Suspending/resuming with an HDMI dongle attached leads to crashes from
an audio regmap.
Unable to handle kernel paging request at virtual address ffffffc018068000
Mem abort info:
ESR = 0x96000047
EC = 0x25: DABT (current EL), IL = 32 bits
SET = 0, FnV = 0
EA = 0, S1PTW = 0
Data abort info:
ISV = 0, ISS = 0x00000047
CM = 0, WnR = 1
swapper pgtable: 4k pages, 39-bit VAs, pgdp=0000000081b12000
[ffffffc018068000] pgd=0000000275d14003, pud=0000000275d14003, pmd=000000026365d003, pte=0000000000000000
Internal error: Oops: 96000047 [#1] PREEMPT SMP
Call trace:
regmap_mmio_write32le+0x2c/0x40
regmap_mmio_write+0x48/0x6c
_regmap_bus_reg_write+0x34/0x44
_regmap_write+0x100/0x150
regcache_default_sync+0xc0/0x138
regcache_sync+0x188/0x26c
lpass_platform_pcmops_resume+0x48/0x54 [snd_soc_lpass_platform]
snd_soc_component_resume+0x28/0x40
soc_resume_deferred+0x6c/0x178
process_one_work+0x208/0x3c8
worker_thread+0x23c/0x3e8
kthread+0x144/0x178
ret_from_fork+0x10/0x18
Code: d503201f d50332bf f94002a8 8b344108 (b9000113)
I can reliably reproduce this problem by running 'tail' on the registers
file in debugfs for the hdmi regmap.
# tail /sys/kernel/debug/regmap/62d87000.lpass-lpass_hdmi/registers
[ 84.658733] Unable to handle kernel paging request at virtual address ffffffd0128e800c
This crash happens because we're trying to read registers from the
regmap beyond the length of the mapping created by ioremap().
The number of hdmi_rdma_channels determines the size of the regmap via
this code in sound/soc/qcom/lpass-cpu.c:
lpass_hdmi_regmap_config.max_register = LPAIF_HDMI_RDMAPER_REG(variant, variant->hdmi_rdma_channels);
According to debugfs the size of the regmap is 0x68010 but according to
the DTS file posted in [1] the size is only 0x68000 (see the first reg
property of the lpass_cpu node). Let's change the number of channels to
be 3 instead of 4 so the math works out to have a max register of
0x67010, nicely fitting inside of the region size of 0x68000.
Note: I tried to bump up the size of the register region to the next
page to include the 0x68010 register but then the tail command caused
SErrors with an async abort, implying that the register region doesn't
exist or it isn't clocked because the bus is telling us that the
register read failed. I reduce the number of channels and played audio
through the HDMI channel and it kept working so I think this is correct.
Fixes: 2ad63dc8df ("ASoC: qcom: sc7180: Add support for audio over DP")
Link: https://lore.kernel.org/r/1601448168-18396-2-git-send-email-srivasam@codeaurora.org [1]
Cc: V Sujith Kumar Reddy <vsujithk@codeaurora.org>
Cc: Srinivasa Rao <srivasam@codeaurora.org>
Cc: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Cc: Cheng-Yi Chiang <cychiang@chromium.org>
Signed-off-by: Stephen Boyd <swboyd@chromium.org>
Link: https://lore.kernel.org/r/20210115203329.846824-1-swboyd@chromium.org
Signed-off-by: Mark Brown <broonie@kernel.org>
LPASS driver is partially broken on DragonBoard DB410c on 5.10 and
its totally broken on other Supported Qualcomm SoCs.
This was due to DAI ids being over written by the SoC specific header files
in the dt-bindings.
Idea of having SoC specific headers is not doable when we are dealing with
a common driver. So this patchset attempts to fix this properly by creating
a common dt-bindings header for lpass which can be updated with new entries
if required. This patchset also add an simple of_xlate function to resolve
the dai names and different SoCs might not have 1:1 mapping for the
dai_driver array with dai ids.
Changes since v1:
- removed array indexes as suggested by Stephan G.
- rebased to sound/for-next branch
- collected Srinivasa tested-by tag for sc7180 platform.
Thanks,
srini
Srinivas Kandagatla (2):
ASoC: dt-bindings: lpass: Fix and common up lpass dai ids
ASoC: qcom: Fix broken support to MI2S TERTIARY and QUATERNARY
include/dt-bindings/sound/apq8016-lpass.h | 7 +++----
include/dt-bindings/sound/qcom,lpass.h | 15 +++++++++++++++
include/dt-bindings/sound/sc7180-lpass.h | 6 ++----
sound/soc/qcom/lpass-cpu.c | 22 ++++++++++++++++++++++
sound/soc/qcom/lpass-platform.c | 12 ++++++++++++
sound/soc/qcom/lpass-sc7180.c | 9 +++------
sound/soc/qcom/lpass.h | 2 +-
7 files changed, 58 insertions(+), 15 deletions(-)
create mode 100644 include/dt-bindings/sound/qcom,lpass.h
--
2.21.0
lpass hdmi support patch totally removed support for MI2S TERTIARY
and QUATERNARY.
One of the major issue was spotted with the design of having
separate SoC specific header files for the common lpass driver.
This design is prone to break as an when new SoC header is added
as the common DAI ids of other SoCs will be overwritten by the
new ones.
Having a common header qcom,lpass.h should fix the issue and any new
DAI ids should be added to the common header.
With this change lpass also needs a new of_xlate function to resolve
dai name.
Fixes: 7cb37b7bd0 ("ASoC: qcom: Add support for lpass hdmi driver")
Reported-by: Jun Nie <jun.nie@linaro.org>
Reported-by: Stephan Gerhold <stephan@gerhold.net>
Tested-by: Srinivasa Rao <srivasam@codeaurora.org>
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Tested-by: Stephan Gerhold <stephan@gerhold.net>
Link: https://lore.kernel.org/r/20210119171527.32145-3-srinivas.kandagatla@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
MI2S and DMA control registers are not volatile, so remove these from volatile registers list.
Registers reset state check by reading non volatile registers makes no use,
so remove error check from cpu and platform trigger callbacks.
Initialized map variable two times in lpass platform trigger API,
so remove redundant initialization.
Fixes commit b182496822 ("ASoC: qcom: Fix enabling BCLK and LRCLK in LPAIF invalid state")
Signed-off-by: V Sujith Kumar Reddy <vsujithk@codeaurora.org>
Signed-off-by: Srinivasa Rao Mandadapu <srivasam@codeaurora.org>
Link: https://lore.kernel.org/r/1608192514-29695-2-git-send-email-srivasam@codeaurora.org
Signed-off-by: Mark Brown <broonie@kernel.org>
setting clock rate on child clocks without a parent clock rate will
result in zero clk rate for child. This also means that when audio
is started dsp will attempt to access registers without enabling
clock resulting in board boot up.
Fix this by adding the missing parent clock rate.
Fixes: 520a1c396d ("ASoC: q6afe-clocks: add q6afe clock controller")
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20201204164228.1826-1-srinivas.kandagatla@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Fix enabling BCLK and LRCLK only when LPAIF is invalid state and
bit clock in enable state.
In device suspend/resume scenario LPAIF is going to reset state.
which is causing LRCLK disable and BCLK enable.
Avoid such inconsitency by removing unnecessary cpu dai prepare API,
which is doing LRCLK enable, and by maintaining BLCK state information.
Fixes: 7e6799d8f8 ("ASoC: qcom: lpass-cpu: Enable MI2S BCLK and LRCLK together")
Signed-off-by: V Sujith Kumar Reddy <vsujithk@codeaurora.org>
Signed-off-by: Srinivasa Rao Mandadapu <srivasam@codeaurora.org>
Link: https://lore.kernel.org/r/1606148273-17325-1-git-send-email-srivasam@codeaurora.org
Signed-off-by: Mark Brown <broonie@kernel.org>
There are two issues in this function.
1) We can't drop the refrences on "cpu", "codec" and "platform" before
we take the reference. This doesn't cause a problem on the first
iteration because those pointers start as NULL so the of_node_put()
is a no-op. But on the subsequent iterations, it will lead to a use
after free.
2) If the devm_kzalloc() allocation failed then the code returned
directly instead of cleaning up.
Fixes: c1e6414cdc ("ASoC: qcom: common: Fix refcount imbalance on error")
Fixes: 1e36ea360a ("ASoC: qcom: common: use modern dai_link style")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/r/20201105125154.GA176426@mwanda
Signed-off-by: Mark Brown <broonie@kernel.org>
Disable MI2S bit clock from PAUSE/STOP/SUSPEND usecase instead of
shutdown time. Acheive this by invoking clk_disable API from
cpu daiops trigger instead of cpu daiops shutdown.
Change non-atomic API "clk_prepare_enable" to atomic API
"clk_enable" in trigger, as trigger is being called from atomic context.
Fixes: 7e6799d8f8 ("ASoC: qcom: lpass-cpu: Enable MI2S BCLK and LRCLK together")
Signed-off-by: V Sujith Kumar Reddy <vsujithk@codeaurora.org>
Signed-off-by: Srinivasa Rao Mandadapu <srivasam@codeaurora.org>
Link: https://lore.kernel.org/r/1603098363-9251-1-git-send-email-srivasam@codeaurora.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Note:
- The machine driver patch is made by the collaboration of
Cheng-Yi Chiang <cychiang@chromium.org>
Rohit kumar <rohitkr@codeaurora.org>
Ajit Pandey <ajitp@codeaurora.org>
But Ajit has left codeaurora.
Changes from v1 to v2:
- Ducumentation: Addressed all suggestions from Doug.
- Machine driver:
- Fix comment style for license.
- Sort includes.
- Remove sc7180_snd_hw_params.
- Remove sc7180_dai_init and use aux device instead for headset jack registration.
- Statically define format for Primary MI2S.
- Atomic is not a concern because there is mutex in card to make sure
startup and shutdown happen sequentially.
- Fix missing return -EINVAL in startup.
- Use static sound card.
- Use devm_kzalloc to avoid kfree.
Changes from v2 to v3:
- Ducumentation: Addressed suggestions from Srini.
- Machine driver:
- Reuse qcom_snd_parse_of to parse properties.
- Remove playback-only and capture-only.
- Misc fixes to address comments.
Changes from v3 to v4:
- Ducumentation: Addressed suggestions from Rob.
- Remove definition of dai.
- Use 'sound-dai: true' for sound-dai schema.
- Add reg property to pass 'make dt_binding_check' check although reg is not used in the driver.
- Machine driver:
- Add Reviewed-by: Tzung-Bi Shih <tzungbi@google.com>
Changes from v4 to v5:
- Documentation: Addressed suggestions from Rob.
- Add definition for "#address-cells" and "#size-cells".
- Add additionalProperties: false
- Add required properties.
Changes from v5 to v6:
- Documentation: Addressed suggestions from Rob.
- Drop contains in compatible strings.
- Only allow dai-link@[0-9]
- Remove reg ref since it has a type definition already.
Changes from v6 to v7
- Documentation:
- Add headset-jack and hdmi-jack to specify the codec
responsible for jack detection.
- HDMI codec driver:
- Use component set_jack ops instead of exporting hdmi_codec_set_jack_detect.
- Machine driver:
- Removed aux device following Stephan's suggestion.
- Use headset-jack and hdmi-jack to specify the codec
responsible for jack detection.
- Add support for HDMI(actually DP) playback.
Changes from v7 to v8
- Documentation:
- Remove headset-jack and hdmi-jack.
- Machine driver:
- Let machine driver decide whether there is a jack on the DAI.
Changes from v8 to v9
- hdmi-codec driver:
- Fixed the naming.
- Machine driver:
- Fixed unused fields.
- Moved snd_soc_card_set_drvdata
- Keep the naming of HDMI as dai name until v5 of lpass-hdmi patches.
Changes from v9 to v10
- Documentation:
- Let compatible string be more specific for board configuration to allow
for future changes.
- Machine driver:
- Fixed unused include and macro.
- Add temporary macro SC7180_LPASS_DP for future change in sc7180-lpass.h.
- Let sound card be dynamically allocated.
- Change compatible string accordingly.
Changes from v10 to v11
- Machine driver:
- Use temporary macro LPASS_DP_RX for future change in sc7180-lpass.h.
Changes from v11 to v12
- Documentation:
- Change the file and title name for new compatible string google,sc7180-trogdor.
- Change the example of model name.
- Machine driver:
- Use the definitaion of index LPASS_DP_RX in sc7180-lpass.h.
- Fix for compatible string.
- Replace a comma with semicolon.
Ajit Pandey (1):
ASoC: qcom: sc7180: Add machine driver for sound card registration
Cheng-Yi Chiang (1):
ASoC: google: dt-bindings: Add sc7180-trogdor machine bindings
.../bindings/sound/google,sc7180-trogdor.yaml | 130 +++++++++
sound/soc/qcom/Kconfig | 12 +
sound/soc/qcom/Makefile | 2 +
sound/soc/qcom/sc7180.c | 266 ++++++++++++++++++
4 files changed, 410 insertions(+)
create mode 100644 Documentation/devicetree/bindings/sound/google,sc7180-trogdor.yaml
create mode 100644 sound/soc/qcom/sc7180.c
--
2.29.0.rc2.309.g374f81d7ae-goog
With the current state of code, we would endup with something like
below in /proc/asound/cards for 2 machines based on this driver.
Machine 1:
0 [DB845c ]: DB845c - DB845c
DB845c
Machine 2:
0 [LenovoYOGAC6301]: Lenovo-YOGA-C63 - Lenovo-YOGA-C630-13Q50
LENOVO-81JL-LenovoYOGAC630_13Q50-LNVNB161216
This is not very UCM friendly both w.r.t to common up configs and
card identification, and UCM2 became totally not usefull with just
one ucm sdm845.conf for two machines which have different setups
w.r.t HDMI and other dais.
Reasons for such thing is partly because Qualcomm machine drivers never
cared to set driver_name.
This patch sets up driver name for the this driver to sort out the
UCM integration issues!
after this patch contents of /proc/asound/cards:
Machine 1:
0 [DB845c ]: sdm845 - DB845c
DB845c
Machine 2:
0 [LenovoYOGAC6301]: sdm845 - Lenovo-YOGA-C630-13Q50
LENOVO-81JL-LenovoYOGAC630_13Q50-LNVNB161216
with this its possible to align with what UCM2 expects and we can have
sdm845/DB845.conf
sdm845/LENOVO-81JL-LenovoYOGAC630_13Q50-LNVNB161216.conf
... for board variants. This should scale much better!
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20201023095849.22894-1-srinivas.kandagatla@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Here are fixes for two warnings types discovered while building qdsp6 drivers
without CONFIG_OF and with W=1
One of them was reported by Intel kernel test robot on q6afe-clocks patch, which
equally applies to rest of the qdsp6 drivers.
changes since v1:
- added ifdef CONFIG_OF instead of removing of_match_ptr
Srinivas Kandagatla (2):
ASoC: qdsp6: add ifdef CONFIG_OF around of_device_id
ASoC: q6asm: fix kernel doc warnings
sound/soc/qcom/qdsp6/q6adm.c | 2 ++
sound/soc/qcom/qdsp6/q6afe-clocks.c | 2 ++
sound/soc/qcom/qdsp6/q6afe-dai.c | 2 ++
sound/soc/qcom/qdsp6/q6afe.c | 2 ++
sound/soc/qcom/qdsp6/q6asm-dai.c | 2 ++
sound/soc/qcom/qdsp6/q6asm.c | 6 ++++++
sound/soc/qcom/qdsp6/q6core.c | 2 ++
sound/soc/qcom/qdsp6/q6routing.c | 2 ++
8 files changed, 20 insertions(+)
--
2.21.0
This patch fixes below kernel doc warnings on not describing all the parmeters
sound/soc/qcom/qdsp6/q6asm.c:927: warning: Function parameter or member
'stream_id' not described in 'q6asm_open_write'
sound/soc/qcom/qdsp6/q6asm.c:927: warning: Function parameter or member
'is_gapless' not described in 'q6asm_open_write'
sound/soc/qcom/qdsp6/q6asm.c:1053: warning: Function parameter or member
'stream_id' not described in 'q6asm_run'
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Link: https://lore.kernel.org/r/20200925163552.20717-3-srinivas.kandagatla@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
