On Wed, 2025-01-22 at 21:11 -0800, Lucas De Marchi wrote:
> Having the exec queue snapshot inside a "GuC CT" section was always
> wrong. Commit c28fd6c358db ("drm/xe/devcoredump: Improve section
> headings and add tile info") tried to fix that bug, but with that also
> broke the mesa tool that parses the devcoredump, hence it was reverted
> in commit 70fb86a85dc9 ("drm/xe: Revert some changes that break a mesa
> debug tool").
>
> With the mesa tool also fixed, this can propagate as a fix on both
> kernel and userspace side to avoid unnecessary headache for a debug
> feature.
This will break older versions of the Mesa parser. Is this really necessary?
Is it worth breaking the tool? In my opinion, it is not.
Also, do we need to discuss this now? Wouldn't it be better to focus on bringing the GuC log in first?
>
> Cc: John Harrison <John.C.Harrison@xxxxxxxxx>
> Cc: Julia Filipchuk <julia.filipchuk@xxxxxxxxx>
> Cc: José Roberto de Souza <jose.souza@xxxxxxxxx>
> Cc: stable@xxxxxxxxxxxxxxx
> Fixes: 70fb86a85dc9 ("drm/xe: Revert some changes that break a mesa debug tool")
> Signed-off-by: Lucas De Marchi <lucas.demarchi@xxxxxxxxx>
> ---
> drivers/gpu/drm/xe/xe_devcoredump.c | 6 +-----
> 1 file changed, 1 insertion(+), 5 deletions(-)
>
> diff --git a/drivers/gpu/drm/xe/xe_devcoredump.c b/drivers/gpu/drm/xe/xe_devcoredump.c
> index 81dc7795c0651..a7946a76777e7 100644
> --- a/drivers/gpu/drm/xe/xe_devcoredump.c
> +++ b/drivers/gpu/drm/xe/xe_devcoredump.c
> @@ -119,11 +119,7 @@ static ssize_t __xe_devcoredump_read(char *buffer, size_t count,
> drm_puts(&p, "\n**** GuC CT ****\n");
> xe_guc_ct_snapshot_print(ss->guc.ct, &p);
>
> - /*
> - * Don't add a new section header here because the mesa debug decoder
> - * tool expects the context information to be in the 'GuC CT' section.
> - */
> - /* drm_puts(&p, "\n**** Contexts ****\n"); */
> + drm_puts(&p, "\n**** Contexts ****\n");
> xe_guc_exec_queue_snapshot_print(ss->ge, &p);
>
> drm_puts(&p, "\n**** Job ****\n");
