All of lore.kernel.org
 help / color / mirror / Atom feed
From: Sebastian Brzezinka <sebastian.brzezinka@intel.com>
To: Kamil Konieczny <kamil.konieczny@linux.intel.com>,
	Sebastian Brzezinka <sebastian.brzezinka@intel.com>
Cc: igt-dev@lists.freedesktop.org, ashutosh.dixit@intel.com,
	"Krzysztof Karas" <krzysztof.karas@intel.com>,
	"Zbigniew Kempczyński" <zbigniew.kempczynski@intel.com>
Subject: Re: [PATCH] tools/igt_power: Fix crash on Xe devices by initializing xe_device cache
Date: Tue, 19 May 2026 13:47:01 +0200	[thread overview]
Message-ID: <DIMMNFLMXWP9.248NDUD1GYKFC@intel.com> (raw)
In-Reply-To: <20260519112825.7c7gpzdpevydhkjt@kamilkon-DESK.igk.intel.com>

Hi,

On Tue May 19, 2026 at 1:28 PM CEST, Kamil Konieczny wrote:
> Hi Sebastian,
> On 2026-05-18 at 14:32:24 +0200, Sebastian Brzezinka wrote:
>> Opening a DRM device with plain open() does not initialize the Xe device
>> cache. When igt_power_open() calls is_intel_dgfx() -> xe_has_vram(), it
>> uses find_in_cache() which asserts the cache entry exists, causing a crash
>> on Xe devices with: xe_query-CRITICAL: Failed assertion: xe_dev
>> 
>> Fix this by opening the device with O_RDWR (required for DRM ioctls),
>> calling xe_device_get() for Xe devices to populate the cache before
>> calling igt_power_open(), and calling xe_device_put() on cleanup for
>> proper cache management.
>> 
>> Signed-off-by: Sebastian Brzezinka <sebastian.brzezinka@intel.com>
>> ---
>>  tools/igt_power.c | 18 +++++++++++++++---
>>  1 file changed, 15 insertions(+), 3 deletions(-)
>> 
>> diff --git a/tools/igt_power.c b/tools/igt_power.c
>> index 6dd180d34..11da0430b 100644
>> --- a/tools/igt_power.c
>> +++ b/tools/igt_power.c
>> @@ -11,6 +11,7 @@
>>  
>>  #include "drmtest.h"
>>  #include "igt_power.h"
>> +#include "xe/xe_query.h"
>>  
>
> While you fix it now, your change make it even more difficult
> to remove 'testing' dependancy in tools, see work item
> https://gitlab.freedesktop.org/drm/igt-gpu-tools/-/work_items/16
> IGT command line tools load redundant GUI libraries
Yeah, I know. I can probably handle it in igt_power_open(), but I’m
not sure it helps much with the dependency issue overall.  Still, it
should be a bit cleaner, so I’ll go with v2.

-- 
Best regards,
Sebastian


  reply	other threads:[~2026-05-19 11:47 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-05-18 12:32 [PATCH] tools/igt_power: Fix crash on Xe devices by initializing xe_device cache Sebastian Brzezinka
2026-05-19  1:22 ` ✓ Xe.CI.BAT: success for " Patchwork
2026-05-19  1:43 ` ✓ i915.CI.BAT: " Patchwork
2026-05-19  9:19 ` ✗ Xe.CI.FULL: failure " Patchwork
2026-05-19 10:23 ` [PATCH] " Krzysztof Karas
2026-05-19 10:28   ` Sebastian Brzezinka
2026-05-19 10:52 ` Zbigniew Kempczyński
2026-05-19 11:28 ` Kamil Konieczny
2026-05-19 11:47   ` Sebastian Brzezinka [this message]
2026-05-19 20:55 ` ✗ i915.CI.Full: failure for " Patchwork

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=DIMMNFLMXWP9.248NDUD1GYKFC@intel.com \
    --to=sebastian.brzezinka@intel.com \
    --cc=ashutosh.dixit@intel.com \
    --cc=igt-dev@lists.freedesktop.org \
    --cc=kamil.konieczny@linux.intel.com \
    --cc=krzysztof.karas@intel.com \
    --cc=zbigniew.kempczynski@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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.