Messages in this thread | | | Date | Mon, 9 Jan 2023 20:24:16 +0200 | Subject | Re: [PATCH v2 4/8] drm/msm/dpu: Disallow unallocated resources to be returned | From | Dmitry Baryshkov <> |
| |
On 09/01/2023 19:12, Marijn Suijten wrote: > On 2023-01-09 11:06:45, Dmitry Baryshkov wrote: >> On Mon, 9 Jan 2023 at 10:24, Marijn Suijten >> <marijn.suijten@somainline.org> wrote: >>> >>> On 2023-01-09 01:30:29, Dmitry Baryshkov wrote: >>>> On 09/01/2023 01:28, Dmitry Baryshkov wrote: >>>>> On 22/12/2022 01:19, Marijn Suijten wrote: >>>>>> In the event that the topology requests resources that have not been >>>>>> created by the system (because they are typically not represented in >>>>>> dpu_mdss_cfg ^1), the resource(s) in global_state (in this case DSC >>>>>> blocks) remain NULL but will still be returned out of >>>>>> dpu_rm_get_assigned_resources, where the caller expects to get an array >>>>>> containing num_blks valid pointers (but instead gets these NULLs). >>>>>> >>>>>> To prevent this from happening, where null-pointer dereferences >>>>>> typically result in a hard-to-debug platform lockup, num_blks shouldn't >>>>>> increase past NULL blocks and will print an error and break instead. >>>>>> After all, max_blks represents the static size of the maximum number of >>>>>> blocks whereas the actual amount varies per platform. >>>>>> >>>>>> ^1: which can happen after a git rebase ended up moving additions to >>>>>> _dpu_cfg to a different struct which has the same patch context. >>>>>> >>>>>> Fixes: bb00a452d6f7 ("drm/msm/dpu: Refactor resource manager") >>>>>> Signed-off-by: Marijn Suijten <marijn.suijten@somainline.org> >>>>>> --- >>>>>> drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c | 5 +++++ >>>>>> 1 file changed, 5 insertions(+) >>>>> >>>>> I think the patch is not fully correct. Please check resource >>>>> availability during allocation. I wouldn't expect an error from >>>>> get_assigned_resources because of resource exhaustion. >>> >>> Theoretically patch 5/8 should take care of this, and we should never >>> reach this failure condition. Emphasis on /should/, this may happen >>> again if/when another block type is added with sub-par resource >>> allocation and assignment implementation. >> >> Yeah. Maybe swapping 4/8 and 5/8 makes sense. > > Ack. > >>>> Another option, since allocation functions (except DSC) already have >>>> these safety checks: check error message to mention internal >>>> inconstency: allocated resource doesn't exist. >>> >>> Is this a suggestion for the wording of the error message? >> >> Yes. Because the current message makes one think that it is output >> during allocation / assignment to encoder, while this is a safety net. > > Good. So the patch is correct, just the wording is off, which I fully > agree on. This isn't allocating anything, just handing out what was > previously allocated (and is a safety net).
Yes. Please excuse me if my original message was not 100% clear.
> > - Marijn
-- With best wishes Dmitry
| |