Intel-XE Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Michal Wajdeczko <michal.wajdeczko@intel.com>
To: John Harrison <john.c.harrison@intel.com>,
	intel-xe@lists.freedesktop.org
Cc: Lucas De Marchi <lucas.demarchi@intel.com>,
	Rodrigo Vivi <rodrigo.vivi@intel.com>
Subject: Re: [PATCH] drm/xe: Allow to trigger GT resets using debugfs writes
Date: Tue, 20 May 2025 11:12:04 +0200	[thread overview]
Message-ID: <e222a156-bc07-45d0-bfd6-f24ff31696a4@intel.com> (raw)
In-Reply-To: <eefda6f4-e81c-43a6-bebc-7d4fe0c3132a@intel.com>



On 20.05.2025 01:44, John Harrison wrote:
> On 5/19/2025 1:09 PM, Michal Wajdeczko wrote:
>> Today we allow to trigger GT resest by reading dedicated debugfs
>> files "force_reset" and "force_reset_sync" that we are exposing
>> using drm_info_list[] and drm_debugfs_create_files().
>>
>> To avoid triggering potentially disruptive actions during otherwise
>> "safe" read operations, expose those two attributes using debugfs
>> function where we can specify file permissions and provide custom
>> "write" handler to trigger the GT resets also from there.
> Would you like to extend this to the GuC log dump-via-dmesg trigger
> entry as well? Dumping that during the IGT read debugfs test also causes
> issues due to the huge amount of output it produces.

yes, that was my plan as I was expecting more discussion there so I
wanted to put that into a separate series

but since you already asked here, my question would be:

do you prefer keeping separate write-only "guc_log_dmesg" file, or keep
single read-write "guc_log" file that would print encoded log on read
and dump it to dmesg on write?

also do we want to keep 0444 mode for "guc_log" or change it to 0400
mode (or 0600 if it would be dual-purpose)

> 
>>
>> This step would allow us to drop triggering GT resets during read
>> operations, which we leave just to give users more time to switch.
> I'm not following this sentence. Give users more time to switch what?

switch to use "echo 1 > force_reset" instead of "cat force_reset"

> 
> Also, would it be more accurate to call the patch set something like
> 'require a write to trigger GT resets via debugfs'. Saying 'allow'
> implies you are just extending the existing interface rather than
> removing the trigger-via-read ability.

this patch extends the existing interface, the legacy way to trigger a
reset by read is still there (due to IGT) - see force_reset_show()

once the IGT will be updated, we would make those files write-only
I didn't want to introduce any CI regression right now, hence two steps

> 
> John.
> 
>>
>> Signed-off-by: Michal Wajdeczko <michal.wajdeczko@intel.com>
>> Cc: Lucas De Marchi <lucas.demarchi@intel.com>
>> Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
>> ---
>>   drivers/gpu/drm/xe/xe_gt_debugfs.c | 96 +++++++++++++++++++++++-------
>>   1 file changed, 76 insertions(+), 20 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/xe/xe_gt_debugfs.c b/drivers/gpu/drm/xe/
>> xe_gt_debugfs.c
>> index 119a55bb7580..848618acdca8 100644
>> --- a/drivers/gpu/drm/xe/xe_gt_debugfs.c
>> +++ b/drivers/gpu/drm/xe/xe_gt_debugfs.c
>> @@ -122,24 +122,6 @@ static int powergate_info(struct xe_gt *gt,
>> struct drm_printer *p)
>>       return ret;
>>   }
>>   -static int force_reset(struct xe_gt *gt, struct drm_printer *p)
>> -{
>> -    xe_pm_runtime_get(gt_to_xe(gt));
>> -    xe_gt_reset_async(gt);
>> -    xe_pm_runtime_put(gt_to_xe(gt));
>> -
>> -    return 0;
>> -}
>> -
>> -static int force_reset_sync(struct xe_gt *gt, struct drm_printer *p)
>> -{
>> -    xe_pm_runtime_get(gt_to_xe(gt));
>> -    xe_gt_reset(gt);
>> -    xe_pm_runtime_put(gt_to_xe(gt));
>> -
>> -    return 0;
>> -}
>> -
>>   static int sa_info(struct xe_gt *gt, struct drm_printer *p)
>>   {
>>       struct xe_tile *tile = gt_to_tile(gt);
>> @@ -306,8 +288,6 @@ static int hwconfig(struct xe_gt *gt, struct
>> drm_printer *p)
>>    * - without access to the PF specific data
>>    */
>>   static const struct drm_info_list vf_safe_debugfs_list[] = {
>> -    {"force_reset", .show = xe_gt_debugfs_simple_show, .data =
>> force_reset},
>> -    {"force_reset_sync", .show = xe_gt_debugfs_simple_show, .data =
>> force_reset_sync},
>>       {"sa_info", .show = xe_gt_debugfs_simple_show, .data = sa_info},
>>       {"topology", .show = xe_gt_debugfs_simple_show, .data = topology},
>>       {"ggtt", .show = xe_gt_debugfs_simple_show, .data = ggtt},
>> @@ -332,6 +312,78 @@ static const struct drm_info_list
>> pf_only_debugfs_list[] = {
>>       {"steering", .show = xe_gt_debugfs_simple_show, .data = steering},
>>   };
>>   +static ssize_t write_to_gt_call(const char __user *userbuf, size_t
>> count, loff_t *ppos,
>> +                void (*call)(struct xe_gt *), struct xe_gt *gt)
>> +{
>> +    bool yes;
>> +    int ret;
>> +
>> +    if (*ppos)
>> +        return -EINVAL;
>> +    ret = kstrtobool_from_user(userbuf, count, &yes);
>> +    if (ret < 0)
>> +        return ret;
>> +    if (yes)
>> +        call(gt);
>> +    return count;
>> +}
>> +
>> +static void force_reset(struct xe_gt *gt)
>> +{
>> +    struct xe_device *xe = gt_to_xe(gt);
>> +
>> +    xe_pm_runtime_get(xe);
>> +    xe_gt_reset_async(gt);
>> +    xe_pm_runtime_put(xe);
>> +}
>> +
>> +static ssize_t force_reset_write(struct file *file,
>> +                 const char __user *userbuf,
>> +                 size_t count, loff_t *ppos)
>> +{
>> +    struct seq_file *s = file->private_data;
>> +    struct xe_gt *gt = s->private;
>> +
>> +    return write_to_gt_call(userbuf, count, ppos, force_reset, gt);
>> +}
>> +
>> +static int force_reset_show(struct seq_file *s, void *unused)
>> +{
>> +    struct xe_gt *gt = s->private;
>> +
>> +    force_reset(gt); /* to be deprecated! */
>> +    return 0;
>> +}
>> +DEFINE_SHOW_STORE_ATTRIBUTE(force_reset);
>> +
>> +static void force_reset_sync(struct xe_gt *gt)
>> +{
>> +    struct xe_device *xe = gt_to_xe(gt);
>> +
>> +    xe_pm_runtime_get(xe);
>> +    xe_gt_reset(gt);
>> +    xe_pm_runtime_put(xe);
>> +}
>> +
>> +static ssize_t force_reset_sync_write(struct file *file,
>> +                      const char __user *userbuf,
>> +                      size_t count, loff_t *ppos)
>> +{
>> +    struct seq_file *s = file->private_data;
>> +    struct xe_gt *gt = s->private;
>> +
>> +    return write_to_gt_call(userbuf, count, ppos, force_reset_sync, gt);
>> +}
>> +
>> +static int force_reset_sync_show(struct seq_file *s, void *unused)
>> +{
>> +    struct xe_gt *gt = s->private;
>> +
>> +    force_reset_sync(gt); /* to be deprecated! */
>> +    return 0;
>> +}
>> +DEFINE_SHOW_STORE_ATTRIBUTE(force_reset_sync);
>> +
>>   void xe_gt_debugfs_register(struct xe_gt *gt)
>>   {
>>       struct xe_device *xe = gt_to_xe(gt);
>> @@ -355,6 +407,10 @@ void xe_gt_debugfs_register(struct xe_gt *gt)
>>        */
>>       root->d_inode->i_private = gt;
>>   +    /* VF safe */
>> +    debugfs_create_file("force_reset", 0600, root, gt,
>> &force_reset_fops);
>> +    debugfs_create_file("force_reset_sync", 0600, root, gt,
>> &force_reset_sync_fops);
>> +
>>       drm_debugfs_create_files(vf_safe_debugfs_list,
>>                    ARRAY_SIZE(vf_safe_debugfs_list),
>>                    root, minor);
> 


  reply	other threads:[~2025-05-20  9:12 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-05-19 20:09 [PATCH] drm/xe: Allow to trigger GT resets using debugfs writes Michal Wajdeczko
2025-05-19 20:15 ` ✓ CI.Patch_applied: success for " Patchwork
2025-05-19 20:15 ` ✓ CI.checkpatch: " Patchwork
2025-05-19 20:16 ` ✓ CI.KUnit: " Patchwork
2025-05-19 20:29 ` ✓ CI.Build: " Patchwork
2025-05-19 20:32 ` ✓ CI.Hooks: " Patchwork
2025-05-19 20:34 ` ✓ CI.checksparse: " Patchwork
2025-05-19 20:59 ` ✓ Xe.CI.BAT: " Patchwork
2025-05-19 21:59 ` [PATCH] " Rodrigo Vivi
2025-05-19 23:44 ` John Harrison
2025-05-20  9:12   ` Michal Wajdeczko [this message]
2025-05-21 21:05     ` John Harrison
2025-05-20  5:03 ` ✗ Xe.CI.Full: failure for " Patchwork
2025-05-28 12:28   ` Michal Wajdeczko
2025-05-27  5:10 ` ✓ CI.Patch_applied: success " Patchwork
2025-05-27  5:10 ` ✓ CI.checkpatch: " Patchwork
2025-05-27  5:12 ` ✓ CI.KUnit: " Patchwork
2025-05-27  5:22 ` ✓ CI.Build: " 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=e222a156-bc07-45d0-bfd6-f24ff31696a4@intel.com \
    --to=michal.wajdeczko@intel.com \
    --cc=intel-xe@lists.freedesktop.org \
    --cc=john.c.harrison@intel.com \
    --cc=lucas.demarchi@intel.com \
    --cc=rodrigo.vivi@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