Skip to contentSkip to conversation

drm/i915: Always program CSR if CSR is uninitialized

PATCH Superseded Patrik Jakobsson 3 participants 3 messages Oct 21, 2015 – Oct 22, 2015

Subsystems (2)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 drm/i915: Always program CSR if CSR is uninitialized

The current CSR loading code depends on the CSR program memory to be cleared after boot. This is unfortunately not true on all hardware. Instead make use of the FW_UNINITIALIZED state in init and check for FW_LOADED to prevent init path from skipping the actual programming.

---
 drivers/gpu/drm/i915/intel_csr.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_csr.c +3 -1
index 9e530a7..0f7c49e 100644
@@ -271,7 +271,7 @@ void intel_csr_load_program(struct drm_device *dev)
 	 * Unfortunately the ACPI subsystem doesn't yet give us a way to
 	 * differentiate this, hence figure it out with this hack.
 	 */
-	if (I915_READ(CSR_PROGRAM(0)))
+	if (I915_READ(CSR_PROGRAM(0)) && dev_priv->csr.state == FW_LOADED)
 		return;
 
 	mutex_lock(&dev_priv->csr_lock);
@@ -425,6 +425,8 @@ void intel_csr_ucode_init(struct drm_device *dev)
 	struct intel_csr *csr = &dev_priv->csr;
 	int ret;
 
+	intel_csr_load_status_set(dev_priv, FW_UNINITIALIZED);
+
 	if (!HAS_CSR(dev))
 		return;
 
Signature · 6 lines
2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Tested-by Reviewed-by on v1 1/1
#1 lore raw

regarding your offline question: yes, I had your patch applied here, so

Quoted 24 lines · Patrik Jakobsson

The current CSR loading code depends on the CSR program memory to be cleared after boot. This is unfortunately not true on all hardware. Instead make use of the FW_UNINITIALIZED state in init and check for FW_LOADED to prevent init path from skipping the actual programming.

---
 drivers/gpu/drm/i915/intel_csr.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_csr.c +1 -1
index 9e530a7..0f7c49e 100644
@@ -271,7 +271,7 @@ void intel_csr_load_program(struct drm_device *dev)
         * Unfortunately the ACPI subsystem doesn't yet give us a way to
         * differentiate this, hence figure it out with this hack.
         */
-       if (I915_READ(CSR_PROGRAM(0)))
+       if (I915_READ(CSR_PROGRAM(0)) && dev_priv->csr.state == FW_LOADED)
                return;
 
        mutex_lock(&dev_priv->csr_lock);

drivers/gpu/drm/i915/intel_csr.c void intel_csr_ucode_init(struct drm_device *dev) · line 430

Quoted 6 lines
drivers/gpu/drm/i915/intel_csr.c +2 -0
@@ -425,6 +425,8 @@ void intel_csr_ucode_init(struct drm_device *dev)
        struct intel_csr *csr = &dev_priv->csr;
        int ret;
 
+       intel_csr_load_status_set(dev_priv, FW_UNINITIALIZED);
+
        if (!HAS_CSR(dev))

My bikesheding here is that I like this kind of HAS_FEATURE() protection before anything else so we are always sure that useless code won't be executed in case you are running it on a platform that doesn't have this feature.

With this change also fell free to use a

Quoted 9 lines
drivers/gpu/drm/i915/intel_csr.c +0 -1
                return;
 
--

2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Signature · 6 lines
Rodrigo Vivi
Blog: http://blog.vivi.eng.br
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#2 lore raw
Quoted 33 lines · Rodrigo Vivi

regarding your offline question: yes, I had your patch applied here, so

Quoted 27 lines · Patrik Jakobsson

The current CSR loading code depends on the CSR program memory to be cleared after boot. This is unfortunately not true on all hardware. Instead make use of the FW_UNINITIALIZED state in init and check for FW_LOADED to prevent init path from skipping the actual programming.

---
 drivers/gpu/drm/i915/intel_csr.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_csr.c +1 -1
index 9e530a7..0f7c49e 100644
@@ -271,7 +271,7 @@ void intel_csr_load_program(struct drm_device *dev)
         * Unfortunately the ACPI subsystem doesn't yet give us a way to
         * differentiate this, hence figure it out with this hack.
         */
-       if (I915_READ(CSR_PROGRAM(0)))
+       if (I915_READ(CSR_PROGRAM(0)) && dev_priv->csr.state == FW_LOADED)
                return;
 
        mutex_lock(&dev_priv->csr_lock);
@@ -425,6 +425,8 @@ void intel_csr_ucode_init(struct drm_device *dev)
        struct intel_csr *csr = &dev_priv->csr;
        int ret;
 
Quoted 8 lines

drivers/gpu/drm/i915/intel_csr.c void intel_csr_ucode_init(struct drm_device *dev) · line 430

Quoted 3 lines
drivers/gpu/drm/i915/intel_csr.c +2 -0
+       intel_csr_load_status_set(dev_priv, FW_UNINITIALIZED);
+
        if (!HAS_CSR(dev))

My bikesheding here is that I like this kind of HAS_FEATURE() protection before anything else so we are always sure that useless code won't be executed in case you are running it on a platform that doesn't have this feature.

This also crossed my mind. My reasoning was that it's nice to initialize this value instead of leaving it undefined but on the other hand FW_UNINITIALIZED = 0 so the untouched state is still good. I don't feel strongly about this so I'll change it and send out a v2.

Thanks for the review!
-Patrik

Quoted 16 lines

With this change also fell free to use a

Quoted 9 lines
drivers/gpu/drm/i915/intel_csr.c +0 -1
                return;
 
--

2.1.4

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Quoted 7 lines

--
Rodrigo Vivi
Blog: http://blog.vivi.eng.br
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx