drm/i915: add stepping macro
Subsystems (2)
PATCH 1/2 drm/i915: add stepping macro
In some cases to enable or disable features & workarounds, we may need to check the GPU stepping. Add a macro to do that based on caching the PCI revision ID reg.
--- drivers/gpu/drm/i915/i915_dma.c | 2 ++ drivers/gpu/drm/i915/i915_drv.h | 2 ++ 2 files changed, 4 insertions(+)
drivers/gpu/drm/i915/i915_dma.c +2 -0
@@ -1565,6 +1565,8 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags) device_info = (struct intel_device_info *)&dev_priv->info; *device_info = *info; +pci_read_config_byte (dev->pdev,PCI_REVISION_ID , &dev_priv->stepping); +spin_lock_init (&dev_priv->irq_lock);spin_lock_init (&dev_priv->gpu_error.lock);spin_lock_init (&dev_priv->backlight_lock);
drivers/gpu/drm/i915/i915_drv.h +2 -0
@@ -1341,6 +1341,7 @@ struct drm_i915_private {struct kmem_cache *slab;const struct intel_device_info info; +u8 stepping;int relative_constants_mode;
@@ -1867,6 +1868,7 @@ struct drm_i915_cmd_table { };#define INTEL_INFO (dev) (&to_i915(dev)->info) +#define INTEL_STEPPING (dev) (&to_i915(dev)->stepping)#define IS_I830 (dev) ((dev)->pdev->device == 0x3577)#define IS_845G (dev) ((dev)->pdev->device == 0x2562)
Signature · 1 line
1.8.4.2
PATCH 2/2 drm/i915: use SNB A step FDI values on A step CPUs
More of an example of where to use the stepping macro than anything else. I think these early steppings never went outside of Intel.
--- drivers/gpu/drm/i915/intel_display.c | 17 +++++++++++++++-- 1 file changed, 15 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/intel_display.c +15 -2
@@ -2883,6 +2883,13 @@ static void ironlake_fdi_link_train(struct drm_crtc *crtc) } +static const int snb_a_fdi_train_param[] = { +FDI_LINK_TRAIN_400MV_0DB_SNB_A , +FDI_LINK_TRAIN_400MV_6DB_SNB_A , +FDI_LINK_TRAIN_600MV_3_5DB_SNB_A , +FDI_LINK_TRAIN_800MV_0DB_SNB_A , +}; +static const int snb_b_fdi_train_param[] = {FDI_LINK_TRAIN_400MV_0DB_SNB_B ,FDI_LINK_TRAIN_400MV_6DB_SNB_B ,
@@ -2898,6 +2905,12 @@ static void gen6_fdi_link_train(struct drm_crtc *crtc)struct intel_crtc *intel_crtc =to_intel_crtc (crtc);int pipe = intel_crtc->pipe;u32 reg, temp, i, retry; +const int *train_param; + +if (INTEL_STEPPING (dev) ==0 ) + train_param = snb_a_fdi_train_param; +else + train_param = snb_b_fdi_train_param;/* Train 1: umask FDI RX Interrupt symbol_lock and bit_lock bit for train result */
@@ -2943,7 +2956,7 @@ static void gen6_fdi_link_train(struct drm_crtc *crtc) reg =FDI_TX_CTL (pipe); temp =I915_READ (reg); temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK ; - temp |= snb_b_fdi_train_param[i]; + temp |= train_param[i];I915_WRITE (reg, temp);POSTING_READ (reg);
@@ -2996,7 +3009,7 @@ static void gen6_fdi_link_train(struct drm_crtc *crtc) reg =FDI_TX_CTL (pipe); temp =I915_READ (reg); temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK ; - temp |= snb_b_fdi_train_param[i]; + temp |= train_param[i];I915_WRITE (reg, temp);POSTING_READ (reg);
Signature · 1 line
1.8.4.2
Quoted 3 lines · Jesse Barnes
In some cases to enable or disable features & workarounds, we may need to check the GPU stepping. Add a macro to do that based on caching the PCI revision ID reg.
What about just using dev->pdev->revision like we already do? -Chris
Signature · 1 line
Chris Wilson, Intel Open Source Technology Centre
Quoted 6 lines · On Wed, 21 May 2014 19:50:53 +0100 Chris Wilson <[email protected]>…
Quoted 3 lines · Jesse Barnes
In some cases to enable or disable features & workarounds, we may need to check the GPU stepping. Add a macro to do that based on caching the PCI revision ID reg.
What about just using
dev->pdev->revisionlike we already do?
Oh missed that, was looking for stepping comments and didn't see it. Yeah that's fine.
Signature · 1 line
Jesse Barnes, Intel Open Source Technology Center