| Message ID | 20180227125230.13000-9-michal.winiarski@intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Tue, 27 Feb 2018 13:52:24 +0100, Michał Winiarski <michal.winiarski@intel.com> wrote: > We only need those fast memcpy_wc when we're using relay to read > continuous GuC log. Let's prevent the user from creating a relay if we > know we won't be able to keep up with GuC. > > Signed-off-by: Michał Winiarski <michal.winiarski@intel.com> > Cc: Chris Wilson <chris@chris-wilson.co.uk> > Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> > Cc: Sagar Arun Kamble <sagar.a.kamble@intel.com> > Cc: Michal Wajdeczko <michal.wajdeczko@intel.com> > --- > drivers/gpu/drm/i915/intel_guc_log.c | 20 ++++++++++---------- > 1 file changed, 10 insertions(+), 10 deletions(-) > > diff --git a/drivers/gpu/drm/i915/intel_guc_log.c > b/drivers/gpu/drm/i915/intel_guc_log.c > index 4dee65692f5f..d2aca10ab986 100644 > --- a/drivers/gpu/drm/i915/intel_guc_log.c > +++ b/drivers/gpu/drm/i915/intel_guc_log.c > @@ -452,16 +452,6 @@ int intel_guc_log_create(struct intel_guc *guc) > GEM_BUG_ON(guc->log.vma); > - /* > - * We require SSE 4.1 for fast reads from the GuC log buffer and > - * it should be present on the chipsets supporting GuC based > - * submisssions. > - */ > - if (WARN_ON(!i915_has_memcpy_from_wc())) { > - ret = -EINVAL; > - goto err; > - } > - > vma = intel_guc_allocate_vma(guc, GUC_LOG_SIZE); > if (IS_ERR(vma)) { > ret = PTR_ERR(vma); > @@ -568,6 +558,16 @@ int intel_guc_log_relay_open(struct intel_guc *guc) > goto out_unlock; > } > + /* > + * We require SSE 4.1 for fast reads from the GuC log buffer and > + * it should be present on the chipsets supporting GuC based > + * submisssions. > + */ > + if (!i915_has_memcpy_from_wc()) { > + ret = -EINVAL; hmm, maybe -ENODEV as this is not user fault > + goto out_unlock; > + } > + > ret = guc_log_relay_create(guc); > if (ret) > goto out_unlock;
On Tue, Feb 27, 2018 at 03:37:25PM +0100, Michal Wajdeczko wrote: > On Tue, 27 Feb 2018 13:52:24 +0100, Michał Winiarski > <michal.winiarski@intel.com> wrote: > > > We only need those fast memcpy_wc when we're using relay to read > > continuous GuC log. Let's prevent the user from creating a relay if we > > know we won't be able to keep up with GuC. > > > > Signed-off-by: Michał Winiarski <michal.winiarski@intel.com> > > Cc: Chris Wilson <chris@chris-wilson.co.uk> > > Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> > > Cc: Sagar Arun Kamble <sagar.a.kamble@intel.com> > > Cc: Michal Wajdeczko <michal.wajdeczko@intel.com> > > --- > > drivers/gpu/drm/i915/intel_guc_log.c | 20 ++++++++++---------- > > 1 file changed, 10 insertions(+), 10 deletions(-) > > > > diff --git a/drivers/gpu/drm/i915/intel_guc_log.c > > b/drivers/gpu/drm/i915/intel_guc_log.c > > index 4dee65692f5f..d2aca10ab986 100644 > > --- a/drivers/gpu/drm/i915/intel_guc_log.c > > +++ b/drivers/gpu/drm/i915/intel_guc_log.c > > @@ -452,16 +452,6 @@ int intel_guc_log_create(struct intel_guc *guc) > > GEM_BUG_ON(guc->log.vma); > > - /* > > - * We require SSE 4.1 for fast reads from the GuC log buffer and > > - * it should be present on the chipsets supporting GuC based > > - * submisssions. > > - */ > > - if (WARN_ON(!i915_has_memcpy_from_wc())) { > > - ret = -EINVAL; > > - goto err; > > - } > > - > > vma = intel_guc_allocate_vma(guc, GUC_LOG_SIZE); > > if (IS_ERR(vma)) { > > ret = PTR_ERR(vma); > > @@ -568,6 +558,16 @@ int intel_guc_log_relay_open(struct intel_guc *guc) > > goto out_unlock; > > } > > + /* > > + * We require SSE 4.1 for fast reads from the GuC log buffer and > > + * it should be present on the chipsets supporting GuC based > > + * submisssions. > > + */ > > + if (!i915_has_memcpy_from_wc()) { > > + ret = -EINVAL; > > hmm, maybe -ENODEV as this is not user fault ENODEV means that we're not using GuC. -ENXIO? -Michał > > > + goto out_unlock; > > + } > > + > > ret = guc_log_relay_create(guc); > > if (ret) > > goto out_unlock;
On 2/27/2018 6:22 PM, Michał Winiarski wrote: > We only need those fast memcpy_wc when we're using relay to read > continuous GuC log. Let's prevent the user from creating a relay if we > know we won't be able to keep up with GuC. > > Signed-off-by: Michał Winiarski <michal.winiarski@intel.com> > Cc: Chris Wilson <chris@chris-wilson.co.uk> > Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> > Cc: Sagar Arun Kamble <sagar.a.kamble@intel.com> > Cc: Michal Wajdeczko <michal.wajdeczko@intel.com> Reviewed-by: Sagar Arun Kamble <sagar.a.kamble@intel.com> > --- > drivers/gpu/drm/i915/intel_guc_log.c | 20 ++++++++++---------- > 1 file changed, 10 insertions(+), 10 deletions(-) > > diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c > index 4dee65692f5f..d2aca10ab986 100644 > --- a/drivers/gpu/drm/i915/intel_guc_log.c > +++ b/drivers/gpu/drm/i915/intel_guc_log.c > @@ -452,16 +452,6 @@ int intel_guc_log_create(struct intel_guc *guc) > > GEM_BUG_ON(guc->log.vma); > > - /* > - * We require SSE 4.1 for fast reads from the GuC log buffer and > - * it should be present on the chipsets supporting GuC based > - * submisssions. > - */ > - if (WARN_ON(!i915_has_memcpy_from_wc())) { > - ret = -EINVAL; > - goto err; > - } > - > vma = intel_guc_allocate_vma(guc, GUC_LOG_SIZE); > if (IS_ERR(vma)) { > ret = PTR_ERR(vma); > @@ -568,6 +558,16 @@ int intel_guc_log_relay_open(struct intel_guc *guc) > goto out_unlock; > } > > + /* > + * We require SSE 4.1 for fast reads from the GuC log buffer and > + * it should be present on the chipsets supporting GuC based > + * submisssions. > + */ > + if (!i915_has_memcpy_from_wc()) { > + ret = -EINVAL; > + goto out_unlock; > + } > + > ret = guc_log_relay_create(guc); > if (ret) > goto out_unlock;
Quoting Michał Winiarski (2018-02-27 15:37:04) > On Tue, Feb 27, 2018 at 03:37:25PM +0100, Michal Wajdeczko wrote: > > On Tue, 27 Feb 2018 13:52:24 +0100, Michał Winiarski > > <michal.winiarski@intel.com> wrote: > > > > > We only need those fast memcpy_wc when we're using relay to read > > > continuous GuC log. Let's prevent the user from creating a relay if we > > > know we won't be able to keep up with GuC. > > > > > > Signed-off-by: Michał Winiarski <michal.winiarski@intel.com> > > > Cc: Chris Wilson <chris@chris-wilson.co.uk> > > > Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> > > > Cc: Sagar Arun Kamble <sagar.a.kamble@intel.com> > > > Cc: Michal Wajdeczko <michal.wajdeczko@intel.com> > > > --- > > > drivers/gpu/drm/i915/intel_guc_log.c | 20 ++++++++++---------- > > > 1 file changed, 10 insertions(+), 10 deletions(-) > > > > > > diff --git a/drivers/gpu/drm/i915/intel_guc_log.c > > > b/drivers/gpu/drm/i915/intel_guc_log.c > > > index 4dee65692f5f..d2aca10ab986 100644 > > > --- a/drivers/gpu/drm/i915/intel_guc_log.c > > > +++ b/drivers/gpu/drm/i915/intel_guc_log.c > > > @@ -452,16 +452,6 @@ int intel_guc_log_create(struct intel_guc *guc) > > > GEM_BUG_ON(guc->log.vma); > > > - /* > > > - * We require SSE 4.1 for fast reads from the GuC log buffer and > > > - * it should be present on the chipsets supporting GuC based > > > - * submisssions. > > > - */ > > > - if (WARN_ON(!i915_has_memcpy_from_wc())) { > > > - ret = -EINVAL; > > > - goto err; > > > - } > > > - > > > vma = intel_guc_allocate_vma(guc, GUC_LOG_SIZE); > > > if (IS_ERR(vma)) { > > > ret = PTR_ERR(vma); > > > @@ -568,6 +558,16 @@ int intel_guc_log_relay_open(struct intel_guc *guc) > > > goto out_unlock; > > > } > > > + /* > > > + * We require SSE 4.1 for fast reads from the GuC log buffer and > > > + * it should be present on the chipsets supporting GuC based > > > + * submisssions. > > > + */ > > > + if (!i915_has_memcpy_from_wc()) { > > > + ret = -EINVAL; > > > > hmm, maybe -ENODEV as this is not user fault > > ENODEV means that we're not using GuC. > -ENXIO? Do we respond to the error at all? It should just result in the relay not being available, right? (i.e. are we just discussing the color of the bikeshed :) -Chris
diff --git a/drivers/gpu/drm/i915/intel_guc_log.c b/drivers/gpu/drm/i915/intel_guc_log.c index 4dee65692f5f..d2aca10ab986 100644 --- a/drivers/gpu/drm/i915/intel_guc_log.c +++ b/drivers/gpu/drm/i915/intel_guc_log.c @@ -452,16 +452,6 @@ int intel_guc_log_create(struct intel_guc *guc) GEM_BUG_ON(guc->log.vma); - /* - * We require SSE 4.1 for fast reads from the GuC log buffer and - * it should be present on the chipsets supporting GuC based - * submisssions. - */ - if (WARN_ON(!i915_has_memcpy_from_wc())) { - ret = -EINVAL; - goto err; - } - vma = intel_guc_allocate_vma(guc, GUC_LOG_SIZE); if (IS_ERR(vma)) { ret = PTR_ERR(vma); @@ -568,6 +558,16 @@ int intel_guc_log_relay_open(struct intel_guc *guc) goto out_unlock; } + /* + * We require SSE 4.1 for fast reads from the GuC log buffer and + * it should be present on the chipsets supporting GuC based + * submisssions. + */ + if (!i915_has_memcpy_from_wc()) { + ret = -EINVAL; + goto out_unlock; + } + ret = guc_log_relay_create(guc); if (ret) goto out_unlock;
We only need those fast memcpy_wc when we're using relay to read continuous GuC log. Let's prevent the user from creating a relay if we know we won't be able to keep up with GuC. Signed-off-by: Michał Winiarski <michal.winiarski@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com> Cc: Sagar Arun Kamble <sagar.a.kamble@intel.com> Cc: Michal Wajdeczko <michal.wajdeczko@intel.com> --- drivers/gpu/drm/i915/intel_guc_log.c | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-)