| Message ID | 20200805122231.23313-1-chris@chris-wilson.co.uk (mailing list archive) |
|---|---|
| Headers | show |
| Series | Replace obj->mm.lock with reservation_ww_class | expand |
Hi, Chris, On 8/5/20 2:21 PM, Chris Wilson wrote: > Long story short, we need to manage evictions using dma_resv & dma_fence > tracking. The backing storage will then be managed using the ww_mutex > borrowed from (and shared via) obj->base.resv, rather than the current > obj->mm.lock. > > Skipping over the breadcrumbs, While perhaps needed fixes, could we submit them as a separate series, since they, from what I can tell, are not a direct part of the locking rework, and some of them were actually part of a series that Dave NaK'ed and may require additional justification? > the first step is to remove the final > crutches of struct_mutex from execbuf and to broaden the hold for the > dma-resv to guard not just publishing the dma-fences, but for the > duration of the execbuf submission (holding all objects and their > backing store from the point of acquisition to publishing of the final > GPU work, after which the guard is delegated to the dma-fences). > > This is of course made complicated by our history. On top of the user's > objects, we also have the HW/kernel objects with their own lifetimes, > and a bunch of auxiliary objects used for working around unhappy HW and > for providing the legacy relocation mechanism. We add every auxiliary > object to the list of user objects required, and attempt to acquire them > en masse. Since all the objects can be known a priori, we can build a > list of those objects and pass that to a routine that can resolve the > -EDEADLK (and evictions). [To avoid relocations imposing a penalty on > sane userspace that avoids them, we do not touch any relocations until > necessary, at will point we have to unroll the state, and rebuild a new > list with more auxiliary buffers to accommodate the extra copy_from_user]. > More examples are included as to how we can break down operations > involving multiple objects into an acquire phase prior to those > operations, keeping the -EDEADLK handling under control. > > execbuf is the unique interface in that it deals with multiple user > and kernel buffers. After that, we have callers that in principle care > about accessing a single buffer, and so can be migrated over to a helper > that permits only holding one such buffer at a time. That enables us to > swap out obj->mm.lock for obj->base.resv->lock, and use lockdep to spot > illegal nesting, and to throw away the temporary pins by replacing them > with holding the ww_mutex for the duration instead. > > What's changed? Some patch splitting and we need to pull in Matthew's > patch to map the page directories under the ww_mutex. I would still like to see a justification for the newly introduced async work, as opposed to add it as an optimizing / regression fixing series follow the locking rework. That async work introduces a bunch of code complexity and it would be beneficial to see a discussion of the tradeoffs and how it alignes with the upstream proposed dma-fence annotations Thanks, Thomas > _______________________________________________ > Intel-gfx mailing list > Intel-gfx@lists.freedesktop.org > https://lists.freedesktop.org/mailman/listinfo/intel-gfx
On 05/08/2020 17:22, Thomas Hellström (Intel) wrote: > Hi, Chris, > > > On 8/5/20 2:21 PM, Chris Wilson wrote: >> Long story short, we need to manage evictions using dma_resv & dma_fence >> tracking. The backing storage will then be managed using the ww_mutex >> borrowed from (and shared via) obj->base.resv, rather than the current >> obj->mm.lock. >> >> Skipping over the breadcrumbs, > > While perhaps needed fixes, could we submit them as a separate series, > since they, from what I can tell, are not a direct part of the locking > rework, and some of them were actually part of a series that Dave NaK'ed > and may require additional justification? > > >> the first step is to remove the final >> crutches of struct_mutex from execbuf and to broaden the hold for the >> dma-resv to guard not just publishing the dma-fences, but for the >> duration of the execbuf submission (holding all objects and their >> backing store from the point of acquisition to publishing of the final >> GPU work, after which the guard is delegated to the dma-fences). >> >> This is of course made complicated by our history. On top of the user's >> objects, we also have the HW/kernel objects with their own lifetimes, >> and a bunch of auxiliary objects used for working around unhappy HW and >> for providing the legacy relocation mechanism. We add every auxiliary >> object to the list of user objects required, and attempt to acquire them >> en masse. Since all the objects can be known a priori, we can build a >> list of those objects and pass that to a routine that can resolve the >> -EDEADLK (and evictions). [To avoid relocations imposing a penalty on >> sane userspace that avoids them, we do not touch any relocations until >> necessary, at will point we have to unroll the state, and rebuild a new >> list with more auxiliary buffers to accommodate the extra >> copy_from_user]. >> More examples are included as to how we can break down operations >> involving multiple objects into an acquire phase prior to those >> operations, keeping the -EDEADLK handling under control. >> >> execbuf is the unique interface in that it deals with multiple user >> and kernel buffers. After that, we have callers that in principle care >> about accessing a single buffer, and so can be migrated over to a helper >> that permits only holding one such buffer at a time. That enables us to >> swap out obj->mm.lock for obj->base.resv->lock, and use lockdep to spot >> illegal nesting, and to throw away the temporary pins by replacing them >> with holding the ww_mutex for the duration instead. >> >> What's changed? Some patch splitting and we need to pull in Matthew's >> patch to map the page directories under the ww_mutex. > > I would still like to see a justification for the newly introduced async > work, as opposed to add it as an optimizing / regression fixing series > follow the locking rework. That async work introduces a bunch of code > complexity and it would be beneficial to see a discussion of the > tradeoffs and how it alignes with the upstream proposed dma-fence > annotations On the topic of annotations, maybe do a trybot run with them enabled with the latest series and then see what pops up. +Daniel, since I noticed last time he was doing that one of the splats (possibly the only one?) was actually caused by dma_fence_is_signaled. Which I think comes under the opportunistic signaling rule for the annotation kerneldoc so looked like a false positive to me. Not sure how to avoid that one, apart from making it call a special, un-annotated, flavours of dma_fence_signal(_locked). Regards, Tvrtko
On Thu, Aug 6, 2020 at 11:21 AM Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com> wrote: > > > On 05/08/2020 17:22, Thomas Hellström (Intel) wrote: > > Hi, Chris, > > > > > > On 8/5/20 2:21 PM, Chris Wilson wrote: > >> Long story short, we need to manage evictions using dma_resv & dma_fence > >> tracking. The backing storage will then be managed using the ww_mutex > >> borrowed from (and shared via) obj->base.resv, rather than the current > >> obj->mm.lock. > >> > >> Skipping over the breadcrumbs, > > > > While perhaps needed fixes, could we submit them as a separate series, > > since they, from what I can tell, are not a direct part of the locking > > rework, and some of them were actually part of a series that Dave NaK'ed > > and may require additional justification? > > > > > >> the first step is to remove the final > >> crutches of struct_mutex from execbuf and to broaden the hold for the > >> dma-resv to guard not just publishing the dma-fences, but for the > >> duration of the execbuf submission (holding all objects and their > >> backing store from the point of acquisition to publishing of the final > >> GPU work, after which the guard is delegated to the dma-fences). > >> > >> This is of course made complicated by our history. On top of the user's > >> objects, we also have the HW/kernel objects with their own lifetimes, > >> and a bunch of auxiliary objects used for working around unhappy HW and > >> for providing the legacy relocation mechanism. We add every auxiliary > >> object to the list of user objects required, and attempt to acquire them > >> en masse. Since all the objects can be known a priori, we can build a > >> list of those objects and pass that to a routine that can resolve the > >> -EDEADLK (and evictions). [To avoid relocations imposing a penalty on > >> sane userspace that avoids them, we do not touch any relocations until > >> necessary, at will point we have to unroll the state, and rebuild a new > >> list with more auxiliary buffers to accommodate the extra > >> copy_from_user]. > >> More examples are included as to how we can break down operations > >> involving multiple objects into an acquire phase prior to those > >> operations, keeping the -EDEADLK handling under control. > >> > >> execbuf is the unique interface in that it deals with multiple user > >> and kernel buffers. After that, we have callers that in principle care > >> about accessing a single buffer, and so can be migrated over to a helper > >> that permits only holding one such buffer at a time. That enables us to > >> swap out obj->mm.lock for obj->base.resv->lock, and use lockdep to spot > >> illegal nesting, and to throw away the temporary pins by replacing them > >> with holding the ww_mutex for the duration instead. > >> > >> What's changed? Some patch splitting and we need to pull in Matthew's > >> patch to map the page directories under the ww_mutex. > > > > I would still like to see a justification for the newly introduced async > > work, as opposed to add it as an optimizing / regression fixing series > > follow the locking rework. That async work introduces a bunch of code > > complexity and it would be beneficial to see a discussion of the > > tradeoffs and how it alignes with the upstream proposed dma-fence > > annotations > > On the topic of annotations, maybe do a trybot run with them enabled > with the latest series and then see what pops up. > > +Daniel, since I noticed last time he was doing that one of the splats > (possibly the only one?) was actually caused by dma_fence_is_signaled. > Which I think comes under the opportunistic signaling rule for the > annotation kerneldoc so looked like a false positive to me. Not sure how > to avoid that one, apart from making it call a special, un-annotated, > flavours of dma_fence_signal(_locked). Yeah that became a bit more constrained due to the switch from recursive locks (which don't catch the actual wait vs signalling issues because lockdep is not as good as it should be) to explicitly recursive read locks (which do catch the wait vs signal side of things, but get more annoyed about creative locking schemes on the read (i.e. signalling side). -Daniel
On 06/08/2020 12:55, Daniel Vetter wrote: > On Thu, Aug 6, 2020 at 11:21 AM Tvrtko Ursulin > <tvrtko.ursulin@linux.intel.com> wrote: >> On 05/08/2020 17:22, Thomas Hellström (Intel) wrote: >>> Hi, Chris, >>> On 8/5/20 2:21 PM, Chris Wilson wrote: >>>> Long story short, we need to manage evictions using dma_resv & dma_fence >>>> tracking. The backing storage will then be managed using the ww_mutex >>>> borrowed from (and shared via) obj->base.resv, rather than the current >>>> obj->mm.lock. >>>> >>>> Skipping over the breadcrumbs, >>> >>> While perhaps needed fixes, could we submit them as a separate series, >>> since they, from what I can tell, are not a direct part of the locking >>> rework, and some of them were actually part of a series that Dave NaK'ed >>> and may require additional justification? >>> >>> >>>> the first step is to remove the final >>>> crutches of struct_mutex from execbuf and to broaden the hold for the >>>> dma-resv to guard not just publishing the dma-fences, but for the >>>> duration of the execbuf submission (holding all objects and their >>>> backing store from the point of acquisition to publishing of the final >>>> GPU work, after which the guard is delegated to the dma-fences). >>>> >>>> This is of course made complicated by our history. On top of the user's >>>> objects, we also have the HW/kernel objects with their own lifetimes, >>>> and a bunch of auxiliary objects used for working around unhappy HW and >>>> for providing the legacy relocation mechanism. We add every auxiliary >>>> object to the list of user objects required, and attempt to acquire them >>>> en masse. Since all the objects can be known a priori, we can build a >>>> list of those objects and pass that to a routine that can resolve the >>>> -EDEADLK (and evictions). [To avoid relocations imposing a penalty on >>>> sane userspace that avoids them, we do not touch any relocations until >>>> necessary, at will point we have to unroll the state, and rebuild a new >>>> list with more auxiliary buffers to accommodate the extra >>>> copy_from_user]. >>>> More examples are included as to how we can break down operations >>>> involving multiple objects into an acquire phase prior to those >>>> operations, keeping the -EDEADLK handling under control. >>>> >>>> execbuf is the unique interface in that it deals with multiple user >>>> and kernel buffers. After that, we have callers that in principle care >>>> about accessing a single buffer, and so can be migrated over to a helper >>>> that permits only holding one such buffer at a time. That enables us to >>>> swap out obj->mm.lock for obj->base.resv->lock, and use lockdep to spot >>>> illegal nesting, and to throw away the temporary pins by replacing them >>>> with holding the ww_mutex for the duration instead. >>>> >>>> What's changed? Some patch splitting and we need to pull in Matthew's >>>> patch to map the page directories under the ww_mutex. >>> >>> I would still like to see a justification for the newly introduced async >>> work, as opposed to add it as an optimizing / regression fixing series >>> follow the locking rework. That async work introduces a bunch of code >>> complexity and it would be beneficial to see a discussion of the >>> tradeoffs and how it alignes with the upstream proposed dma-fence >>> annotations >> >> On the topic of annotations, maybe do a trybot run with them enabled >> with the latest series and then see what pops up. >> >> +Daniel, since I noticed last time he was doing that one of the splats >> (possibly the only one?) was actually caused by dma_fence_is_signaled. >> Which I think comes under the opportunistic signaling rule for the >> annotation kerneldoc so looked like a false positive to me. Not sure how >> to avoid that one, apart from making it call a special, un-annotated, >> flavours of dma_fence_signal(_locked). > > Yeah that became a bit more constrained due to the switch from > recursive locks (which don't catch the actual wait vs signalling > issues because lockdep is not as good as it should be) to explicitly > recursive read locks (which do catch the wait vs signal side of > things, but get more annoyed about creative locking schemes on the > read (i.e. signalling side). I did not get you, what became more constrained? Allowance for opportunistic signaling got retracted? But dma_fence_is_signaled is not even that, I don't see how it makes sense to flag it as foul. Or you agree dma_fence_is_signaled should be made call new non-annotated helpers as I suggested? Regards, Tvrtko
Quoting Tvrtko Ursulin (2020-08-06 10:21:38) > > On 05/08/2020 17:22, Thomas Hellström (Intel) wrote: > > Hi, Chris, > > > > > > On 8/5/20 2:21 PM, Chris Wilson wrote: > >> Long story short, we need to manage evictions using dma_resv & dma_fence > >> tracking. The backing storage will then be managed using the ww_mutex > >> borrowed from (and shared via) obj->base.resv, rather than the current > >> obj->mm.lock. > >> > >> Skipping over the breadcrumbs, > > > > While perhaps needed fixes, could we submit them as a separate series, > > since they, from what I can tell, are not a direct part of the locking > > rework, and some of them were actually part of a series that Dave NaK'ed > > and may require additional justification? > > > > > >> the first step is to remove the final > >> crutches of struct_mutex from execbuf and to broaden the hold for the > >> dma-resv to guard not just publishing the dma-fences, but for the > >> duration of the execbuf submission (holding all objects and their > >> backing store from the point of acquisition to publishing of the final > >> GPU work, after which the guard is delegated to the dma-fences). > >> > >> This is of course made complicated by our history. On top of the user's > >> objects, we also have the HW/kernel objects with their own lifetimes, > >> and a bunch of auxiliary objects used for working around unhappy HW and > >> for providing the legacy relocation mechanism. We add every auxiliary > >> object to the list of user objects required, and attempt to acquire them > >> en masse. Since all the objects can be known a priori, we can build a > >> list of those objects and pass that to a routine that can resolve the > >> -EDEADLK (and evictions). [To avoid relocations imposing a penalty on > >> sane userspace that avoids them, we do not touch any relocations until > >> necessary, at will point we have to unroll the state, and rebuild a new > >> list with more auxiliary buffers to accommodate the extra > >> copy_from_user]. > >> More examples are included as to how we can break down operations > >> involving multiple objects into an acquire phase prior to those > >> operations, keeping the -EDEADLK handling under control. > >> > >> execbuf is the unique interface in that it deals with multiple user > >> and kernel buffers. After that, we have callers that in principle care > >> about accessing a single buffer, and so can be migrated over to a helper > >> that permits only holding one such buffer at a time. That enables us to > >> swap out obj->mm.lock for obj->base.resv->lock, and use lockdep to spot > >> illegal nesting, and to throw away the temporary pins by replacing them > >> with holding the ww_mutex for the duration instead. > >> > >> What's changed? Some patch splitting and we need to pull in Matthew's > >> patch to map the page directories under the ww_mutex. > > > > I would still like to see a justification for the newly introduced async > > work, as opposed to add it as an optimizing / regression fixing series > > follow the locking rework. That async work introduces a bunch of code > > complexity and it would be beneficial to see a discussion of the > > tradeoffs and how it alignes with the upstream proposed dma-fence > > annotations > > On the topic of annotations, maybe do a trybot run with them enabled > with the latest series and then see what pops up. It didn't change since the run Daniel did. In that run there were two splats I found, vm->mutex -> dma_fence_is_signaled/dma_fence_signal -> async work + vm->mutex This is the overconstraint Daniel mentioned, it's an entirely false coupling from the assumption that the async work runs inside the parent's signaling context. Gut feeling says that should be resolvable by using lockdep's dependency analysis between local execution/signal contexts. The other was userptr mmu_notifier -> i915_vma_unbind + i915_vma_revoke -> mmu_notifier [with the interplay between the two global lockmap's, two strikes?] And that is the well known false positive for userptr that we break by never allowing the userptr to in be the dev->mapping. A path not exercised but it should find is synchronous fence eviction. While binding a fresh fence ensures the vma is idle before starting, we do not do the same for the fence we steal. That's why care was taken for execbuf, and it just means teaching the synchronous path to similarly move the wait outside of the mutex (the simplest will be to wrap the async handling). -Chris
On 10/08/2020 10:51, Chris Wilson wrote: > Quoting Tvrtko Ursulin (2020-08-06 10:21:38) >> >> On 05/08/2020 17:22, Thomas Hellström (Intel) wrote: >>> Hi, Chris, >>> >>> >>> On 8/5/20 2:21 PM, Chris Wilson wrote: >>>> Long story short, we need to manage evictions using dma_resv & dma_fence >>>> tracking. The backing storage will then be managed using the ww_mutex >>>> borrowed from (and shared via) obj->base.resv, rather than the current >>>> obj->mm.lock. >>>> >>>> Skipping over the breadcrumbs, >>> >>> While perhaps needed fixes, could we submit them as a separate series, >>> since they, from what I can tell, are not a direct part of the locking >>> rework, and some of them were actually part of a series that Dave NaK'ed >>> and may require additional justification? >>> >>> >>>> the first step is to remove the final >>>> crutches of struct_mutex from execbuf and to broaden the hold for the >>>> dma-resv to guard not just publishing the dma-fences, but for the >>>> duration of the execbuf submission (holding all objects and their >>>> backing store from the point of acquisition to publishing of the final >>>> GPU work, after which the guard is delegated to the dma-fences). >>>> >>>> This is of course made complicated by our history. On top of the user's >>>> objects, we also have the HW/kernel objects with their own lifetimes, >>>> and a bunch of auxiliary objects used for working around unhappy HW and >>>> for providing the legacy relocation mechanism. We add every auxiliary >>>> object to the list of user objects required, and attempt to acquire them >>>> en masse. Since all the objects can be known a priori, we can build a >>>> list of those objects and pass that to a routine that can resolve the >>>> -EDEADLK (and evictions). [To avoid relocations imposing a penalty on >>>> sane userspace that avoids them, we do not touch any relocations until >>>> necessary, at will point we have to unroll the state, and rebuild a new >>>> list with more auxiliary buffers to accommodate the extra >>>> copy_from_user]. >>>> More examples are included as to how we can break down operations >>>> involving multiple objects into an acquire phase prior to those >>>> operations, keeping the -EDEADLK handling under control. >>>> >>>> execbuf is the unique interface in that it deals with multiple user >>>> and kernel buffers. After that, we have callers that in principle care >>>> about accessing a single buffer, and so can be migrated over to a helper >>>> that permits only holding one such buffer at a time. That enables us to >>>> swap out obj->mm.lock for obj->base.resv->lock, and use lockdep to spot >>>> illegal nesting, and to throw away the temporary pins by replacing them >>>> with holding the ww_mutex for the duration instead. >>>> >>>> What's changed? Some patch splitting and we need to pull in Matthew's >>>> patch to map the page directories under the ww_mutex. >>> >>> I would still like to see a justification for the newly introduced async >>> work, as opposed to add it as an optimizing / regression fixing series >>> follow the locking rework. That async work introduces a bunch of code >>> complexity and it would be beneficial to see a discussion of the >>> tradeoffs and how it alignes with the upstream proposed dma-fence >>> annotations >> >> On the topic of annotations, maybe do a trybot run with them enabled >> with the latest series and then see what pops up. > > It didn't change since the run Daniel did. In that run there were two > splats I found, > > vm->mutex -> dma_fence_is_signaled/dma_fence_signal -> async work + vm->mutex Coming back to this after a few weeks - was there any further discussion on the dma_fence_is_signaled problem? I don't think Daniel you ever replied, or at least I did not see it. Can the signaling section in there possibly remain and not be obviously incorrect? Regards, Tvrtko > > This is the overconstraint Daniel mentioned, it's an entirely false > coupling from the assumption that the async work runs inside the > parent's signaling context. Gut feeling says that should be resolvable > by using lockdep's dependency analysis between local execution/signal > contexts. > > The other was > > userptr mmu_notifier -> i915_vma_unbind + i915_vma_revoke -> mmu_notifier > [with the interplay between the two global lockmap's, two strikes?] > > And that is the well known false positive for userptr that we break by > never allowing the userptr to in be the dev->mapping. > > A path not exercised but it should find is synchronous fence eviction. > While binding a fresh fence ensures the vma is idle before starting, we > do not do the same for the fence we steal. That's why care was taken for > execbuf, and it just means teaching the synchronous path to similarly > move the wait outside of the mutex (the simplest will be to wrap the > async handling). > -Chris >