At Mon, 17 Dec 2012 16:33:10 +0100, David Henningsson wrote:
On 12/12/2012 06:15 PM, Takashi Iwai wrote:
At Wed, 12 Dec 2012 18:02:04 +0100, David Henningsson wrote:
The bug reporter reports that setting the speaker pin to D3 before turning off its pinctl fixes the clicking noise on powersave for Thinkpad T61.
Thanks to c4pp4 for doing most of the work with this bug, including reporting and testing, and writing preliminary patches.
BugLink: https://bugs.launchpad.net/bugs/886975 Signed-off-by: David Henningsson david.henningsson@canonical.com
sound/pci/hda/patch_conexant.c | 36 ++++++++++++++++++++++++++++++++---- 1 file changed, 32 insertions(+), 4 deletions(-)
Note: c4pp4 pointed out that in his case, only the speaker needed to go to D3 before pinctl, but noticed no regressions from the below implementation, so I made the fixup more generic (easier to reuse later).
Well, if going to D3 is already achieved there, there is no point to call snd_hda_shutup_pins(). snd_hda_shutup_pins() itself is for reducing the click noise while the codec is going down to D3.
Thus, for this particular machine, simply disable snd_hda_shutup_pins() call from suspend, and make D3 call in reboot_notify.
Looking at the bug report, c4pp4 has just reported back that reverting the old commit 697c373e34613609cb5 improves the situation - there are less clicks compared to the current situation.
Quoting c4pp4 below:
"The patch [697c373e34613609cb5] in fact caused click noise hell because the initial click noise is still alive and in addition there were born new click noises, louder click noise and double click noise. They start to appear when using power-save mode, when suspending computer and when rebooting computer."
So maybe revert that patch?
Yes, feel free to submit the patch.
Takashi