public inbox for intel-gfx@lists.freedesktop.org
 help / color / mirror / Atom feed
From: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
To: Fernando Pacheco <fernando.pacheco@intel.com>,
	intel-gfx@lists.freedesktop.org
Subject: Re: [PATCH] drm/i915/uc: Fini hw even if GuC is not running
Date: Tue, 13 Aug 2019 13:16:53 -0700	[thread overview]
Message-ID: <39d7f313-dec7-6432-af08-a9127d3032ce@intel.com> (raw)
In-Reply-To: <20190813162628.18531-1-fernando.pacheco@intel.com>



On 8/13/19 9:26 AM, Fernando Pacheco wrote:
> We should not be skipping uc_fini_hw on finding GuC
> is no longer running. There is plenty of hw and internal
> state that can be cleaned up without having to communicate
> with GuC.
> 

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=110943

> Signed-off-by: Fernando Pacheco <fernando.pacheco@intel.com>
> Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
> Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
> ---
>   drivers/gpu/drm/i915/gt/uc/intel_uc.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc.c b/drivers/gpu/drm/i915/gt/uc/intel_uc.c
> index 0dc2b0cf4604..c698cddc14dc 100644
> --- a/drivers/gpu/drm/i915/gt/uc/intel_uc.c
> +++ b/drivers/gpu/drm/i915/gt/uc/intel_uc.c
> @@ -521,7 +521,7 @@ void intel_uc_fini_hw(struct intel_uc *uc)
>   {
>   	struct intel_guc *guc = &uc->guc;
>   
> -	if (!intel_guc_is_running(guc))
> +	if (!intel_uc_supports_guc(uc))

Both calls below handle the case where GuC is already not running so 
we're safe, but now that we wedge when we hit a guc error we can also do 
something like:

	-EIO load error -> uc_fini_hw() -> wedge
and then
	unload -> uc_fini_hw()

it should be all be handled safely (the fault injection test is passing 
where we've run it), but we end up with "GuC communication disabled" 
multiple times in the logs:

https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_13999/fi-skl-guc/igt@i915_module_load@reload-with-fault-injection.html

<6> [237.818905] [drm] GuC communication enabled
....
<6> [237.822739] i915 0000:00:02.0: [drm:__i915_inject_load_error 
[i915]] Injecting failure -5 at checkpoint 44 [i915_gem_init:1503]
<6> [237.840808] [drm] GuC communication disabled
...
<6> [238.160935] [drm] GuC communication disabled

Maybe return early from guc_disable_communication if the communication 
was already disabled?

Daniele

>   		return;
>   
>   	if (intel_uc_supports_guc_submission(uc))
> 
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

  parent reply	other threads:[~2019-08-13 20:16 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-08-13 16:26 [PATCH] drm/i915/uc: Fini hw even if GuC is not running Fernando Pacheco
2019-08-13 18:38 ` ✗ Fi.CI.BAT: failure for " Patchwork
2019-08-13 20:16 ` Daniele Ceraolo Spurio [this message]
2019-08-14 14:20   ` [PATCH] " Fernando Pacheco
2019-08-13 20:18 ` Michal Wajdeczko
2019-08-14 14:33   ` Fernando Pacheco

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=39d7f313-dec7-6432-af08-a9127d3032ce@intel.com \
    --to=daniele.ceraolospurio@intel.com \
    --cc=fernando.pacheco@intel.com \
    --cc=intel-gfx@lists.freedesktop.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox