| Message ID | 20231213004237.20375-3-ville.syrjala@linux.intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | drm/i915: (stolen) memory region related fixes | expand |
On 13.12.2023 01:42, Ville Syrjala wrote: > From: Ville Syrjälä <ville.syrjala@linux.intel.com> > > Dump the details about every memory region into dmesg at probe time. > Avoids having to dig those out from random places when debugging stuff. > > Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> > --- > drivers/gpu/drm/i915/intel_memory_region.c | 18 ++++++++++++++++++ > 1 file changed, 18 insertions(+) > > diff --git a/drivers/gpu/drm/i915/intel_memory_region.c b/drivers/gpu/drm/i915/intel_memory_region.c > index b2708f8cac2a..52d998e5c21a 100644 > --- a/drivers/gpu/drm/i915/intel_memory_region.c > +++ b/drivers/gpu/drm/i915/intel_memory_region.c > @@ -372,6 +372,24 @@ int intel_memory_regions_hw_probe(struct drm_i915_private *i915) > i915->mm.regions[i] = mem; > } > > + for (i = 0; i < ARRAY_SIZE(i915->mm.regions); i++) { > + struct intel_memory_region *mem = i915->mm.regions[i]; > + u64 region_size, io_size; > + > + if (!mem) > + continue; > + > + region_size = resource_size(&mem->region) >> 20; > + io_size = resource_size(&mem->io) >> 20; > + > + if (resource_size(&mem->io)) > + drm_dbg(&i915->drm, "Memory region(%d): %s: %llu MiB %pR, io: %llu MiB %pR\n", > + mem->id, mem->name, region_size, &mem->region, io_size, &mem->io); > + else > + drm_dbg(&i915->drm, "Memory region(%d): %s: %llu MiB %pR, io: n/a\n", > + mem->id, mem->name, region_size, &mem->region); Doesn't printk handle properly 0-length resources? Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com> Regards Andrzej > + } > + > return 0; > > out_cleanup:
On Wed, Dec 13, 2023 at 05:05:21PM +0100, Andrzej Hajda wrote: > On 13.12.2023 01:42, Ville Syrjala wrote: > > From: Ville Syrjälä <ville.syrjala@linux.intel.com> > > > > Dump the details about every memory region into dmesg at probe time. > > Avoids having to dig those out from random places when debugging stuff. > > > > Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> > > --- > > drivers/gpu/drm/i915/intel_memory_region.c | 18 ++++++++++++++++++ > > 1 file changed, 18 insertions(+) > > > > diff --git a/drivers/gpu/drm/i915/intel_memory_region.c b/drivers/gpu/drm/i915/intel_memory_region.c > > index b2708f8cac2a..52d998e5c21a 100644 > > --- a/drivers/gpu/drm/i915/intel_memory_region.c > > +++ b/drivers/gpu/drm/i915/intel_memory_region.c > > @@ -372,6 +372,24 @@ int intel_memory_regions_hw_probe(struct drm_i915_private *i915) > > i915->mm.regions[i] = mem; > > } > > > > + for (i = 0; i < ARRAY_SIZE(i915->mm.regions); i++) { > > + struct intel_memory_region *mem = i915->mm.regions[i]; > > + u64 region_size, io_size; > > + > > + if (!mem) > > + continue; > > + > > + region_size = resource_size(&mem->region) >> 20; > > + io_size = resource_size(&mem->io) >> 20; > > + > > + if (resource_size(&mem->io)) > > + drm_dbg(&i915->drm, "Memory region(%d): %s: %llu MiB %pR, io: %llu MiB %pR\n", > > + mem->id, mem->name, region_size, &mem->region, io_size, &mem->io); > > + else > > + drm_dbg(&i915->drm, "Memory region(%d): %s: %llu MiB %pR, io: n/a\n", > > + mem->id, mem->name, region_size, &mem->region); > > Doesn't printk handle properly 0-length resources? Not without extra help. Apparently there is IORESOURCE_UNSET that would just print the size, but you have to explicitly set that yourself, which I suppose we could do. > > Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com> > > Regards > Andrzej > > > > + } > > + > > return 0; > > > > out_cleanup:
On Wed, Dec 13, 2023 at 06:26:13PM +0200, Ville Syrjälä wrote: > On Wed, Dec 13, 2023 at 05:05:21PM +0100, Andrzej Hajda wrote: > > On 13.12.2023 01:42, Ville Syrjala wrote: > > > From: Ville Syrjälä <ville.syrjala@linux.intel.com> > > > > > > Dump the details about every memory region into dmesg at probe time. > > > Avoids having to dig those out from random places when debugging stuff. > > > > > > Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> > > > --- > > > drivers/gpu/drm/i915/intel_memory_region.c | 18 ++++++++++++++++++ > > > 1 file changed, 18 insertions(+) > > > > > > diff --git a/drivers/gpu/drm/i915/intel_memory_region.c b/drivers/gpu/drm/i915/intel_memory_region.c > > > index b2708f8cac2a..52d998e5c21a 100644 > > > --- a/drivers/gpu/drm/i915/intel_memory_region.c > > > +++ b/drivers/gpu/drm/i915/intel_memory_region.c > > > @@ -372,6 +372,24 @@ int intel_memory_regions_hw_probe(struct drm_i915_private *i915) > > > i915->mm.regions[i] = mem; > > > } > > > > > > + for (i = 0; i < ARRAY_SIZE(i915->mm.regions); i++) { > > > + struct intel_memory_region *mem = i915->mm.regions[i]; > > > + u64 region_size, io_size; > > > + > > > + if (!mem) > > > + continue; > > > + > > > + region_size = resource_size(&mem->region) >> 20; > > > + io_size = resource_size(&mem->io) >> 20; > > > + > > > + if (resource_size(&mem->io)) > > > + drm_dbg(&i915->drm, "Memory region(%d): %s: %llu MiB %pR, io: %llu MiB %pR\n", > > > + mem->id, mem->name, region_size, &mem->region, io_size, &mem->io); > > > + else > > > + drm_dbg(&i915->drm, "Memory region(%d): %s: %llu MiB %pR, io: n/a\n", > > > + mem->id, mem->name, region_size, &mem->region); > > > > Doesn't printk handle properly 0-length resources? > > Not without extra help. Apparently there is IORESOURCE_UNSET > that would just print the size, but you have to explicitly > set that yourself, which I suppose we could do. This is what we get with nothing: Memory region(6): stolen-local: 20 MiB [mem 0x3fe800000-0x3ffbfffff], io: 0 MiB [mem 0x00000000-0xffffffffffffffff] This is with IORESOURCE_UNSET: Memory region(6): stolen-local: 20 MiB [mem 0x3fe800000-0x3ffbfffff], io: 0 MiB [mem size 0x00000000] And this is the original: Memory region(6): stolen-local: 20 MiB [mem 0x3fe800000-0x3ffbfffff], io: n/a I kinda like the explicit 'n/a' to make it obvious there is noting there. Also using IORESOURCE_UNSET would be a bit more annoying as we'd have to explicitly preserve it in places where we resize the resource (stolen reseved handling and selftests). But I suppose it would be just a matter of using the full DEFINE_RES_NAMED() instead of DEFINE_RES_MEM(). Hmm, maybe I should do that anyway just to be safe in case someone starts to use some extra flags on these... > > > > > Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com> > > > > Regards > > Andrzej > > > > > > > + } > > > + > > > return 0; > > > > > > out_cleanup: > > -- > Ville Syrjälä > Intel
diff --git a/drivers/gpu/drm/i915/intel_memory_region.c b/drivers/gpu/drm/i915/intel_memory_region.c index b2708f8cac2a..52d998e5c21a 100644 --- a/drivers/gpu/drm/i915/intel_memory_region.c +++ b/drivers/gpu/drm/i915/intel_memory_region.c @@ -372,6 +372,24 @@ int intel_memory_regions_hw_probe(struct drm_i915_private *i915) i915->mm.regions[i] = mem; } + for (i = 0; i < ARRAY_SIZE(i915->mm.regions); i++) { + struct intel_memory_region *mem = i915->mm.regions[i]; + u64 region_size, io_size; + + if (!mem) + continue; + + region_size = resource_size(&mem->region) >> 20; + io_size = resource_size(&mem->io) >> 20; + + if (resource_size(&mem->io)) + drm_dbg(&i915->drm, "Memory region(%d): %s: %llu MiB %pR, io: %llu MiB %pR\n", + mem->id, mem->name, region_size, &mem->region, io_size, &mem->io); + else + drm_dbg(&i915->drm, "Memory region(%d): %s: %llu MiB %pR, io: n/a\n", + mem->id, mem->name, region_size, &mem->region); + } + return 0; out_cleanup: