Intel-GFX Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: "Teres Alexis, Alan Previn" <alan.previn.teres.alexis@intel.com>
To: "Harrison, John C" <john.c.harrison@intel.com>,
	"intel-gfx@lists.freedesktop.org"
	<intel-gfx@lists.freedesktop.org>
Subject: Re: [Intel-gfx] [PATCH v2 1/1] drm/i915/guc: Fix GuC error capture sizing estimation and reporting
Date: Mon, 3 Oct 2022 18:36:20 +0000	[thread overview]
Message-ID: <6db74d66964d1db51649fe92003a9fe0dab52fae.camel@intel.com> (raw)
In-Reply-To: <20220930211836.148041-2-alan.previn.teres.alexis@intel.com>

Hi John - how would you like to proceed? I have re-rev'd as per your original review comment on rev1.

Shall we adopt this rev2's "drm_warn" for the worst-case (knowing well that gpu_core_dump is still an external subsystem
that can cull our data, but at least within this subsystem we are adding this warning for worst case and merely a debug
when we want the 3x capture). As we know GuC operation is unaware of the gpu-core-dump restriction so its not hard to
imagine the GuC capturing the amount of data for the worst case scenario if have big problems in the workloads or hw.

Additionally, would you prefer to completely drop the spare size? Some context: with the calculation fix we are
allocating 4MB but we only need 78k as min-est.

...alan

On Fri, 2022-09-30 at 14:18 -0700, Alan Previn wrote:
> +	/*
> +	 * Don't drm_warn or drm_error here on "possible" insufficient size because we only run out
> +	 * of space if all engines were to suffer resets all at once AND the driver is not able to
> +	 * extract that data fast enough in the interrupt handler worker (moving them to the
> +	 * larger pool of pre-allocated capture nodes. If GuC does run out of space, we will
> +	 * print an error when processing the G2H event capture-notification (search for
> +	 * "INTEL_GUC_STATE_CAPTURE_EVENT_STATUS_NOSPACE").
> +	 */
>  	if (min_size < 0)
>  		drm_warn(&i915->drm, "Failed to calculate GuC error state capture buffer minimum size: %d!\n",
>  			 min_size);
>  	else if (min_size > buffer_size)
> -		drm_warn(&i915->drm, "GuC error state capture buffer is too small: %d < %d\n",
> +		drm_warn(&i915->drm, "GuC error state capture buffer maybe small: %d < %d\n",
>  			 buffer_size, min_size);
>  	else if (spare_size > buffer_size)
> -		drm_notice(&i915->drm, "GuC error state capture buffer maybe too small: %d < %d (min = %d)\n",
> -			   buffer_size, spare_size, min_size);
> +		drm_dbg(&i915->drm, "GuC error state capture buffer lacks spare size: %d < %d (min = %d)\n",
> +			buffer_size, spare_size, min_size);
>  }
>  
>  /*
> -- 
> 2.34.1
> 


  reply	other threads:[~2022-10-03 18:36 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-09-30 21:18 [Intel-gfx] [PATCH v2 0/1] Fix Guc-Err-Capture sizing warning Alan Previn
2022-09-30 21:18 ` [Intel-gfx] [PATCH v2 1/1] drm/i915/guc: Fix GuC error capture sizing estimation and reporting Alan Previn
2022-10-03 18:36   ` Teres Alexis, Alan Previn [this message]
2022-09-30 21:54 ` [Intel-gfx] ✓ Fi.CI.BAT: success for Fix Guc-Err-Capture sizing warning (rev2) Patchwork
2022-10-01 20:05 ` [Intel-gfx] ✗ Fi.CI.IGT: failure " Patchwork
2022-10-03 20:23   ` Teres Alexis, Alan Previn

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=6db74d66964d1db51649fe92003a9fe0dab52fae.camel@intel.com \
    --to=alan.previn.teres.alexis@intel.com \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=john.c.harrison@intel.com \
    /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