[v2] drm/i915: Use the engine name directly in the error_state file

Submitted by Michel Thierry on Jan. 10, 2018, 1:21 a.m.

Details

Message ID 20180110012151.28261-1-michel.thierry@intel.com
State New
Headers show
Series "drm/i915: Use the engine name directly in the error_state file" ( rev: 2 ) in Intel GFX

Not browsing as part of any series.

Commit Message

Michel Thierry Jan. 10, 2018, 1:21 a.m.
Instead of using local string names that we will have to keep
maintaining, use the engine->name directly.

v2: Better invalid engine_id handling, capture_bo will not be able know
the engine_id and end up with -1 (Michal).

Suggested-by: Michal Wajdeczko <michal.wajdeczko@intel.com>
Signed-off-by: Michel Thierry <michel.thierry@intel.com>
Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
---
 drivers/gpu/drm/i915/i915_gpu_error.c | 33 ++++++++++++++++++++-------------
 1 file changed, 20 insertions(+), 13 deletions(-)

Patch hide | download patch | download mbox

diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c
index 94499c24f279..422e302161e5 100644
--- a/drivers/gpu/drm/i915/i915_gpu_error.c
+++ b/drivers/gpu/drm/i915/i915_gpu_error.c
@@ -34,16 +34,22 @@ 
 
 #include "i915_drv.h"
 
-static const char *engine_str(int engine)
-{
-	switch (engine) {
-	case RCS: return "render";
-	case VCS: return "bsd";
-	case BCS: return "blt";
-	case VECS: return "vebox";
-	case VCS2: return "bsd2";
-	default: return "";
-	}
+static inline const char *intel_engine_name(struct intel_engine_cs *engine)
+{
+	return engine ? engine->name : "";
+}
+
+static inline struct intel_engine_cs *
+intel_engine_lookup(struct drm_i915_private *i915, int engine_id)
+{
+	if (engine_id < 0 || engine_id >= I915_NUM_ENGINES)
+		return NULL;
+	return i915->engine[engine_id];
+}
+
+static const char *engine_str(struct drm_i915_private *i915, int engine_id)
+{
+	return intel_engine_name(intel_engine_lookup(i915, engine_id));
 }
 
 static const char *tiling_flag(int tiling)
@@ -345,7 +351,7 @@  static void print_error_buffers(struct drm_i915_error_state_buf *m,
 		err_puts(m, purgeable_flag(err->purgeable));
 		err_puts(m, err->userptr ? " userptr" : "");
 		err_puts(m, err->engine != -1 ? " " : "");
-		err_puts(m, engine_str(err->engine));
+		err_puts(m, engine_str(m->i915, err->engine));
 		err_puts(m, i915_cache_level_str(m->i915, err->cache_level));
 
 		if (err->name)
@@ -417,7 +423,8 @@  static void error_print_engine(struct drm_i915_error_state_buf *m,
 {
 	int n;
 
-	err_printf(m, "%s command stream:\n", engine_str(ee->engine_id));
+	err_printf(m, "%s command stream:\n", engine_str(m->i915,
+							 ee->engine_id));
 	err_printf(m, "  IDLE?: %s\n", yesno(ee->idle));
 	err_printf(m, "  START: 0x%08x\n", ee->start);
 	err_printf(m, "  HEAD:  0x%08x [0x%08x]\n", ee->head, ee->rq_head);
@@ -633,7 +640,7 @@  int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
 		if (error->engine[i].hangcheck_stalled &&
 		    error->engine[i].context.pid) {
 			err_printf(m, "Active process (on ring %s): %s [%d], score %d\n",
-				   engine_str(i),
+				   engine_str(m->i915, i),
 				   error->engine[i].context.comm,
 				   error->engine[i].context.pid,
 				   error->engine[i].context.ban_score);

Comments

On 1/9/2018 5:21 PM, Michel Thierry wrote:
> Instead of using local string names that we will have to keep
> maintaining, use the engine->name directly.
> 
> v2: Better invalid engine_id handling, capture_bo will not be able know
> the engine_id and end up with -1 (Michal).
> 
Hi,

Fi.CI.IGT didn't catch any failure, apart from intel_error_decode (which 
didn't complain), I'm not sure which other tool could be using these names.

And now that there are icl patches, this will help there too.


> Suggested-by: Michal Wajdeczko <michal.wajdeczko@intel.com>
> Signed-off-by: Michel Thierry <michel.thierry@intel.com>
> Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
> Cc: Chris Wilson <chris@chris-wilson.co.uk>
> ---
>   drivers/gpu/drm/i915/i915_gpu_error.c | 33 ++++++++++++++++++++-------------
>   1 file changed, 20 insertions(+), 13 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c
> index 94499c24f279..422e302161e5 100644
> --- a/drivers/gpu/drm/i915/i915_gpu_error.c
> +++ b/drivers/gpu/drm/i915/i915_gpu_error.c
> @@ -34,16 +34,22 @@
> 
>   #include "i915_drv.h"
> 
> -static const char *engine_str(int engine)
> -{
> -       switch (engine) {
> -       case RCS: return "render";
> -       case VCS: return "bsd";
> -       case BCS: return "blt";
> -       case VECS: return "vebox";
> -       case VCS2: return "bsd2";
> -       default: return "";
> -       }
> +static inline const char *intel_engine_name(struct intel_engine_cs *engine)
> +{
> +       return engine ? engine->name : "";
> +}
> +
> +static inline struct intel_engine_cs *
> +intel_engine_lookup(struct drm_i915_private *i915, int engine_id)
> +{
> +       if (engine_id < 0 || engine_id >= I915_NUM_ENGINES)
> +               return NULL;
> +       return i915->engine[engine_id];
> +}
> +
> +static const char *engine_str(struct drm_i915_private *i915, int engine_id)
> +{
> +       return intel_engine_name(intel_engine_lookup(i915, engine_id));
>   }
> 
>   static const char *tiling_flag(int tiling)
> @@ -345,7 +351,7 @@ static void print_error_buffers(struct drm_i915_error_state_buf *m,
>                  err_puts(m, purgeable_flag(err->purgeable));
>                  err_puts(m, err->userptr ? " userptr" : "");
>                  err_puts(m, err->engine != -1 ? " " : "");
> -               err_puts(m, engine_str(err->engine));
> +               err_puts(m, engine_str(m->i915, err->engine));
>                  err_puts(m, i915_cache_level_str(m->i915, err->cache_level));
> 
>                  if (err->name)
> @@ -417,7 +423,8 @@ static void error_print_engine(struct drm_i915_error_state_buf *m,
>   {
>          int n;
> 
> -       err_printf(m, "%s command stream:\n", engine_str(ee->engine_id));
> +       err_printf(m, "%s command stream:\n", engine_str(m->i915,
> +                                                        ee->engine_id));
>          err_printf(m, "  IDLE?: %s\n", yesno(ee->idle));
>          err_printf(m, "  START: 0x%08x\n", ee->start);
>          err_printf(m, "  HEAD:  0x%08x [0x%08x]\n", ee->head, ee->rq_head);
> @@ -633,7 +640,7 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
>                  if (error->engine[i].hangcheck_stalled &&
>                      error->engine[i].context.pid) {
>                          err_printf(m, "Active process (on ring %s): %s [%d], score %d\n",
> -                                  engine_str(i),
> +                                  engine_str(m->i915, i),
>                                     error->engine[i].context.comm,
>                                     error->engine[i].context.pid,
>                                     error->engine[i].context.ban_score);
> --
> 2.15.1
> 
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
>
On 10/01/2018 01:21, Michel Thierry wrote:
> Instead of using local string names that we will have to keep
> maintaining, use the engine->name directly.
> 
> v2: Better invalid engine_id handling, capture_bo will not be able know
> the engine_id and end up with -1 (Michal).
> 
> Suggested-by: Michal Wajdeczko <michal.wajdeczko@intel.com>
> Signed-off-by: Michel Thierry <michel.thierry@intel.com>
> Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
> Cc: Chris Wilson <chris@chris-wilson.co.uk>
> ---
>   drivers/gpu/drm/i915/i915_gpu_error.c | 33 ++++++++++++++++++++-------------
>   1 file changed, 20 insertions(+), 13 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c
> index 94499c24f279..422e302161e5 100644
> --- a/drivers/gpu/drm/i915/i915_gpu_error.c
> +++ b/drivers/gpu/drm/i915/i915_gpu_error.c
> @@ -34,16 +34,22 @@
>   
>   #include "i915_drv.h"
>   
> -static const char *engine_str(int engine)
> -{
> -	switch (engine) {
> -	case RCS: return "render";
> -	case VCS: return "bsd";
> -	case BCS: return "blt";
> -	case VECS: return "vebox";
> -	case VCS2: return "bsd2";
> -	default: return "";
> -	}
> +static inline const char *intel_engine_name(struct intel_engine_cs *engine)
> +{
> +	return engine ? engine->name : "";
> +}
> +
> +static inline struct intel_engine_cs *
> +intel_engine_lookup(struct drm_i915_private *i915, int engine_id)
> +{
> +	if (engine_id < 0 || engine_id >= I915_NUM_ENGINES)
> +		return NULL;
> +	return i915->engine[engine_id];
> +}
> +
> +static const char *engine_str(struct drm_i915_private *i915, int engine_id)
> +{
> +	return intel_engine_name(intel_engine_lookup(i915, engine_id));
>   }

Feels like a bit of an overkill to have three functions to this trivial 
thing but meh. Could also maybe cheat and have engine_id as unsigned int 
and so would only need to check for >= I915_NUM_ENGINES.

Anyway, I peeked in intel_error_decode source and couldn't spot anything 
that looked it would break. You checked it by running it? Assuming you did:

Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>

Regards,

Tvrtko

>   
>   static const char *tiling_flag(int tiling)
> @@ -345,7 +351,7 @@ static void print_error_buffers(struct drm_i915_error_state_buf *m,
>   		err_puts(m, purgeable_flag(err->purgeable));
>   		err_puts(m, err->userptr ? " userptr" : "");
>   		err_puts(m, err->engine != -1 ? " " : "");
> -		err_puts(m, engine_str(err->engine));
> +		err_puts(m, engine_str(m->i915, err->engine));
>   		err_puts(m, i915_cache_level_str(m->i915, err->cache_level));
>   
>   		if (err->name)
> @@ -417,7 +423,8 @@ static void error_print_engine(struct drm_i915_error_state_buf *m,
>   {
>   	int n;
>   
> -	err_printf(m, "%s command stream:\n", engine_str(ee->engine_id));
> +	err_printf(m, "%s command stream:\n", engine_str(m->i915,
> +							 ee->engine_id));
>   	err_printf(m, "  IDLE?: %s\n", yesno(ee->idle));
>   	err_printf(m, "  START: 0x%08x\n", ee->start);
>   	err_printf(m, "  HEAD:  0x%08x [0x%08x]\n", ee->head, ee->rq_head);
> @@ -633,7 +640,7 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
>   		if (error->engine[i].hangcheck_stalled &&
>   		    error->engine[i].context.pid) {
>   			err_printf(m, "Active process (on ring %s): %s [%d], score %d\n",
> -				   engine_str(i),
> +				   engine_str(m->i915, i),
>   				   error->engine[i].context.comm,
>   				   error->engine[i].context.pid,
>   				   error->engine[i].context.ban_score);
>
On 1/15/2018 9:15 AM, Tvrtko Ursulin wrote:
> 
> On 10/01/2018 01:21, Michel Thierry wrote:
>> Instead of using local string names that we will have to keep
>> maintaining, use the engine->name directly.
>>
>> v2: Better invalid engine_id handling, capture_bo will not be able know
>> the engine_id and end up with -1 (Michal).
>>
>> Suggested-by: Michal Wajdeczko <michal.wajdeczko@intel.com>
>> Signed-off-by: Michel Thierry <michel.thierry@intel.com>
>> Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
>> Cc: Chris Wilson <chris@chris-wilson.co.uk>
>> ---
>>   drivers/gpu/drm/i915/i915_gpu_error.c | 33 
>> ++++++++++++++++++++-------------
>>   1 file changed, 20 insertions(+), 13 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c 
>> b/drivers/gpu/drm/i915/i915_gpu_error.c
>> index 94499c24f279..422e302161e5 100644
>> --- a/drivers/gpu/drm/i915/i915_gpu_error.c
>> +++ b/drivers/gpu/drm/i915/i915_gpu_error.c
>> @@ -34,16 +34,22 @@
>>   #include "i915_drv.h"
>> -static const char *engine_str(int engine)
>> -{
>> -    switch (engine) {
>> -    case RCS: return "render";
>> -    case VCS: return "bsd";
>> -    case BCS: return "blt";
>> -    case VECS: return "vebox";
>> -    case VCS2: return "bsd2";
>> -    default: return "";
>> -    }
>> +static inline const char *intel_engine_name(struct intel_engine_cs 
>> *engine)
>> +{
>> +    return engine ? engine->name : "";
>> +}
>> +
>> +static inline struct intel_engine_cs *
>> +intel_engine_lookup(struct drm_i915_private *i915, int engine_id)
>> +{
>> +    if (engine_id < 0 || engine_id >= I915_NUM_ENGINES)
>> +        return NULL;
>> +    return i915->engine[engine_id];
>> +}
>> +
>> +static const char *engine_str(struct drm_i915_private *i915, int 
>> engine_id)
>> +{
>> +    return intel_engine_name(intel_engine_lookup(i915, engine_id));
>>   }
> 
> Feels like a bit of an overkill to have three functions to this trivial 
> thing but meh. Could also maybe cheat and have engine_id as unsigned int 
> and so would only need to check for >= I915_NUM_ENGINES.
> 
> Anyway, I peeked in intel_error_decode source and couldn't spot anything 
> that looked it would break. You checked it by running it? Assuming you did:
> 

Thanks. Yes, I did and also compared the error state output 
(intel_error_decode also still works). The only change is that the batch 
/ ring / HWSP sections now use the engine->name (e.g. they will use rcs0 
instead of render, vcs0 instead of bsd, etc.).

I didn't see anything in IGT complaining (probably because IGT usually 
reads dmesg and there we already use engine->name), but there must be 
something else out there that will complain.


> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
> 
> Regards,
> 
> Tvrtko
> 
>>   static const char *tiling_flag(int tiling)
>> @@ -345,7 +351,7 @@ static void print_error_buffers(struct 
>> drm_i915_error_state_buf *m,
>>           err_puts(m, purgeable_flag(err->purgeable));
>>           err_puts(m, err->userptr ? " userptr" : "");
>>           err_puts(m, err->engine != -1 ? " " : "");
>> -        err_puts(m, engine_str(err->engine));
>> +        err_puts(m, engine_str(m->i915, err->engine));
>>           err_puts(m, i915_cache_level_str(m->i915, err->cache_level));
>>           if (err->name)
>> @@ -417,7 +423,8 @@ static void error_print_engine(struct 
>> drm_i915_error_state_buf *m,
>>   {
>>       int n;
>> -    err_printf(m, "%s command stream:\n", engine_str(ee->engine_id));
>> +    err_printf(m, "%s command stream:\n", engine_str(m->i915,
>> +                             ee->engine_id));
>>       err_printf(m, "  IDLE?: %s\n", yesno(ee->idle));
>>       err_printf(m, "  START: 0x%08x\n", ee->start);
>>       err_printf(m, "  HEAD:  0x%08x [0x%08x]\n", ee->head, ee->rq_head);
>> @@ -633,7 +640,7 @@ int i915_error_state_to_str(struct 
>> drm_i915_error_state_buf *m,
>>           if (error->engine[i].hangcheck_stalled &&
>>               error->engine[i].context.pid) {
>>               err_printf(m, "Active process (on ring %s): %s [%d], 
>> score %d\n",
>> -                   engine_str(i),
>> +                   engine_str(m->i915, i),
>>                      error->engine[i].context.comm,
>>                      error->engine[i].context.pid,
>>                      error->engine[i].context.ban_score);
>>
Quoting Michel Thierry (2018-01-16 18:33:32)
> 
> 
> On 1/15/2018 9:15 AM, Tvrtko Ursulin wrote:
> > 
> > On 10/01/2018 01:21, Michel Thierry wrote:
> >> Instead of using local string names that we will have to keep
> >> maintaining, use the engine->name directly.
> >>
> >> v2: Better invalid engine_id handling, capture_bo will not be able know
> >> the engine_id and end up with -1 (Michal).
> >>
> >> Suggested-by: Michal Wajdeczko <michal.wajdeczko@intel.com>
> >> Signed-off-by: Michel Thierry <michel.thierry@intel.com>
> >> Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
> >> Cc: Chris Wilson <chris@chris-wilson.co.uk>
> >> ---
> >>   drivers/gpu/drm/i915/i915_gpu_error.c | 33 
> >> ++++++++++++++++++++-------------
> >>   1 file changed, 20 insertions(+), 13 deletions(-)
> >>
> >> diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c 
> >> b/drivers/gpu/drm/i915/i915_gpu_error.c
> >> index 94499c24f279..422e302161e5 100644
> >> --- a/drivers/gpu/drm/i915/i915_gpu_error.c
> >> +++ b/drivers/gpu/drm/i915/i915_gpu_error.c
> >> @@ -34,16 +34,22 @@
> >>   #include "i915_drv.h"
> >> -static const char *engine_str(int engine)
> >> -{
> >> -    switch (engine) {
> >> -    case RCS: return "render";
> >> -    case VCS: return "bsd";
> >> -    case BCS: return "blt";
> >> -    case VECS: return "vebox";
> >> -    case VCS2: return "bsd2";
> >> -    default: return "";
> >> -    }
> >> +static inline const char *intel_engine_name(struct intel_engine_cs 
> >> *engine)
> >> +{
> >> +    return engine ? engine->name : "";
> >> +}
> >> +
> >> +static inline struct intel_engine_cs *
> >> +intel_engine_lookup(struct drm_i915_private *i915, int engine_id)
> >> +{
> >> +    if (engine_id < 0 || engine_id >= I915_NUM_ENGINES)
> >> +        return NULL;
> >> +    return i915->engine[engine_id];
> >> +}
> >> +
> >> +static const char *engine_str(struct drm_i915_private *i915, int 
> >> engine_id)
> >> +{
> >> +    return intel_engine_name(intel_engine_lookup(i915, engine_id));
> >>   }
> > 
> > Feels like a bit of an overkill to have three functions to this trivial 
> > thing but meh. Could also maybe cheat and have engine_id as unsigned int 
> > and so would only need to check for >= I915_NUM_ENGINES.

Overkill? Yes! If you are concerned about the back mapping, just put a
pointer back to the engine (ee->engine). We are not going to reallocate
it after module load, right? RIGHT?!!! :)

> > Anyway, I peeked in intel_error_decode source and couldn't spot anything 
> > that looked it would break. You checked it by running it? Assuming you did:
> > 
> 
> Thanks. Yes, I did and also compared the error state output 
> (intel_error_decode also still works). The only change is that the batch 
> / ring / HWSP sections now use the engine->name (e.g. they will use rcs0 
> instead of render, vcs0 instead of bsd, etc.).
> 
> I didn't see anything in IGT complaining (probably because IGT usually 
> reads dmesg and there we already use engine->name), but there must be 
> something else out there that will complain.

Hmm, we will need to fixup mesa/aubinator_error_decode to expect the new
name, as it is using them to map various registers from the error state.
-Chris
On 17/01/18 07:15, Chris Wilson wrote:
> Quoting Michel Thierry (2018-01-16 18:33:32)
>>
>>
>> On 1/15/2018 9:15 AM, Tvrtko Ursulin wrote:
>>>
>>> On 10/01/2018 01:21, Michel Thierry wrote:
>>>> Instead of using local string names that we will have to keep
>>>> maintaining, use the engine->name directly.
>>>>
>>>> v2: Better invalid engine_id handling, capture_bo will not be able know
>>>> the engine_id and end up with -1 (Michal).
>>>>
>>>> Suggested-by: Michal Wajdeczko <michal.wajdeczko@intel.com>
>>>> Signed-off-by: Michel Thierry <michel.thierry@intel.com>
>>>> Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
>>>> Cc: Chris Wilson <chris@chris-wilson.co.uk>
>>>> ---
>>>>    drivers/gpu/drm/i915/i915_gpu_error.c | 33
>>>> ++++++++++++++++++++-------------
>>>>    1 file changed, 20 insertions(+), 13 deletions(-)
>>>>
>>>> diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c
>>>> b/drivers/gpu/drm/i915/i915_gpu_error.c
>>>> index 94499c24f279..422e302161e5 100644
>>>> --- a/drivers/gpu/drm/i915/i915_gpu_error.c
>>>> +++ b/drivers/gpu/drm/i915/i915_gpu_error.c
>>>> @@ -34,16 +34,22 @@
>>>>    #include "i915_drv.h"
>>>> -static const char *engine_str(int engine)
>>>> -{
>>>> -    switch (engine) {
>>>> -    case RCS: return "render";
>>>> -    case VCS: return "bsd";
>>>> -    case BCS: return "blt";
>>>> -    case VECS: return "vebox";
>>>> -    case VCS2: return "bsd2";
>>>> -    default: return "";
>>>> -    }
>>>> +static inline const char *intel_engine_name(struct intel_engine_cs
>>>> *engine)
>>>> +{
>>>> +    return engine ? engine->name : "";
>>>> +}
>>>> +
>>>> +static inline struct intel_engine_cs *
>>>> +intel_engine_lookup(struct drm_i915_private *i915, int engine_id)
>>>> +{
>>>> +    if (engine_id < 0 || engine_id >= I915_NUM_ENGINES)
>>>> +        return NULL;
>>>> +    return i915->engine[engine_id];
>>>> +}
>>>> +
>>>> +static const char *engine_str(struct drm_i915_private *i915, int
>>>> engine_id)
>>>> +{
>>>> +    return intel_engine_name(intel_engine_lookup(i915, engine_id));
>>>>    }
>>>
>>> Feels like a bit of an overkill to have three functions to this trivial
>>> thing but meh. Could also maybe cheat and have engine_id as unsigned int
>>> and so would only need to check for >= I915_NUM_ENGINES.
> 
> Overkill? Yes! If you are concerned about the back mapping, just put a
> pointer back to the engine (ee->engine). We are not going to reallocate
> it after module load, right? RIGHT?!!! :)
> 
>>> Anyway, I peeked in intel_error_decode source and couldn't spot anything
>>> that looked it would break. You checked it by running it? Assuming you did:
>>>
>>
>> Thanks. Yes, I did and also compared the error state output
>> (intel_error_decode also still works). The only change is that the batch
>> / ring / HWSP sections now use the engine->name (e.g. they will use rcs0
>> instead of render, vcs0 instead of bsd, etc.).
>>
>> I didn't see anything in IGT complaining (probably because IGT usually
>> reads dmesg and there we already use engine->name), but there must be
>> something else out there that will complain.
> 
> Hmm, we will need to fixup mesa/aubinator_error_decode to expect the new
> name, as it is using them to map various registers from the error state.

Thanks for taking care of aubinator, hopefully next time I won't forget 
about mesa.
Quoting Michel Thierry (2018-01-18 00:39:55)
> On 17/01/18 07:15, Chris Wilson wrote:
> > Hmm, we will need to fixup mesa/aubinator_error_decode to expect the new
> > name, as it is using them to map various registers from the error state.
> 
> Thanks for taking care of aubinator, hopefully next time I won't forget 
> about mesa.

Would be good if you could review my string handling there to make sure
it does match. :)
-Chris
Quoting Michel Thierry (2018-01-10 01:21:51)
> Instead of using local string names that we will have to keep
> maintaining, use the engine->name directly.
> 
> v2: Better invalid engine_id handling, capture_bo will not be able know
> the engine_id and end up with -1 (Michal).
> 
> Suggested-by: Michal Wajdeczko <michal.wajdeczko@intel.com>
> Signed-off-by: Michel Thierry <michel.thierry@intel.com>
> Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
> Cc: Chris Wilson <chris@chris-wilson.co.uk>

Massaged the patch just ever so slightly and applied. Thanks for the
patch and review,
-Chris