On Tue, 23 Oct 2018 10:49:27 +0200, Vinod wrote:

On 22-10-18, 12:22, Pierre-Louis Bossart wrote:

...

Looking at the latest for-next branch, by accident I saw a number of GFP_ATOMIC references.

With my extended break I don't remember much but didn't we agree that none of those were needed? If yes I'll send a patch to fix all this.

I thought that someone had fixed up all these instances,

Looking back at the git history, "someone" was mostly you :)

A few of them below seem to be new codes that copied the bad past examples, some are overlooked one, and some are mailbox codes that aren't 100% sure to safe to replace (although I guess it is).

looks like few missed, so yes please do fix them up and Welcome back :-)

I hope you enjoyed the time-off, Pierre.

thanks,

Takashi

...

Thanks

-Pierre

/data/pbossart/ktest/broonie-next/sound/soc/intel$ git grep ATOMIC atom/sst/sst_pvt.c:     msg = kzalloc(sizeof(struct ipc_post), GFP_ATOMIC); atom/sst/sst_pvt.c:             msg->mailbox_data = kzalloc(SST_MAILBOX_SIZE, GFP_ATOMIC); boards/cht_bsw_rt5672.c:        drv = devm_kzalloc(&pdev->dev, sizeof(*drv), GFP_ATOMIC); boards/glk_rt5682_max98357a.c:  ctx = devm_kzalloc(&pdev->dev, sizeof(*ctx), GFP_ATOMIC); boards/kbl_da7219_max98927.c:   ctx = devm_kzalloc(&pdev->dev, sizeof(*ctx), GFP_ATOMIC); boards/skl_hda_dsp_generic.c:   ctx = devm_kzalloc(&pdev->dev, sizeof(*ctx), GFP_ATOMIC);

-- ~Vinod