| Message ID | 20230421134654.273182-1-andi.shyti@linux.intel.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | Use i915 instead of dev_priv | expand |
Hi, On Fri, Apr 21, 2023 at 03:46:52PM +0200, Andi Shyti wrote: > Hi, > > just another "Friday patch". While reviewing some patches from > Tejas I found a bit confusing the use of dev_priv__ inside the > for_each_engine(), perhaps it should be moved inside the gt/? > > As I was at it I made the /dev_priv/i915/ change which is still > harmless. Next in queue is to change the i915_irq.h, which is a > bit tricky (but not much) as the "dev_priv" is hardcoded inside > some defines. > > Andi > > Andi Shyti (2): > drm/i915/i915_drv: Use proper parameter naming in for_each_gt() > drm/i915/i915_drv: Use i915 instead of dev_priv insied the file_priv > structure > > drivers/gpu/drm/i915/i915_drv.h | 462 ++++++++++++++++---------------- > 1 file changed, 231 insertions(+), 231 deletions(-) Pushed to dmr-intel-gt-next. Thanks Rodrigo and Andrzej for your review. Andi
On Sun, 23 Apr 2023, Andi Shyti <andi.shyti@linux.intel.com> wrote: > Hi, > > On Fri, Apr 21, 2023 at 03:46:52PM +0200, Andi Shyti wrote: >> Hi, >> >> just another "Friday patch". While reviewing some patches from >> Tejas I found a bit confusing the use of dev_priv__ inside the >> for_each_engine(), perhaps it should be moved inside the gt/? >> >> As I was at it I made the /dev_priv/i915/ change which is still >> harmless. Next in queue is to change the i915_irq.h, which is a >> bit tricky (but not much) as the "dev_priv" is hardcoded inside >> some defines. >> >> Andi >> >> Andi Shyti (2): >> drm/i915/i915_drv: Use proper parameter naming in for_each_gt() >> drm/i915/i915_drv: Use i915 instead of dev_priv insied the file_priv >> structure >> >> drivers/gpu/drm/i915/i915_drv.h | 462 ++++++++++++++++---------------- >> 1 file changed, 231 insertions(+), 231 deletions(-) > > Pushed to dmr-intel-gt-next. That's going to create problems for us for weeks to come. I'm actually tempted to ask Joonas or Tvrtko to just force push that out of there. Only use drm-intel-gt-next for stuff that's specifically about gt or gem, and touches files used by gt or gem only. For everything else, use drm-intel-next. When in doubt, err on the side of drm-intel-next. It's not enough that one of the patches changes parameters of for_each_gt() macro. We can cross-merge drm-intel-next to drm-intel-gt-next, but we can't cross-merge drm-intel-gt-next to drm-intel-next. This means the only way to sync those i915_drv.h changes to drm-intel-next is to have a drm-intel-gt-next pull request merged to drm-next, and then backmerged to drm-intel-next. That's not going to happen for several weeks. Any change aimed at drm-intel-next that conflicts with the i915_drv.h changes will now be pending on those merges. BR, Jani. > > Thanks Rodrigo and Andrzej for your review. > > Andi
Hi Jani, > >> just another "Friday patch". While reviewing some patches from > >> Tejas I found a bit confusing the use of dev_priv__ inside the > >> for_each_engine(), perhaps it should be moved inside the gt/? > >> > >> As I was at it I made the /dev_priv/i915/ change which is still > >> harmless. Next in queue is to change the i915_irq.h, which is a > >> bit tricky (but not much) as the "dev_priv" is hardcoded inside > >> some defines. > >> > >> Andi > >> > >> Andi Shyti (2): > >> drm/i915/i915_drv: Use proper parameter naming in for_each_gt() > >> drm/i915/i915_drv: Use i915 instead of dev_priv insied the file_priv > >> structure > >> > >> drivers/gpu/drm/i915/i915_drv.h | 462 ++++++++++++++++---------------- > >> 1 file changed, 231 insertions(+), 231 deletions(-) > > > > Pushed to dmr-intel-gt-next. > > That's going to create problems for us for weeks to come. I'm actually > tempted to ask Joonas or Tvrtko to just force push that out of there. > > Only use drm-intel-gt-next for stuff that's specifically about gt or > gem, and touches files used by gt or gem only. For everything else, use > drm-intel-next. When in doubt, err on the side of drm-intel-next. sorry, I did think of it. But... > It's not enough that one of the patches changes parameters of > for_each_gt() macro. ... I was fooled by the gt/i915 parameter. Thanks and sorry, Andi > We can cross-merge drm-intel-next to drm-intel-gt-next, but we can't > cross-merge drm-intel-gt-next to drm-intel-next. This means the only way > to sync those i915_drv.h changes to drm-intel-next is to have a > drm-intel-gt-next pull request merged to drm-next, and then backmerged > to drm-intel-next. That's not going to happen for several weeks. > > Any change aimed at drm-intel-next that conflicts with the i915_drv.h > changes will now be pending on those merges.
Hi, just another "Friday patch". While reviewing some patches from Tejas I found a bit confusing the use of dev_priv__ inside the for_each_engine(), perhaps it should be moved inside the gt/? As I was at it I made the /dev_priv/i915/ change which is still harmless. Next in queue is to change the i915_irq.h, which is a bit tricky (but not much) as the "dev_priv" is hardcoded inside some defines. Andi Andi Shyti (2): drm/i915/i915_drv: Use proper parameter naming in for_each_gt() drm/i915/i915_drv: Use i915 instead of dev_priv insied the file_priv structure drivers/gpu/drm/i915/i915_drv.h | 462 ++++++++++++++++---------------- 1 file changed, 231 insertions(+), 231 deletions(-)