Am 11.02.21 um 16:02 schrieb Jani Nikula:

On Thu, 11 Feb 2021, Christian König christian.koenig@amd.com wrote:

...

Hi guys,

I had a conflict this morning in the ttm pool while pushing an important fix to drm-misc-fixes.

I'm not 100% sure if I correctly fixed up drm-tip. How can this be double checked? And how can it be fixed if the merge has gone bad?

I'm afraid there's a problem; bb52cb0dec8d ("drm/ttm: make the pool shrinker lock a mutex") in upstream and drm-misc-fixes creates a silent conflict with ba051901d10f ("drm/ttm: add a debugfs file for the global page pools") in drm-misc-next, causing the latter to use spin_lock/unlock on a mutex.

But while you hit a conflict, it does look like the conflict breaking the build is silent, AFAICT the spinlock part does not conflict. So a fixup patch in drm-rerere is probably needed until there are some backmerges.

Well exactly that's the issue. I've already had a fixup in drm-rerere for this.

But today I've pushed another fix to drm-misc-fixes which also conflicts with ba051901d10f ("drm/ttm: add a debugfs file for the global page pools").

I've fixed this up as well and committed the solution. But dim then complained that the original fixup is now not applicable any more (which is true as far as I know).

This somehow went away when I said that dim should assume patch reversal, but I'm not sure if that's the right thing to do.

Thanks, Christian.

BR, Jani.

Am 11.02.21 um 18:22 schrieb Daniel Vetter:

On Thu, Feb 11, 2021 at 4:27 PM Christian König christian.koenig@amd.com wrote:

...

Am 11.02.21 um 16:02 schrieb Jani Nikula:

...

On Thu, 11 Feb 2021, Christian König christian.koenig@amd.com wrote:

...

Hi guys,

I had a conflict this morning in the ttm pool while pushing an important fix to drm-misc-fixes.

I'm not 100% sure if I correctly fixed up drm-tip. How can this be double checked? And how can it be fixed if the merge has gone bad?

I'm afraid there's a problem; bb52cb0dec8d ("drm/ttm: make the pool shrinker lock a mutex") in upstream and drm-misc-fixes creates a silent conflict with ba051901d10f ("drm/ttm: add a debugfs file for the global page pools") in drm-misc-next, causing the latter to use spin_lock/unlock on a mutex.

But while you hit a conflict, it does look like the conflict breaking the build is silent, AFAICT the spinlock part does not conflict. So a fixup patch in drm-rerere is probably needed until there are some backmerges.

Well exactly that's the issue. I've already had a fixup in drm-rerere for this.

But today I've pushed another fix to drm-misc-fixes which also conflicts with ba051901d10f ("drm/ttm: add a debugfs file for the global page pools").

I've fixed this up as well and committed the solution. But dim then complained that the original fixup is now not applicable any more (which is true as far as I know).

This somehow went away when I said that dim should assume patch reversal, but I'm not sure if that's the right thing to do.

Nothing, it's all still broken. You need to delete the now unecessary fixup. As a rule, when rebuilding drm-tip failed always try again to confirm that what you've done actually fixed things (since sometimes git can also not remember where to apply the resolution for some odd reasons).

My question is how to I remove the fixup now?

E.g. where can I find it?

Thanks, Christian.

I'm fixing this now. -Daniel

...

Thanks, Christian.

...

BR, Jani.

Am 12.02.21 um 09:27 schrieb Daniel Vetter:

On Fri, Feb 12, 2021 at 08:51:19AM +0100, Christian König wrote:

...

Am 11.02.21 um 18:22 schrieb Daniel Vetter:

...

On Thu, Feb 11, 2021 at 4:27 PM Christian König christian.koenig@amd.com wrote:

...

Am 11.02.21 um 16:02 schrieb Jani Nikula:

...

On Thu, 11 Feb 2021, Christian König christian.koenig@amd.com wrote:

...

Hi guys,

I had a conflict this morning in the ttm pool while pushing an important fix to drm-misc-fixes.

I'm not 100% sure if I correctly fixed up drm-tip. How can this be double checked? And how can it be fixed if the merge has gone bad?

I'm afraid there's a problem; bb52cb0dec8d ("drm/ttm: make the pool shrinker lock a mutex") in upstream and drm-misc-fixes creates a silent conflict with ba051901d10f ("drm/ttm: add a debugfs file for the global page pools") in drm-misc-next, causing the latter to use spin_lock/unlock on a mutex.

But while you hit a conflict, it does look like the conflict breaking the build is silent, AFAICT the spinlock part does not conflict. So a fixup patch in drm-rerere is probably needed until there are some backmerges.

Well exactly that's the issue. I've already had a fixup in drm-rerere for this.

But today I've pushed another fix to drm-misc-fixes which also conflicts with ba051901d10f ("drm/ttm: add a debugfs file for the global page pools").

I've fixed this up as well and committed the solution. But dim then complained that the original fixup is now not applicable any more (which is true as far as I know).

This somehow went away when I said that dim should assume patch reversal, but I'm not sure if that's the right thing to do.

Nothing, it's all still broken. You need to delete the now unecessary fixup. As a rule, when rebuilding drm-tip failed always try again to confirm that what you've done actually fixed things (since sometimes git can also not remember where to apply the resolution for some odd reasons).

My question is how to I remove the fixup now?

E.g. where can I find it?

In the rerere branch under the fixup/ directory. That's also the place where rerere solutions are carried around. I guess docs don't explain this well how drm-tip is rebuilt :-/

Ok found it. drm-tip looks clean now again.

Thanks, Christian.

-Daniel

...

Thanks, Christian.

...

I'm fixing this now. -Daniel

...

Thanks, Christian.

...

BR, Jani.