On Wed, Oct 4, 2017 at 4:16 AM, sathyanarayanan kuppuswamy <sathyanarayanan.kuppuswamy@xxxxxxxxxxxxxxx> wrote: > On 10/01/2017 07:48 AM, Andy Shevchenko wrote: >> On Tue, Sep 5, 2017 at 8:37 AM, >> <sathyanarayanan.kuppuswamy@xxxxxxxxxxxxxxx> wrote: >> Since it sounds as candidate for stable, > > Yes. >> >> can we have split it to just >> as less as possible intrusive fix + moving to regmap as a separate >> change? > > If we have to split it into two patches then, > > Patch #1 will fix the "sleep in atomic context issue" by replacing > mutex_lock() with spin_lock() > in GCR read/write APIs to protect the GCR memory updates. > Patch #2 will remove GCR read/write/update APIs and replace it with regmap > APIs. But along with this > change we will also remove the spin_lock() added in previous patch because > regmap calls are already > protected by its own locking mechanism. > > Since Patch #2 will clean up what we do in Patch #1, Do we need to split it > into two patches? Yes, please do... >> It should include Fixes: tag as well I suppose. > > Agree. I will add Fixes tag in next version. ...because this one will go alone to stable releases. Be also sure patch #1 will be applied on current vanilla (w/o PDx86 queue involvement). -- With Best Regards, Andy Shevchenko
