[Intel-gfx] [PATCH 11/15] drm/i915: Trim i915_guc_info() stack usage

[Tvrtko Ursulin]

On 28/11/2016 11:35, Chris Wilson wrote:
> On Mon, Nov 28, 2016 at 11:15:27AM +0000, Tvrtko Ursulin wrote:
>>
>> On 25/11/2016 09:30, Chris Wilson wrote:
>>> i915_guc_info() (part of debugfs output) tries to avoid holding
>>> struct_mutex for a long period by copying onto the stack. This causes a
>>> warning that the stack frame is massive, so stop doing that. We can even
>>> forgo holding the struct_mutex here as that doesn't serialise the values
>>> being read (and the lists used exist for the device lifetime).
>>>
>>> Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
>>> ---
>>> drivers/gpu/drm/i915/i915_debugfs.c | 39 +++++++++++++------------------------
>>> 1 file changed, 14 insertions(+), 25 deletions(-)
>>>
>>> diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
>>> index 8eb8c29b7492..7676e88ae5f2 100644
>>> --- a/drivers/gpu/drm/i915/i915_debugfs.c
>>> +++ b/drivers/gpu/drm/i915/i915_debugfs.c
>>> @@ -2434,47 +2434,36 @@ static void i915_guc_client_info(struct seq_file *m,
>>> static int i915_guc_info(struct seq_file *m, void *data)
>>> {
>>> 	struct drm_i915_private *dev_priv = node_to_i915(m->private);
>>> -	struct drm_device *dev = &dev_priv->drm;
>>> -	struct intel_guc guc;
>>> -	struct i915_guc_client client = {};
>>> +	const struct intel_guc *guc = &dev_priv->guc;
>>> 	struct intel_engine_cs *engine;
>>> 	enum intel_engine_id id;
>>> -	u64 total = 0;
>>> +	u64 total;
>>>
>>> 	if (!HAS_GUC_SCHED(dev_priv))
>>> 		return 0;
>>>
>>> -	if (mutex_lock_interruptible(&dev->struct_mutex))
>>> -		return 0;
>>> -
>>> -	/* Take a local copy of the GuC data, so we can dump it at leisure */
>>> -	guc = dev_priv->guc;
>>> -	if (guc.execbuf_client)
>>> -		client = *guc.execbuf_client;
>>
>> So this used to print out all zeros when GuC submission is disabled.
>> Should we instead just skip all the counter dumping if
>> execbuf_client == NULL and just print "disabled" or something?
>
> That would be the sensible escape instead of HAS_GUC_SCHED.
>
> if (!dev_priv->guc.execbuf_client) {
> 	seq_puts(m, "GuC scheduling %s\n",
> 		 HAS_GUC_SCHED(dev_priv) ?  "disabled" : "not supported");
> 	return 0;
> }

That looks good to me. (Just printf, not puts.)

Regards,

Tvrtko