[alsa-devel] [PATCH] [PATCH v2] ALSA: hda: Retry codec read on first time failure
From: Vani Ganji vani.ganji@intel.com
Sometimes HDMI audio fails while fast switching between HDMI and Speaker due to codec Get Power state verb read failure.
Verb is codec driver to access codec HW. Since issue scenario involves fast switching between HDMI and speaker, there might be chance to fail read from HW and same happened in the error scenario.
Based on Display/Graphics architect suggestion, added a retry mechanism for failure on first attempt.
Changes from v1: -Corrected missing braces.
Signed-off-by: Vani Ganji vani.ganji@intel.com --- sound/hda/hdac_device.c | 15 +++++++++++---- 1 file changed, 11 insertions(+), 4 deletions(-)
diff --git a/sound/hda/hdac_device.c b/sound/hda/hdac_device.c index b26cc93e7e10..3874d54aa686 100644 --- a/sound/hda/hdac_device.c +++ b/sound/hda/hdac_device.c @@ -984,10 +984,17 @@ static unsigned int codec_read(struct hdac_device *hdac, hda_nid_t nid, { unsigned int cmd = snd_hdac_make_cmd(hdac, nid, verb, parm); unsigned int res; - - if (snd_hdac_exec_verb(hdac, cmd, flags, &res)) - return -1; - + unsigned int count; + + /* Retry verb once more if it fails first time */ + /* recommended to retry since it is HW related */ + for (count = 0; count <= 1; count++) { + if (snd_hdac_exec_verb(hdac, cmd, flags, &res)) { + if (count == 1) + return -1; + } else + break; + } return res; }
On Fri, 30 Aug 2019 07:23:18 +0200, vani.ganji@intel.com wrote:
From: Vani Ganji vani.ganji@intel.com
Sometimes HDMI audio fails while fast switching between HDMI and Speaker due to codec Get Power state verb read failure.
Verb is codec driver to access codec HW. Since issue scenario involves fast switching between HDMI and speaker, there might be chance to fail read from HW and same happened in the error scenario.
Based on Display/Graphics architect suggestion, added a retry mechanism for failure on first attempt.
Changes from v1: -Corrected missing braces.
Signed-off-by: Vani Ganji vani.ganji@intel.com
Do you still need this even with the very latest upstream code? As mentioned, the recent change to use the sync-write should reduce such errors.
If it's about the SOF/SST, you can try to set the hdac_bus.sync_write flag in the corresponding drivers, too.
thanks,
Takashi
sound/hda/hdac_device.c | 15 +++++++++++---- 1 file changed, 11 insertions(+), 4 deletions(-)
diff --git a/sound/hda/hdac_device.c b/sound/hda/hdac_device.c index b26cc93e7e10..3874d54aa686 100644 --- a/sound/hda/hdac_device.c +++ b/sound/hda/hdac_device.c @@ -984,10 +984,17 @@ static unsigned int codec_read(struct hdac_device *hdac, hda_nid_t nid, { unsigned int cmd = snd_hdac_make_cmd(hdac, nid, verb, parm); unsigned int res;
- if (snd_hdac_exec_verb(hdac, cmd, flags, &res))
return -1;
- unsigned int count;
- /* Retry verb once more if it fails first time */
- /* recommended to retry since it is HW related */
- for (count = 0; count <= 1; count++) {
if (snd_hdac_exec_verb(hdac, cmd, flags, &res)) {if (count == 1)return -1;} elsebreak;- } return res;
}
-- 2.17.1
Alsa-devel mailing list Alsa-devel@alsa-project.org https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
participants (2)
-
Takashi Iwai -
vani.ganji＠intel.com