The check before alone is not enough for a case where there is another bug introduced so that
context->stream_count is not in sync with actual number of streams across entire resource_context.
At least assert indeed should be there.
Andrey
On 03/12/2018 07:06 PM, Li, Roman wrote:
There is a check just before for-loop that should ensure pipe_ctx is not null:
/* Only supports single display */ if (context->stream_count != 1) return false;
To remove the subject warning - we can rather add an assert: assert(pipe_ctx);
Thanks, Roman
-----Original Message----- From: Grodzovsky, Andrey Sent: Monday, March 12, 2018 6:23 PM To: David Binderman dcb314@hotmail.com; Deucher, Alexander Alexander.Deucher@amd.com; Koenig, Christian Christian.Koenig@amd.com; Zhou, David(ChunMing) David1.Zhou@amd.com; airlied@linux.ie; amd-gfx@lists.freedesktop.org; dri-devel@lists.freedesktop.org; linux-kernel@vger.kernel.org; Wentland, Harry Harry.Wentland@amd.com; Li, Roman Roman.Li@amd.com Subject: Re: linux-4.16-rc5/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c:1723]: (warning) Possible null pointer dereference: pipe_ctx
On 03/12/2018 06:22 AM, David Binderman wrote:
hello there,
Source code is
for (i = 0; i < dc->res_pool->pipe_count; i++) { if (res_ctx->pipe_ctx[i].stream) { pipe_ctx = &res_ctx->pipe_ctx[i]; *pipe_idx = i; break; } }
Indeed something like
if (!pipe_ctx) return false;
seems the right thing here.
Harry, Roman , your opinion ?
Andrey
/* Only supports eDP */ if (pipe_ctx->stream->sink->link->connector_signal != SIGNAL_TYPE_EDP) return false;
Suggest add some code to deal with the case that the for loop doesn't find what it is looking for and so pipe_ctx is NULL.
Regards
David Binderman
amd-gfx mailing list amd-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/amd-gfx