz:
[ 310.276924] wm8904 1-001a: Target BCLK is 1411200Hz
[ 310.276990] wm8904 1-001a: Using 25000000Hz MCLK
[ 310.277001] wm8904 1-001a: CLK_SYS is 12500000Hz
[ 310.277018] wm8904 1-001a: Selected CLK_SYS_RATIO of 256
[ 310.277026] wm8904 1-001a: Selected SAMPLE_RATE of 44100Hz
[ 310.277034] wm8904 1-001a: Selected BCLK_DIV of 80 for 1562500Hz BCLK
[ 310.277044] wm8904 1-001a: LRCLK_RATE is 35
Deviation = 1411200 vs 1562500 = 10.721%
Also, LRCLK_RATE is 35, should be 32.
48000Hz:
[ 302.449970] wm8904 1-001a: Target BCLK is 1536000Hz
[ 302.450037] wm8904 1-001a: Using 25000000Hz MCLK
[ 302.450049] wm8904 1-001a: CLK_SYS is 12500000Hz
[ 302.450065] wm8904 1-001a: Selected CLK_SYS_RATIO of 256
[ 302.450074] wm8904 1-001a: Selected SAMPLE_RATE of 48000Hz
[ 302.450083] wm8904 1-001a: Selected BCLK_DIV of 80 for 1562500Hz BCLK
[ 302.450092] wm8904 1-001a: LRCLK_RATE is 32
Deviation = 1536000 vs 1562500 = 1.725%
Enabling wm8904 FLL via providing mclk-fs property to simple-audio-card
configures clocks properly, but also adjusts audio reference clock
(mclk), which in case of TI AM62 should be avoided, as it only
supports 25MHz output [1][2].
This change enables FLL on wm8904 by providing mclk-fs, and drops
audio reference clock out of DAI configuration, which prevents
simple-audio-card to adjust it before every playback [3].
41000Hz:
[ 111.820533] wm8904 1-001a: FLL configured for 25000000Hz->11289600Hz
[ 111.820597] wm8904 1-001a: Clock source is 0 at 11289600Hz
[ 111.820651] wm8904 1-001a: Using 11289600Hz FLL clock
[ 111.820703] wm8904 1-001a: CLK_SYS is 11289600Hz
[ 111.820798] wm8904 1-001a: Target BCLK is 1411200Hz
[ 111.820847] wm8904 1-001a: Using 11289600Hz FLL clock
[ 111.820894] wm8904 1-001a: CLK_SYS is 11289600Hz
[ 111.820933] wm8904 1-001a: Selected CLK_SYS_RATIO of 256
[ 111.820971] wm8904 1-001a: Selected SAMPLE_RATE of 44100Hz
[ 111.821009] wm8904 1-001a: Selected BCLK_DIV of 80 for 1411200Hz BCLK
[ 111.821051] wm8904 1-001a: LRCLK_RATE is 32
48000Hz:
[ 144.119254] wm8904 1-001a: FLL configured for 25000000Hz->12288000Hz
[ 144.119309] wm8904 1-001a: Clock source is 0 at 12288000Hz
[ 144.119364] wm8904 1-001a: Using 12288000Hz FLL clock
[ 144.119413] wm8904 1-001a: CLK_SYS is 12288000Hz
[ 144.119512] wm8904 1-001a: Target BCLK is 1536000Hz
[ 144.119561] wm8904 1-001a: Using 12288000Hz FLL clock
[ 144.119608] wm8904 1-001a: CLK_SYS is 12288000Hz
[ 144.119646] wm8904 1-001a: Selected CLK_SYS_RATIO of 256
[ 144.119685] wm8904 1-001a: Selected SAMPLE_RATE of 48000Hz
[ 144.119723] wm8904 1-001a: Selected BCLK_DIV of 80 for 1536000Hz BCLK
[ 144.119764] wm8904 1-001a: LRCLK_RATE is 32
[1]: https://e2e.ti.com/support/processors-group/processors/f/processors-forum/1175479/processor-sdk-am62x-output-audio_ext_refclk0-as-mclk-for-codec-and-mcbsp/4444986#4444986
[2]: https://e2e.ti.com/support/processors-group/processors/f/processors-forum/1188051/am625-audio_ext_refclk1-clock-output---dts-support/4476322#4476322
[3]: sound/soc/generic/simple-card-utils.c#L441
Fixes: f5bf894c86 ("arm64: dts: ti: verdin-am62: dahlia: add sound card")
Suggested-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: Andrejs Cainikovs <andrejs.cainikovs@toradex.com>
Reviewed-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Reviewed-by: Francesco Dolcini <francesco.dolcini@toradex.com>
Link: https://lore.kernel.org/r/20240315102500.18492-1-andrejs.cainikovs@gmail.com
Signed-off-by: Nishanth Menon <nm@ti.com>
If the max open zones of zoned devices are less than
the active logs of F2FS, the device may error due to
insufficient zone resources when multiple active logs
are being written at the same time.
Signed-off-by: Wenjie Qi <qwjhust@gmail.com>
Signed-off-by: Chao Yu <chao@kernel.org>
Reviewed-by: Daeho Jeong <daehojeong@google.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
There's no need to carve out bits of the NOP optimization functionality
and look at JMP opcodes - simply do one more NOPs optimization pass
at the end of patching.
A lot simpler code.
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/r/20240130105941.19707-3-bp@alien8.de
mbox_send_message() sends a u32 bit message, not a pointer to a message.
We only convert to a pointer type as a generic type. If we want to send
a dummy message of 0, then simply send 0 (NULL).
Signed-off-by: Andrew Davis <afd@ti.com>
Link: https://lore.kernel.org/r/20240325165507.30323-1-afd@ti.com
Signed-off-by: Nishanth Menon <nm@ti.com>
Instead of optimizing NOPs in-place, use a temporary buffer like the
usual alternatives patching flow does. This obviates the need to grab
locks when patching, see
6778977590 ("x86/alternatives: Disable interrupts and sync when optimizing NOPs in place")
While at it, add nomenclature definitions clarifying and simplifying the
naming of function-local variables in the alternatives code.
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/r/20240130105941.19707-2-bp@alien8.de
This property was only ever used in one device. It is no longer needed as
what it signaled is now default. Remove this unneeded/unused property.
Signed-off-by: Andrew Davis <afd@ti.com>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20240326223730.54639-5-afd@ti.com
Signed-off-by: Nishanth Menon <nm@ti.com>
The property ti,system-reboot-controller is no longer needed as the reboot
handler is now always registered. Remove this property.
While here remove the comment about delete-property, all K2G platforms use
PMMC, and it wasn't good advice anyway.
Signed-off-by: Andrew Davis <afd@ti.com>
Link: https://lore.kernel.org/r/20240326223730.54639-4-afd@ti.com
Signed-off-by: Nishanth Menon <nm@ti.com>
There was once a limitation that there could only be one system
reset handler. Due to that we only would register this handler
when a non-standard device tree property was found, else we left
the default handler in place (usually PSCI). Now that we can
have multiple handlers, and TI-SCI reset is always available
in the firmware, register this handler unconditionally.
This priority is left at the default so higher priority handlers
(like PSCI) are still attempted first.
Signed-off-by: Andrew Davis <afd@ti.com>
Reviewed-by: Markus Schneider-Pargmann <msp@baylibre.com>
Link: https://lore.kernel.org/r/20240326223730.54639-3-afd@ti.com
Signed-off-by: Nishanth Menon <nm@ti.com>
Use device life-cycle managed register function to simplify probe.
Signed-off-by: Andrew Davis <afd@ti.com>
Reviewed-by: Gabriel Somlo <gsomlo@gmail.com>
Reviewed-by: Markus Schneider-Pargmann <msp@baylibre.com>
Link: https://lore.kernel.org/r/20240326223730.54639-2-afd@ti.com
Signed-off-by: Nishanth Menon <nm@ti.com>
After alternatives have been patched, changes to the X86_FEATURE flags
won't take effect and could potentially even be wrong.
Warn about it.
This is something which has been long overdue.
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Srikanth Aithal <sraithal@amd.com>
Link: https://lore.kernel.org/r/20240327154317.29909-3-bp@alien8.de
McBSP's data delay can be configured from 0 to 2 bit clock periods. 0 is
used for DSP_B format, 1 is used for DSP_A format, 2 is unused.
A data delay of 2 bit clock periods can be used to interface to
'T1 framing' devices where data stream is preceded by a 'framing bit'. On
transmission, McBSP inserts a blank period (high-impedance period)
before the first data bit to leave an opportunity for other devices to
set this 'framing bit'. On reception, McBSP discards the 'framing bit'
that precedes the data stream.
Add support for the 'framing bit' according to the
'ti,T1-framing-[tx/rx]' device-tree properties. If a flag is present,
the data delay is set to 2 bit clock periods regardless of the selected
DAI format.
Signed-off-by: Bastien Curutchet <bastien.curutchet@bootlin.com>
Acked-by: Peter Ujfalusi <peter.ujfalusi@gmail.com>
Link: https://msgid.link/r/20240402071213.11671-14-bastien.curutchet@bootlin.com
Signed-off-by: Mark Brown <broonie@kernel.org>
McBSP's data delay can be configured from 0 to 2 bit clock periods. 0 is
used for DSP_B format, 1 for DSP_A format. A data delay of 2 bit clock
periods can be used to interface to 'T1 framing' devices where data
stream is preceded by a 'framing bit'. This 2 bit clock data delay is
not described in the bindings.
Add two flags 'ti,T1-framing-[rx/tx]' to enable a data delay of 2
bit clock periods in reception or transmission.
Signed-off-by: Bastien Curutchet <bastien.curutchet@bootlin.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Acked-by: Peter Ujfalusi <peter.ujfalusi@gmail.com>
Link: https://msgid.link/r/20240402071213.11671-13-bastien.curutchet@bootlin.com
Signed-off-by: Mark Brown <broonie@kernel.org>
S24_LE is supported by McBSP but not by the driver.
Add S24_LE to driver's supported formats. Using it enables the sign
extension in DRR (Data Receive Register). The other formats are kept
with the zero extension in DRR.
Remove data_type table as it is no longer used.
Signed-off-by: Bastien Curutchet <bastien.curutchet@bootlin.com>
Acked-by: Peter Ujfalusi <peter.ujfalusi@gmail.com>
Link: https://msgid.link/r/20240402071213.11671-12-bastien.curutchet@bootlin.com
Signed-off-by: Mark Brown <broonie@kernel.org>
McBSP has free-running mode where serial clocks continue to run during
emulation halts. This mode is always enabled by the driver.
Set free-running mode when SND_SOC_DAIFMT_CONT is selected by DAI
format, unset it when SND_SOC_DAIFMT_GATED is selected.
Signed-off-by: Bastien Curutchet <bastien.curutchet@bootlin.com>
Acked-by: Peter Ujfalusi <peter.ujfalusi@gmail.com>
Link: https://msgid.link/r/20240402071213.11671-11-bastien.curutchet@bootlin.com
Signed-off-by: Mark Brown <broonie@kernel.org>
McBSP can generate a SYNCERR when unexpected frame pulses are
detected. The driver always disables this feature and ignore the
unexpected frame pulses.
Enable the generation of SYNCERR by the McBSP. Unexpected frame
pulses are not ignored anymore.
Signed-off-by: Bastien Curutchet <bastien.curutchet@bootlin.com>
Acked-by: Peter Ujfalusi <peter.ujfalusi@gmail.com>
Link: https://msgid.link/r/20240402071213.11671-10-bastien.curutchet@bootlin.com
Signed-off-by: Mark Brown <broonie@kernel.org>
McBSP is able to drive bit clock and consume frame clock but BP_FC
format is not handled by McBSP driver.
Add BP_FC format support.
When BP_FC is selected:
- CLKX and CLKR are configured as outputs
- The sample rate generator is configured to be able to provide bit
clock.
Signed-off-by: Bastien Curutchet <bastien.curutchet@bootlin.com>
Acked-by: Peter Ujfalusi <peter.ujfalusi@gmail.com>
Link: https://msgid.link/r/20240402071213.11671-9-bastien.curutchet@bootlin.com
Signed-off-by: Mark Brown <broonie@kernel.org>
TDM is not supported by the McBSP driver. The McBSP datasheet does not
name explicitly TDM as a supported format but it is possible to configure
the McBSP to do TDM if all slots are used by McBSP.
Add TDM support. It uses single-phase frame. Slot width is used to
compute the McBSP's word length.
Implement the set_tdm_slot() hook of snd_soc_dai_ops struct. It only
supports TDM if all slots are used by McBSP.
The snd_soc_dai_driver's channels_max is updated from 2 to 128.
This was tested with BP_FC format on a platform designed off of
DAVINCI/OMAP_L138. A check is done in davinci_i2s_set_dai_fmt() to
prevent TDM to be used with BC_FC and BC_FP formats.
Signed-off-by: Bastien Curutchet <bastien.curutchet@bootlin.com>
Acked-by: Peter Ujfalusi <peter.ujfalusi@gmail.com>
Link: https://msgid.link/r/20240402071213.11671-8-bastien.curutchet@bootlin.com
Signed-off-by: Mark Brown <broonie@kernel.org>
In davinci_i2s_hw_params(), mcbsp_word_length is set twice to
asp_word_length[fmt].
Remove second unnecessary assignment.
Signed-off-by: Bastien Curutchet <bastien.curutchet@bootlin.com>
Acked-by: Peter Ujfalusi <peter.ujfalusi@gmail.com>
Link: https://msgid.link/r/20240402071213.11671-7-bastien.curutchet@bootlin.com
Signed-off-by: Mark Brown <broonie@kernel.org>
McBSP's internal sample rate generator can be programed to be driven by
its internal clock or by an external clock source located on CLKS pin.
The external clock source case is not handled by the driver.
Handle an optional clock related to this external clock source. If
present, the driver uses the clock located on CLKS pin as input for the
sample rate generator. Thus, the external clock rate is used to compute
divisors. If this optional clock is not present, the sample rate
generator is driven by the McBSP's functional clock.
Signed-off-by: Bastien Curutchet <bastien.curutchet@bootlin.com>
Acked-by: Peter Ujfalusi <peter.ujfalusi@gmail.com>
Link: https://msgid.link/r/20240402071213.11671-6-bastien.curutchet@bootlin.com
Signed-off-by: Mark Brown <broonie@kernel.org>
In probe(), the dev_err() is used for every returned error.
Replace dev_err() with dev_err_probe() where -EPROBE_DEFER can be
returned.
Signed-off-by: Bastien Curutchet <bastien.curutchet@bootlin.com>
Acked-by: Peter Ujfalusi <peter.ujfalusi@gmail.com>
Link: https://msgid.link/r/20240402071213.11671-5-bastien.curutchet@bootlin.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The clk_input_pin attribute of davinci_mcbsp_dev struct is not set since
commit 257ade78b6 ("ASoC: davinci-i2s: Convert to use edma-pcm").
Remove the attribute.
Keep the behaviour of the MCBSP_CLKR case as MCBSP_CLKR == 0.
I can't test the BC_FP format so I added back the initial comment that
was removed by commit ec63755337 ("ASoC: DaVinci: Added selection of
clk input pin for McBSP"). This was the last dependency to
linux/platform_data/davinci_asp.h so it is not included anymore.
Remove the enum mcbsp_clk_input_pin from davinci_asp.h as it is not used
anywhere else.
Signed-off-by: Bastien Curutchet <bastien.curutchet@bootlin.com>
Acked-by: Peter Ujfalusi <peter.ujfalusi@gmail.com>
Link: https://msgid.link/r/20240402071213.11671-4-bastien.curutchet@bootlin.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The McBSP uses an internal sample rate generator to provide bit clock
or frame clock. This sample rate generator can be programmed to be
driven by McBSP's internal clock source or by an external clock source
(located on CLKS pin). The external clock source is not described in
the bindings.
Add an optional clock item that allows to select an external clock as
sample rate generator's input.
Signed-off-by: Bastien Curutchet <bastien.curutchet@bootlin.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Acked-by: Peter Ujfalusi <peter.ujfalusi@gmail.com>
Link: https://msgid.link/r/20240402071213.11671-3-bastien.curutchet@bootlin.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Convert the binding for McBSP controllers for TI SoCs from txt
to YAML schema.
Add properties 'clocks', 'clock-names', 'power-domains' and
'#sound-dai-cells' which were missing from the txt file.
Add '#sound-dai-cells' and 'clocks' in the example which were missing
from the txt file.
Signed-off-by: Bastien Curutchet <bastien.curutchet@bootlin.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Acked-by: Peter Ujfalusi <peter.ujfalusi@gmail.com>
Link: https://msgid.link/r/20240402071213.11671-2-bastien.curutchet@bootlin.com
Signed-off-by: Mark Brown <broonie@kernel.org>
gcc-10 warns about a strncpy() that does not enforce zero-termination:
In file included from include/linux/string.h:369,
from drivers/staging/greybus/fw-management.c:9:
In function 'strncpy',
inlined from 'fw_mgmt_backend_fw_update_operation' at drivers/staging/greybus/fw-management.c:306:2:
include/linux/fortify-string.h:108:30: error: '__builtin_strncpy' specified bound 10 equals destination size [-Werror=stringop-truncation]
108 | #define __underlying_strncpy __builtin_strncpy
| ^
include/linux/fortify-string.h:187:9: note: in expansion of macro '__underlying_strncpy'
187 | return __underlying_strncpy(p, q, size);
| ^~~~~~~~~~~~~~~~~~~~
For some reason, I cannot reproduce this with gcc-9 or gcc-11, and I only
get a warning for one of the four related strncpy()s, so I'm not
sure what's going on.
Change all four to strscpy_pad(), which is the safest replacement here,
as it avoids ending up with uninitialized stack data in the tag name.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Dan Carpenter <dan.carpenter@linaro.org>
Reviewed-by: Justin Stitt <justinstitt@google.com>
Link: https://github.com/KSPP/linux/issues/90 [1]
Link: https://lore.kernel.org/r/20240408194821.3183462-3-arnd@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
