Skip to contentSkip to conversation

i915 fixes for sparse warnings.

PATCH In review Ben Widawsky 4 participants 16 messages Apr 16, 2012 – Apr 21, 2012

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

cover i915 fixes for sparse warnings.

Patch 1/2 is a mix of fairly trivial stuff. I see no reason to not take these.
Patch 3-9 Are all __iomem related. These should run through some regression testing.

Most of the __iomem stuff fell into place fairly easily, but it's definitely subject to a lot of copy/paste errors, since trying to automate the replacement didn't workout well. Throughout the patches, I tried really hard to not force cast things, but a couple of situations cropped up which gave me no choice. What I hope to achieve by that is there should be no functional or performance changes (aside from copy/paste).

I've given up on fixing intel_bios.c completely.

These are based off of dinq as of yesterday.

Ben Widawsky (9):
drm/i915: [sparse] trivial sparse fixes drm/i915: [sparse] don't use variable size arrays drm/i915: [sparse] __iomem fixes for opregion drm/i915: [sparse] __iomem fixes for overlay drm/i915: [sparse] __iomem fixes for debugfs drm/i915: [sparse] __iomem fixes for ringbuffer drm/i915: [sparse] forced __iomem ringbuffer fixes drm/i915: [sparse] __iomem fixes for gem drm/i915: [sparse] __iomem fixes for intel_bios

 drivers/gpu/drm/i915/i915_debugfs.c      |   28 ++++---
 drivers/gpu/drm/i915/i915_drv.h          |   12 +--
 drivers/gpu/drm/i915/i915_gem.c          |   11 +--
 drivers/gpu/drm/i915/i915_ioc32.c        |    5 +-
 drivers/gpu/drm/i915/i915_trace_points.c |    2 +
 drivers/gpu/drm/i915/intel_acpi.c        |    1 +
 drivers/gpu/drm/i915/intel_bios.c        |   20 +++--
 drivers/gpu/drm/i915/intel_display.c     |    2 +-
 drivers/gpu/drm/i915/intel_fb.c          |    2 +-
 drivers/gpu/drm/i915/intel_opregion.c    |   67 ++++++++-------
 drivers/gpu/drm/i915/intel_overlay.c     |  130 +++++++++++++++++-------------
 drivers/gpu/drm/i915/intel_ringbuffer.c  |   28 ++++---
 drivers/gpu/drm/i915/intel_sdvo.c        |   26 ++++--
 13 files changed, 197 insertions(+), 137 deletions(-)
Signature · 1 line
1.7.10
#1 lore raw

PATCH 1/9 drm/i915: [sparse] trivial sparse fixes

This should contain all the changes which require no thought to make sparse happy.

---
 drivers/gpu/drm/i915/i915_debugfs.c      |    2 +-
 drivers/gpu/drm/i915/i915_drv.h          |    2 ++
 drivers/gpu/drm/i915/i915_ioc32.c        |    5 ++++-
 drivers/gpu/drm/i915/i915_trace_points.c |    2 ++
 drivers/gpu/drm/i915/intel_acpi.c        |    1 +
 drivers/gpu/drm/i915/intel_display.c     |    2 +-
 drivers/gpu/drm/i915/intel_fb.c          |    2 +-
 drivers/gpu/drm/i915/intel_ringbuffer.c  |    4 ++--
 drivers/gpu/drm/i915/intel_sdvo.c        |    2 +-
 9 files changed, 15 insertions(+), 7 deletions(-)
drivers/gpu/drm/i915/i915_debugfs.c +1 -1
index 97c9630..35462df 100644
@@ -1832,7 +1832,7 @@ static int i915_forcewake_open(struct inode *inode, struct file *file)
 	return 0;
 }
 
-int i915_forcewake_release(struct inode *inode, struct file *file)
+static int i915_forcewake_release(struct inode *inode, struct file *file)
 {
 	struct drm_device *dev = inode->i_private;
 	struct drm_i915_private *dev_priv = dev->dev_private;
drivers/gpu/drm/i915/i915_drv.h +2 -0
index 422f424..303cee7 100644
@@ -1128,8 +1128,10 @@ extern void i915_driver_preclose(struct drm_device *dev,
 extern void i915_driver_postclose(struct drm_device *dev,
 				  struct drm_file *file_priv);
 extern int i915_driver_device_is_agp(struct drm_device * dev);
+#ifdef CONFIG_COMPAT
 extern long i915_compat_ioctl(struct file *filp, unsigned int cmd,
 			      unsigned long arg);
+#endif
 extern int i915_emit_box(struct drm_device *dev,
 			 struct drm_clip_rect *box,
 			 int DR1, int DR4);
drivers/gpu/drm/i915/i915_ioc32.c +4 -1
index 13b0289..0e72abb 100644
@@ -34,6 +34,7 @@
 #include "drmP.h"
 #include "drm.h"
 #include "i915_drm.h"
+#include "i915_drv.h"
 
 typedef struct _drm_i915_batchbuffer32 {
 	int start;		/* agp offset */
@@ -181,7 +182,7 @@ static int compat_i915_alloc(struct file *file, unsigned int cmd,
 			 (unsigned long)request);
 }
 
-drm_ioctl_compat_t *i915_compat_ioctls[] = {
+static drm_ioctl_compat_t *i915_compat_ioctls[] = {
 	[DRM_I915_BATCHBUFFER] = compat_i915_batchbuffer,
 	[DRM_I915_CMDBUFFER] = compat_i915_cmdbuffer,
 	[DRM_I915_GETPARAM] = compat_i915_getparam,
@@ -189,6 +190,7 @@ drm_ioctl_compat_t *i915_compat_ioctls[] = {
 	[DRM_I915_ALLOC] = compat_i915_alloc
 };
 
+#ifdef CONFIG_COMPAT
 /**
  * Called whenever a 32-bit process running under a 64-bit kernel
  * performs an ioctl on /dev/dri/card<n>.
@@ -217,3 +219,4 @@ long i915_compat_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
 
 	return ret;
 }
+#endif
drivers/gpu/drm/i915/i915_trace_points.c +2 -0
index ead876e..f1df2bd 100644
@@ -7,5 +7,7 @@
 
 #include "i915_drv.h"
 
+#ifndef __CHECKER__
 #define CREATE_TRACE_POINTS
 #include "i915_trace.h"
+#endif
drivers/gpu/drm/i915/intel_acpi.c +1 -0
index f152b2a..f413899 100644
@@ -9,6 +9,7 @@
 #include <acpi/acpi_drivers.h>
 
 #include "drmP.h"
+#include "i915_drv.h"
 
 #define INTEL_DSM_REVISION_ID 1 /* For Calpella anyway... */
 
drivers/gpu/drm/i915/intel_display.c +1 -1
index 33aaad3..c457716 100644
@@ -9460,7 +9460,7 @@ struct intel_quirk {
 	void (*hook)(struct drm_device *dev);
 };
 
-struct intel_quirk intel_quirks[] = {
+static struct intel_quirk intel_quirks[] = {
 	/* HP Mini needs pipe A force quirk (LP: #322104) */
 	{ 0x27ae, 0x103c, 0x361a, quirk_pipea_force },
 
drivers/gpu/drm/i915/intel_fb.c +1 -1
index 19ecd78..71ef289 100644
@@ -94,7 +94,7 @@ static int intelfb_create(struct intel_fbdev *ifbdev,
 	mutex_lock(&dev->struct_mutex);
 
 	/* Flush everything out, we'll be doing GTT only from now on */
-	ret = intel_pin_and_fence_fb_obj(dev, obj, false);
+	ret = intel_pin_and_fence_fb_obj(dev, obj, NULL);
 	if (ret) {
 		DRM_ERROR("failed to pin fb: %d\n", ret);
 		goto out_unref;
drivers/gpu/drm/i915/intel_ringbuffer.c +2 -2
index 6c14457..caec699 100644
@@ -891,8 +891,8 @@ err:
 	return ret;
 }
 
-int intel_init_ring_buffer(struct drm_device *dev,
-			   struct intel_ring_buffer *ring)
+static int intel_init_ring_buffer(struct drm_device *dev,
+				  struct intel_ring_buffer *ring)
 {
 	struct drm_i915_gem_object *obj;
 	int ret;
drivers/gpu/drm/i915/intel_sdvo.c +1 -1
index 6898145..aa1a0d5 100644
@@ -1258,7 +1258,7 @@ intel_sdvo_get_analog_edid(struct drm_connector *connector)
 						    dev_priv->crt_ddc_pin));
 }
 
-enum drm_connector_status
+static enum drm_connector_status
 intel_sdvo_tmds_sink_detect(struct drm_connector *connector)
 {
 	struct intel_sdvo *intel_sdvo = intel_attached_sdvo(connector);
Signature · 1 line
1.7.10
Reviewed-by on v1 1/9
#2 lore raw

Re: [PATCH 1/9] drm/i915: [sparse] trivial sparse fixes

Quoted 2 lines · Ben Widawsky

This should contain all the changes which require no thought to make sparse happy.

In a couple of spots, we make a structure/array static. Are they suitable for const as well?

Otherwise, looks sane:

-Chris

Signature · 1 line
Chris Wilson, Intel Open Source Technology Centre
#3 lore raw

PATCH 2/9 drm/i915: [sparse] don't use variable size arrays

In reply to Ben Widawsky

Sparse doesn't like:
"error: bad constant expression"

---
 drivers/gpu/drm/i915/intel_sdvo.c |   24 ++++++++++++++++++------
 1 file changed, 18 insertions(+), 6 deletions(-)
drivers/gpu/drm/i915/intel_sdvo.c +18 -6
index aa1a0d5..1daa198 100644
@@ -443,9 +443,17 @@ static const char *cmd_status_names[] = {
 static bool intel_sdvo_write_cmd(struct intel_sdvo *intel_sdvo, u8 cmd,
 				 const void *args, int args_len)
 {
-	u8 buf[args_len*2 + 2], status;
-	struct i2c_msg msgs[args_len + 3];
-	int i, ret;
+	u8 *buf, status;
+	struct i2c_msg *msgs;
+	int i, ret = true;
+
+	buf = (u8 *)kzalloc(args_len * 2 + 2, GFP_KERNEL);
+	if (!buf)
+		return false;
+
+	msgs = drm_malloc_ab(args_len + 3, sizeof(*msgs));
+	if (!msgs)
+		return false;
 
 	intel_sdvo_debug_write(intel_sdvo, cmd, args, args_len);
 
@@ -479,15 +487,19 @@ static bool intel_sdvo_write_cmd(struct intel_sdvo *intel_sdvo, u8 cmd,
 	ret = i2c_transfer(intel_sdvo->i2c, msgs, i+3);
 	if (ret < 0) {
 		DRM_DEBUG_KMS("I2c transfer returned %d\n", ret);
-		return false;
+		ret = false;
+		goto out;
 	}
 	if (ret != i+3) {
 		/* failure in I2C transfer */
 		DRM_DEBUG_KMS("I2c transfer returned %d/%d\n", ret, i+3);
-		return false;
+		ret = false;
 	}
 
-	return true;
+out:
+	drm_free_large(msgs);
+	kfree(buf);
+	return ret;
 }
 
 static bool intel_sdvo_read_response(struct intel_sdvo *intel_sdvo,
Signature · 1 line
1.7.10
#4 lore raw

Re: [PATCH 2/9] drm/i915: [sparse] don't use variable size arrays

2012/4/16 Ben Widawsky <[email protected]>:

Quoted 3 lines

Sparse doesn't like:
"error: bad constant expression"

<bikeshedding>
I know you'll hate me for asking, but: how difficult is it to fix sparse? Adding those mallocs/frees increases the code complexity, making it harder to read...
</bikeshedding>

Signature · 1 line
Paulo Zanoni
#5 lore raw

Re: [PATCH 2/9] drm/i915: [sparse] don't use variable size arrays

Quoted 11 lines · On Tue, 17 Apr 2012 22:39:15 -0300 Paulo Zanoni <[email protected]> wrote:

2012/4/16 Ben Widawsky <[email protected]>:

Quoted 3 lines

Sparse doesn't like:
"error: bad constant expression"

<bikeshedding>
I know you'll hate me for asking, but: how difficult is it to fix sparse? Adding those mallocs/frees increases the code complexity, making it harder to read...
</bikeshedding>

I don't consider this a bikeshed. I've always been "under the impression" C99 was sort of taboo in the kernel. In this case specifically, it's never a great idea to allocate an unknown amount of stack space as it probably messes with some of the static tools and such.

In other words, I believe the right thing to do here is not to fix sparse. Plus there is precedent in other drivers to fix this kind of thing for sparse. I originally had this patch create an arbitrarily large object on the stack and fail if the args_len was too big. I can go back to that certainly if people prefer.

#6 lore raw

Re: [PATCH 2/9] drm/i915: [sparse] don't use variable size arrays

Quoted 21 lines · Ben Widawsky
Quoted 11 lines · On Tue, 17 Apr 2012 22:39:15 -0300 Paulo Zanoni <[email protected]> wrote:

2012/4/16 Ben Widawsky <[email protected]>:

Quoted 3 lines

Sparse doesn't like:
"error: bad constant expression"

<bikeshedding>
I know you'll hate me for asking, but: how difficult is it to fix sparse? Adding those mallocs/frees increases the code complexity, making it harder to read...
</bikeshedding>

I don't consider this a bikeshed. I've always been "under the impression" C99 was sort of taboo in the kernel. In this case specifically, it's never a great idea to allocate an unknown amount of stack space as it probably messes with some of the static tools and such.

Quoted 5 lines

In other words, I believe the right thing to do here is not to fix sparse. Plus there is precedent in other drivers to fix this kind of thing for sparse. I originally had this patch create an arbitrarily large object on the stack and fail if the args_len was too big. I can go back to that certainly if people prefer.

I've picked up the first 2 patches of this series for -next, with a tiny bikeshed on the 2nd one. I've gotten stuck while reviewing the next one, I think it'd be good if we can quickly discuss these on irc. -Daniel

Signature · 3 lines
Daniel Vetter
Mail: [email protected]
Mobile: +41 (0)79 365 57 48
#7 lore raw

PATCH 3/9 drm/i915: [sparse] __iomem fixes for opregion

In reply to Ben Widawsky

Almost all of the errors related __iomem problems.

Most of the changes here are trivial, however there is plenty of chance for yank/paste errors.

---
 drivers/gpu/drm/i915/i915_drv.h       |   10 ++---
 drivers/gpu/drm/i915/intel_opregion.c |   67 +++++++++++++++++++--------------
 2 files changed, 43 insertions(+), 34 deletions(-)
drivers/gpu/drm/i915/i915_drv.h +5 -5
index 303cee7..1ca8d8d 100644
@@ -122,11 +122,11 @@ struct opregion_asle;
 struct drm_i915_private;
 
 struct intel_opregion {
-	struct opregion_header *header;
-	struct opregion_acpi *acpi;
-	struct opregion_swsci *swsci;
-	struct opregion_asle *asle;
-	void *vbt;
+	struct opregion_header __iomem *header;
+	struct opregion_acpi __iomem *acpi;
+	struct opregion_swsci __iomem *swsci;
+	struct opregion_asle __iomem *asle;
+	void __iomem *vbt;
 	u32 __iomem *lid_state;
 };
 #define OPREGION_SIZE            (8*1024)
drivers/gpu/drm/i915/intel_opregion.c +38 -29

207 lines not on this page · raw message

Signature · 1 line
1.7.10
#8 lore raw

PATCH 4/9 drm/i915: [sparse] __iomem fixes for overlay

In reply to Ben Widawsky

With the exception of a forced cast for phys_obj stuff (a problem in other patches as well) all of these are fairly simple __iomem compliance fixes.

As with other patches, yank/paste errors may exist.

---
 drivers/gpu/drm/i915/intel_overlay.c |  130 +++++++++++++++++++---------------
 1 file changed, 72 insertions(+), 58 deletions(-)
drivers/gpu/drm/i915/intel_overlay.c +72 -58
index 80b331c..95cc21a 100644
@@ -187,14 +187,14 @@ struct intel_overlay {
 	void (*flip_tail)(struct intel_overlay *);
 };
 
-static struct overlay_registers *
+static struct overlay_registers __iomem *
 intel_overlay_map_regs(struct intel_overlay *overlay)
 {
 	drm_i915_private_t *dev_priv = overlay->dev->dev_private;
-	struct overlay_registers *regs;
+	struct overlay_registers __iomem *regs;
 
 	if (OVERLAY_NEEDS_PHYSICAL(overlay->dev))
-		regs = overlay->reg_bo->phys_obj->handle->vaddr;
+		regs = (struct overlay_registers __iomem *)overlay->reg_bo->phys_obj->handle->vaddr;
 	else
 		regs = io_mapping_map_wc(dev_priv->mm.gtt_mapping,
 					 overlay->reg_bo->gtt_offset);
@@ -203,7 +203,7 @@ intel_overlay_map_regs(struct intel_overlay *overlay)
 }
 
 static void intel_overlay_unmap_regs(struct intel_overlay *overlay,
-				     struct overlay_registers *regs)
+				     struct overlay_registers __iomem *regs)
 {
 	if (!OVERLAY_NEEDS_PHYSICAL(overlay->dev))
 		io_mapping_unmap(regs);
@@ -619,14 +619,15 @@ static const u16 uv_static_hcoeffs[N_HORIZ_UV_TAPS * N_PHASES] = {
 	0x3000, 0x0800, 0x3000
 };
 
-static void update_polyphase_filter(struct overlay_registers *regs)
+static void update_polyphase_filter(struct overlay_registers __iomem *regs)
 {
-	memcpy(regs->Y_HCOEFS, y_static_hcoeffs, sizeof(y_static_hcoeffs));
-	memcpy(regs->UV_HCOEFS, uv_static_hcoeffs, sizeof(uv_static_hcoeffs));
+	memcpy_toio(regs->Y_HCOEFS, y_static_hcoeffs, sizeof(y_static_hcoeffs));
+	memcpy_toio(regs->UV_HCOEFS, uv_static_hcoeffs,
+		    sizeof(uv_static_hcoeffs));
 }
 
 static bool update_scaling_factors(struct intel_overlay *overlay,
-				   struct overlay_registers *regs,
+				   struct overlay_registers __iomem *regs,
 				   struct put_image_params *params)
 {
 	/* fixed point with a 12 bit shift */
@@ -665,16 +666,19 @@ static bool update_scaling_factors(struct intel_overlay *overlay,
 	overlay->old_xscale = xscale;
 	overlay->old_yscale = yscale;
 
-	regs->YRGBSCALE = (((yscale & FRACT_MASK) << 20) |
-			   ((xscale >> FP_SHIFT)  << 16) |
-			   ((xscale & FRACT_MASK) << 3));
+	iowrite32(((yscale & FRACT_MASK) << 20) |
+		  ((xscale >> FP_SHIFT)  << 16) |
+		  ((xscale & FRACT_MASK) << 3),
+		 &regs->YRGBSCALE);
 
-	regs->UVSCALE = (((yscale_UV & FRACT_MASK) << 20) |
-			 ((xscale_UV >> FP_SHIFT)  << 16) |
-			 ((xscale_UV & FRACT_MASK) << 3));
+	iowrite32(((yscale_UV & FRACT_MASK) << 20) |
+		  ((xscale_UV >> FP_SHIFT)  << 16) |
+		  ((xscale_UV & FRACT_MASK) << 3),
+		 &regs->UVSCALE);
 
-	regs->UVSCALEV = ((((yscale    >> FP_SHIFT) << 16) |
-			   ((yscale_UV >> FP_SHIFT) << 0)));
+	iowrite32((((yscale    >> FP_SHIFT) << 16) |
+		   ((yscale_UV >> FP_SHIFT) << 0)),
+		 &regs->UVSCALEV);
 
 	if (scale_changed)
 		update_polyphase_filter(regs);
@@ -683,30 +687,32 @@ static bool update_scaling_factors(struct intel_overlay *overlay,
 }
 
 static void update_colorkey(struct intel_overlay *overlay,
-			    struct overlay_registers *regs)
+			    struct overlay_registers __iomem *regs)
 {
 	u32 key = overlay->color_key;
 
 	switch (overlay->crtc->base.fb->bits_per_pixel) {
 	case 8:
-		regs->DCLRKV = 0;
-		regs->DCLRKM = CLK_RGB8I_MASK | DST_KEY_ENABLE;
+		iowrite32(0, &regs->DCLRKV);
+		iowrite32(CLK_RGB8I_MASK | DST_KEY_ENABLE, &regs->DCLRKM);
 		break;
 
 	case 16:
 		if (overlay->crtc->base.fb->depth == 15) {
-			regs->DCLRKV = RGB15_TO_COLORKEY(key);
-			regs->DCLRKM = CLK_RGB15_MASK | DST_KEY_ENABLE;
+			iowrite32(RGB15_TO_COLORKEY(key), &regs->DCLRKV);
+			iowrite32(CLK_RGB15_MASK | DST_KEY_ENABLE,
+				  &regs->DCLRKM);
 		} else {
-			regs->DCLRKV = RGB16_TO_COLORKEY(key);
-			regs->DCLRKM = CLK_RGB16_MASK | DST_KEY_ENABLE;
+			iowrite32(RGB16_TO_COLORKEY(key), &regs->DCLRKV);
+			iowrite32(CLK_RGB16_MASK | DST_KEY_ENABLE,
+				  &regs->DCLRKM);
 		}
 		break;
 
 	case 24:
 	case 32:
-		regs->DCLRKV = key;
-		regs->DCLRKM = CLK_RGB24_MASK | DST_KEY_ENABLE;
+		iowrite32(key, &regs->DCLRKV);
+		iowrite32(CLK_RGB24_MASK | DST_KEY_ENABLE, &regs->DCLRKM);
 		break;
 	}
 }
@@ -761,9 +767,10 @@ static int intel_overlay_do_put_image(struct intel_overlay *overlay,
 				      struct put_image_params *params)
 {
 	int ret, tmp_width;
-	struct overlay_registers *regs;
+	struct overlay_registers __iomem *regs;
 	bool scale_changed = false;
 	struct drm_device *dev = overlay->dev;
+	u32 swidth, swidthsw, sheight, ostride;
 
 	BUG_ON(!mutex_is_locked(&dev->struct_mutex));
 	BUG_ON(!mutex_is_locked(&dev->mode_config.mutex));
@@ -782,16 +789,18 @@ static int intel_overlay_do_put_image(struct intel_overlay *overlay,
 		goto out_unpin;
 
 	if (!overlay->active) {
+		u32 oconfig;
 		regs = intel_overlay_map_regs(overlay);
 		if (!regs) {
 			ret = -ENOMEM;
 			goto out_unpin;
 		}
-		regs->OCONFIG = OCONF_CC_OUT_8BIT;
+		oconfig = OCONF_CC_OUT_8BIT;
 		if (IS_GEN4(overlay->dev))
-			regs->OCONFIG |= OCONF_CSC_MODE_BT709;
-		regs->OCONFIG |= overlay->crtc->pipe == 0 ?
+			oconfig |= OCONF_CSC_MODE_BT709;
+		oconfig |= overlay->crtc->pipe == 0 ?
 			OCONF_PIPE_A : OCONF_PIPE_B;
+		iowrite32(oconfig, &regs->OCONFIG);
 		intel_overlay_unmap_regs(overlay, regs);
 
 		ret = intel_overlay_on(overlay);
@@ -805,42 +814,46 @@ static int intel_overlay_do_put_image(struct intel_overlay *overlay,
 		goto out_unpin;
 	}
 
-	regs->DWINPOS = (params->dst_y << 16) | params->dst_x;
-	regs->DWINSZ = (params->dst_h << 16) | params->dst_w;
+	iowrite32((params->dst_y << 16) | params->dst_x, &regs->DWINPOS);
+	iowrite32((params->dst_h << 16) | params->dst_w, &regs->DWINSZ);
 
 	if (params->format & I915_OVERLAY_YUV_PACKED)
 		tmp_width = packed_width_bytes(params->format, params->src_w);
 	else
 		tmp_width = params->src_w;
 
-	regs->SWIDTH = params->src_w;
-	regs->SWIDTHSW = calc_swidthsw(overlay->dev,
-				       params->offset_Y, tmp_width);
-	regs->SHEIGHT = params->src_h;
-	regs->OBUF_0Y = new_bo->gtt_offset + params->offset_Y;
-	regs->OSTRIDE = params->stride_Y;
+	swidth = params->src_w;
+	swidthsw = calc_swidthsw(overlay->dev, params->offset_Y, tmp_width);
+	sheight = params->src_h;
+	iowrite32(new_bo->gtt_offset + params->offset_Y, &regs->OBUF_0Y);
+	ostride = params->stride_Y;
 
 	if (params->format & I915_OVERLAY_YUV_PLANAR) {
 		int uv_hscale = uv_hsubsampling(params->format);
 		int uv_vscale = uv_vsubsampling(params->format);
 		u32 tmp_U, tmp_V;
-		regs->SWIDTH |= (params->src_w/uv_hscale) << 16;
+		swidth |= (params->src_w/uv_hscale) << 16;
 		tmp_U = calc_swidthsw(overlay->dev, params->offset_U,
 				      params->src_w/uv_hscale);
 		tmp_V = calc_swidthsw(overlay->dev, params->offset_V,
 				      params->src_w/uv_hscale);
-		regs->SWIDTHSW |= max_t(u32, tmp_U, tmp_V) << 16;
-		regs->SHEIGHT |= (params->src_h/uv_vscale) << 16;
-		regs->OBUF_0U = new_bo->gtt_offset + params->offset_U;
-		regs->OBUF_0V = new_bo->gtt_offset + params->offset_V;
-		regs->OSTRIDE |= params->stride_UV << 16;
+		swidthsw |= max_t(u32, tmp_U, tmp_V) << 16;
+		sheight |= (params->src_h/uv_vscale) << 16;
+		iowrite32(new_bo->gtt_offset + params->offset_U, &regs->OBUF_0U);
+		iowrite32(new_bo->gtt_offset + params->offset_V, &regs->OBUF_0V);
+		ostride |= params->stride_UV << 16;
 	}
 
+	iowrite32(swidth, &regs->SWIDTH);
+	iowrite32(swidthsw, &regs->SWIDTHSW);
+	iowrite32(sheight, &regs->SHEIGHT);
+	iowrite32(ostride, &regs->OSTRIDE);
+
 	scale_changed = update_scaling_factors(overlay, regs, params);
 
 	update_colorkey(overlay, regs);
 
-	regs->OCMD = overlay_cmd_reg(params);
+	iowrite32(overlay_cmd_reg(params), &regs->OCMD);
 
 	intel_overlay_unmap_regs(overlay, regs);
 
@@ -860,7 +873,7 @@ out_unpin:
 
 int intel_overlay_switch_off(struct intel_overlay *overlay)
 {
-	struct overlay_registers *regs;
+	struct overlay_registers __iomem *regs;
 	struct drm_device *dev = overlay->dev;
 	int ret;
 
@@ -879,7 +892,7 @@ int intel_overlay_switch_off(struct intel_overlay *overlay)
 		return ret;
 
 	regs = intel_overlay_map_regs(overlay);
-	regs->OCMD = 0;
+	iowrite32(0, &regs->OCMD);
 	intel_overlay_unmap_regs(overlay, regs);
 
 	ret = intel_overlay_off(overlay);
@@ -1250,10 +1263,11 @@ out_free:
 }
 
 static void update_reg_attrs(struct intel_overlay *overlay,
-			     struct overlay_registers *regs)
+			     struct overlay_registers __iomem *regs)
 {
-	regs->OCLRC0 = (overlay->contrast << 18) | (overlay->brightness & 0xff);
-	regs->OCLRC1 = overlay->saturation;
+	iowrite32((overlay->contrast << 18) | (overlay->brightness & 0xff),
+		  &regs->OCLRC0);
+	iowrite32(overlay->saturation, &regs->OCLRC1);
 }
 
 static bool check_gamma_bounds(u32 gamma1, u32 gamma2)
@@ -1306,7 +1320,7 @@ int intel_overlay_attrs(struct drm_device *dev, void *data,
 	struct drm_intel_overlay_attrs *attrs = data;
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	struct intel_overlay *overlay;
-	struct overlay_registers *regs;
+	struct overlay_registers __iomem *regs;
 	int ret;
 
 	if (!dev_priv) {
@@ -1396,7 +1410,7 @@ void intel_setup_overlay(struct drm_device *dev)
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	struct intel_overlay *overlay;
 	struct drm_i915_gem_object *reg_bo;
-	struct overlay_registers *regs;
+	struct overlay_registers __iomem *regs;
 	int ret;
 
 	if (!HAS_OVERLAY(dev))
@@ -1451,7 +1465,7 @@ void intel_setup_overlay(struct drm_device *dev)
 	if (!regs)
 		goto out_unpin_bo;
 
-	memset(regs, 0, sizeof(struct overlay_registers));
+	memset_io(regs, 0, sizeof(struct overlay_registers));
 	update_polyphase_filter(regs);
 	update_reg_attrs(overlay, regs);
 
@@ -1499,14 +1513,14 @@ struct intel_overlay_error_state {
 	u32 isr;
 };
 
-static struct overlay_registers *
+static struct overlay_registers __iomem *
 intel_overlay_map_regs_atomic(struct intel_overlay *overlay)
 {
 	drm_i915_private_t *dev_priv = overlay->dev->dev_private;
-	struct overlay_registers *regs;
+	struct overlay_registers __iomem *regs;
 
 	if (OVERLAY_NEEDS_PHYSICAL(overlay->dev))
-		regs = overlay->reg_bo->phys_obj->handle->vaddr;
+		regs = (struct overlay_registers __iomem *)overlay->reg_bo->phys_obj->handle->vaddr;
 	else
 		regs = io_mapping_map_atomic_wc(dev_priv->mm.gtt_mapping,
 						overlay->reg_bo->gtt_offset);
@@ -1515,7 +1529,7 @@ intel_overlay_map_regs_atomic(struct intel_overlay *overlay)
 }
 
 static void intel_overlay_unmap_regs_atomic(struct intel_overlay *overlay,
-					    struct overlay_registers *regs)
+					struct overlay_registers __iomem *regs)
 {
 	if (!OVERLAY_NEEDS_PHYSICAL(overlay->dev))
 		io_mapping_unmap_atomic(regs);
@@ -1540,9 +1554,9 @@ intel_overlay_capture_error_state(struct drm_device *dev)
 	error->dovsta = I915_READ(DOVSTA);
 	error->isr = I915_READ(ISR);
 	if (OVERLAY_NEEDS_PHYSICAL(overlay->dev))
-		error->base = (long) overlay->reg_bo->phys_obj->handle->vaddr;
+		error->base = (__force long)overlay->reg_bo->phys_obj->handle->vaddr;
 	else
-		error->base = (long) overlay->reg_bo->gtt_offset;
+		error->base = overlay->reg_bo->gtt_offset;
 
 	regs = intel_overlay_map_regs_atomic(overlay);
 	if (!regs)
Signature · 1 line
1.7.10
#9 lore raw

PATCH 5/9 drm/i915: [sparse] __iomem fixes for debugfs

In reply to Ben Widawsky

These were mostly straight forward. No forced casting needed.

---
 drivers/gpu/drm/i915/i915_debugfs.c |   26 +++++++++++++++++---------
 1 file changed, 17 insertions(+), 9 deletions(-)
drivers/gpu/drm/i915/i915_debugfs.c +17 -9
index 35462df..cfe3c82 100644
@@ -586,18 +586,18 @@ static int i915_hws_info(struct seq_file *m, void *data)
 	struct drm_device *dev = node->minor->dev;
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	struct intel_ring_buffer *ring;
-	const volatile u32 __iomem *hws;
+	void __iomem *hws;
 	int i;
 
 	ring = &dev_priv->ring[(uintptr_t)node->info_ent->data];
-	hws = (volatile u32 __iomem *)ring->status_page.page_addr;
+	hws = (void __iomem *)ring->status_page.page_addr;
 	if (hws == NULL)
 		return 0;
 
 	for (i = 0; i < 4096 / sizeof(u32) / 4; i += 4) {
 		seq_printf(m, "0x%08x: 0x%08x 0x%08x 0x%08x 0x%08x\n",
-			   i * 4,
-			   hws[i], hws[i + 1], hws[i + 2], hws[i + 3]);
+			   i * 4, ioread32(hws), ioread32(hws + 4),
+			   ioread32(hws + 8), ioread32(hws + 0xc));
 	}
 	return 0;
 }
@@ -622,8 +622,8 @@ static int i915_ringbuffer_data(struct seq_file *m, void *data)
 		uint32_t off;
 
 		for (off = 0; off < ring->size; off += 4) {
-			uint32_t *ptr = (uint32_t *)(virt + off);
-			seq_printf(m, "%08x :  %08x\n", off, *ptr);
+			uint32_t __iomem *ptr = (uint32_t __iomem *)(virt + off);
+			seq_printf(m, "%08x :  %08x\n", off, ioread32(ptr));
 		}
 	}
 	mutex_unlock(&dev->struct_mutex);
@@ -1353,17 +1353,25 @@ static int i915_opregion(struct seq_file *m, void *unused)
 	struct drm_device *dev = node->minor->dev;
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	struct intel_opregion *opregion = &dev_priv->opregion;
+	void *data = kmalloc(OPREGION_SIZE, GFP_KERNEL);
 	int ret;
 
+	if (data == NULL)
+		return -ENOMEM;
+
 	ret = mutex_lock_interruptible(&dev->struct_mutex);
 	if (ret)
-		return ret;
+		goto out;
 
-	if (opregion->header)
-		seq_write(m, opregion->header, OPREGION_SIZE);
+	if (opregion->header) {
+		memcpy_fromio(data, opregion->header, OPREGION_SIZE);
+		seq_write(m, data, OPREGION_SIZE);
+	}
 
 	mutex_unlock(&dev->struct_mutex);
 
+out:
+	kfree(data);
 	return 0;
 }
 
Signature · 1 line
1.7.10
#10 lore raw

PATCH 6/9 drm/i915: [sparse] __iomem fixes for ringbuffer

In reply to Ben Widawsky

Only one case of a "forced" casting where we use ring->map.handle and cast it to __iomem. As with other patches, this should be safe always.

---
 drivers/gpu/drm/i915/intel_ringbuffer.c |   16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)
drivers/gpu/drm/i915/intel_ringbuffer.c +9 -7
index caec699..62e11c5 100644
@@ -875,7 +875,7 @@ static int init_status_page(struct intel_ring_buffer *ring)
 		goto err_unpin;
 	}
 	ring->status_page.obj = obj;
-	memset(ring->status_page.page_addr, 0, PAGE_SIZE);
+	memset_io(ring->status_page.page_addr, 0, PAGE_SIZE);
 
 	intel_ring_setup_status_page(ring);
 	DRM_DEBUG_DRIVER("%s hws offset: 0x%08x\n",
@@ -938,7 +938,7 @@ static int intel_init_ring_buffer(struct drm_device *dev,
 		goto err_unpin;
 	}
 
-	ring->virtual_start = ring->map.handle;
+	ring->virtual_start = (void __iomem *)ring->map.handle;
 	ret = ring->init(ring);
 	if (ret)
 		goto err_unmap;
@@ -996,7 +996,7 @@ void intel_cleanup_ring_buffer(struct intel_ring_buffer *ring)
 
 static int intel_wrap_ring_buffer(struct intel_ring_buffer *ring)
 {
-	unsigned int *virt;
+	void __iomem *virt;
 	int rem = ring->size - ring->tail;
 
 	if (ring->space < rem) {
@@ -1005,11 +1005,13 @@ static int intel_wrap_ring_buffer(struct intel_ring_buffer *ring)
 			return ret;
 	}
 
-	virt = (unsigned int *)(ring->virtual_start + ring->tail);
+	virt = ring->virtual_start + ring->tail;
 	rem /= 8;
 	while (rem--) {
-		*virt++ = MI_NOOP;
-		*virt++ = MI_NOOP;
+		iowrite32(MI_NOOP, virt);
+		virt+=4;
+		iowrite32(MI_NOOP, virt);
+		virt+=4;
 	}
 
 	ring->tail = 0;
@@ -1308,7 +1310,7 @@ int intel_init_render_ring_buffer(struct drm_device *dev)
 
 	if (!I915_NEED_GFX_HWS(dev)) {
 		ring->status_page.page_addr = dev_priv->status_page_dmah->vaddr;
-		memset(ring->status_page.page_addr, 0, PAGE_SIZE);
+		memset_io(ring->status_page.page_addr, 0, PAGE_SIZE);
 	}
 
 	return intel_init_ring_buffer(dev, ring);
Signature · 1 line
1.7.10
#11 lore raw

Re: [PATCH 6/9] drm/i915: [sparse] __iomem fixes for ringbuffer

drivers/gpu/drm/i915/intel_ringbuffer.c static int intel_wrap_ring_buffer(struct intel_ring_buffer *ring) · line 1014

Quoted 8 lines · Ben Widawsky
drivers/gpu/drm/i915/intel_ringbuffer.c +4 -2
 	rem /= 8;
 	while (rem--) {
-		*virt++ = MI_NOOP;
-		*virt++ = MI_NOOP;
+		iowrite32(MI_NOOP, virt);
+		virt+=4;
+		iowrite32(MI_NOOP, virt);
+		virt+=4;

Might as well use memset_io().
-Chris

Signature · 1 line
Chris Wilson, Intel Open Source Technology Centre
#12 lore raw

PATCH 7/9 drm/i915: [sparse] forced __iomem ringbuffer fixes

In reply to Ben Widawsky

It should be safe to do this since GEN driver is only ever for x86, an __iomem and dma mem should be the same. However, these changes caused me to really think and try multiple solutions. Someone may come along and decide to do something better with these.

---
 drivers/gpu/drm/i915/intel_ringbuffer.c |    8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)
drivers/gpu/drm/i915/intel_ringbuffer.c +5 -3
index 62e11c5..8a0050f 100644
@@ -869,7 +869,7 @@ static int init_status_page(struct intel_ring_buffer *ring)
 	}
 
 	ring->status_page.gfx_addr = obj->gtt_offset;
-	ring->status_page.page_addr = kmap(obj->pages[0]);
+	ring->status_page.page_addr = (uint32_t __iomem *)kmap(obj->pages[0]);
 	if (ring->status_page.page_addr == NULL) {
 		memset(&dev_priv->hws_map, 0, sizeof(dev_priv->hws_map));
 		goto err_unpin;
@@ -1309,7 +1309,8 @@ int intel_init_render_ring_buffer(struct drm_device *dev)
 
 
 	if (!I915_NEED_GFX_HWS(dev)) {
-		ring->status_page.page_addr = dev_priv->status_page_dmah->vaddr;
+		ring->status_page.page_addr =
+			(void __iomem *) dev_priv->status_page_dmah->vaddr;
 		memset_io(ring->status_page.page_addr, 0, PAGE_SIZE);
 	}
 
@@ -1350,7 +1351,8 @@ int intel_render_ring_init_dri(struct drm_device *dev, u64 start, u32 size)
 	ring->cleanup = render_ring_cleanup;
 
 	if (!I915_NEED_GFX_HWS(dev))
-		ring->status_page.page_addr = dev_priv->status_page_dmah->vaddr;
+		ring->status_page.page_addr =
+			(void __iomem *) dev_priv->status_page_dmah->vaddr;
 
 	ring->dev = dev;
 	INIT_LIST_HEAD(&ring->active_list);
Signature · 1 line
1.7.10
#13 lore raw

PATCH 8/9 drm/i915: [sparse] __iomem fixes for gem

In reply to Ben Widawsky

As with one of the earlier patches in the series, we're forced to cast for copy_[to|from]_user. Again because of the nature of the GEN x86 exclusivity, this should be safe.

---
 drivers/gpu/drm/i915/i915_gem.c |   11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)
drivers/gpu/drm/i915/i915_gem.c +6 -5
index 9415c07..96527db 100644
@@ -266,8 +266,8 @@ __copy_to_user_swizzled(char __user *cpu_vaddr,
 }
 
 static inline int
-__copy_from_user_swizzled(char __user *gpu_vaddr, int gpu_offset,
-			  const char *cpu_vaddr,
+__copy_from_user_swizzled(char *gpu_vaddr, int gpu_offset,
+			  const char __user *cpu_vaddr,
 			  int length)
 {
 	int ret, cpu_offset = 0;
@@ -542,12 +542,13 @@ fast_user_write(struct io_mapping *mapping,
 		char __user *user_data,
 		int length)
 {
-	char *vaddr_atomic;
+	void __iomem *vaddr_atomic;
 	unsigned long unwritten;
 
 	vaddr_atomic = io_mapping_map_atomic_wc(mapping, page_base);
-	unwritten = __copy_from_user_inatomic_nocache(vaddr_atomic + page_offset,
-						      user_data, length);
+	unwritten = __copy_from_user_inatomic_nocache(
+				(void __force*)vaddr_atomic + page_offset,
+				user_data, length);
 	io_mapping_unmap_atomic(vaddr_atomic);
 	return unwritten;
 }
Signature · 1 line
1.7.10
#14 lore raw

PATCH 9/9 drm/i915: [sparse] __iomem fixes for intel_bios

In reply to Ben Widawsky

This is only the first part of the patch which does the correct types. The fix for dereferencing stuff was taking too long, so I've stopped.

---
 drivers/gpu/drm/i915/intel_bios.c |   20 ++++++++++++--------
 1 file changed, 12 insertions(+), 8 deletions(-)
drivers/gpu/drm/i915/intel_bios.c +12 -8
index 3534593..7233e72 100644
@@ -697,24 +697,27 @@ intel_parse_bios(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct pci_dev *pdev = dev->pdev;
-	struct bdb_header *bdb = NULL;
+	struct bdb_header __iomem *bdb = NULL;
 	u8 __iomem *bios = NULL;
 
 	init_vbt_defaults(dev_priv);
 
 	/* XXX Should this validation be moved to intel_opregion.c? */
 	if (!dmi_check_system(intel_no_opregion_vbt) && dev_priv->opregion.vbt) {
-		struct vbt_header *vbt = dev_priv->opregion.vbt;
-		if (memcmp(vbt->signature, "$VBT", 4) == 0) {
+		char sig[20];
+		struct vbt_header __iomem *vbt = dev_priv->opregion.vbt;
+		memcpy_fromio(sig, &vbt->signature, sizeof(sig));
+		if (memcmp(sig, "$VBT", 4) == 0) {
 			DRM_DEBUG_KMS("Using VBT from OpRegion: %20s\n",
 					 vbt->signature);
-			bdb = (struct bdb_header *)((char *)vbt + vbt->bdb_offset);
+			bdb = (struct bdb_header __iomem *)((char __iomem *)vbt +
+				ioread32(&vbt->bdb_offset));
 		} else
 			dev_priv->opregion.vbt = NULL;
 	}
 
 	if (bdb == NULL) {
-		struct vbt_header *vbt = NULL;
+		struct vbt_header __iomem *vbt = NULL;
 		size_t size;
 		int i;
 
@@ -724,8 +727,9 @@ intel_parse_bios(struct drm_device *dev)
 
 		/* Scour memory looking for the VBT signature */
 		for (i = 0; i + 4 < size; i++) {
-			if (!memcmp(bios + i, "$VBT", 4)) {
-				vbt = (struct vbt_header *)(bios + i);
+			u32 temp = ioread32(bios + i);
+			if (!memcmp(&temp, "$VBT", 4)) {
+				vbt = (struct vbt_header __iomem *)(bios + i);
 				break;
 			}
 		}
@@ -736,7 +740,7 @@ intel_parse_bios(struct drm_device *dev)
 			return -1;
 		}
 
-		bdb = (struct bdb_header *)(bios + i + vbt->bdb_offset);
+		bdb = (struct bdb_header __iomem *)(bios + i + vbt->bdb_offset);
 	}
 
 	/* Grab useful general definitions */
Signature · 1 line
1.7.10
#15 lore raw
In reply to Ben Widawsky
Quoted 13 lines · Ben Widawsky

Patch 1/2 is a mix of fairly trivial stuff. I see no reason to not take these.
Patch 3-9 Are all __iomem related. These should run through some regression testing.

Most of the __iomem stuff fell into place fairly easily, but it's definitely subject to a lot of copy/paste errors, since trying to automate the replacement didn't workout well. Throughout the patches, I tried really hard to not force cast things, but a couple of situations cropped up which gave me no choice. What I hope to achieve by that is there should be no functional or performance changes (aside from copy/paste).

Quoted 3 lines

I've given up on fixing intel_bios.c completely.

These are based off of dinq as of yesterday.

Ok, after some lengthy irc bikeshedding with Ben and Chris I've picked up a few more of these, but changed a few things:
- Any hunk that touched the status_page got dropped. I've looked around and I think I have a plan to fix that disaster for real.
- Added comments about pointer casts due to sparse.
- Dropped the intel_bios.c patch, we agreed that to be too horrible. This needs some serious care to be sparse-compliant.

Thanks for digging through this mess.
-Daniel

Quoted 35 lines

Ben Widawsky (9):
drm/i915: [sparse] trivial sparse fixes drm/i915: [sparse] don't use variable size arrays drm/i915: [sparse] __iomem fixes for opregion drm/i915: [sparse] __iomem fixes for overlay drm/i915: [sparse] __iomem fixes for debugfs drm/i915: [sparse] __iomem fixes for ringbuffer drm/i915: [sparse] forced __iomem ringbuffer fixes drm/i915: [sparse] __iomem fixes for gem drm/i915: [sparse] __iomem fixes for intel_bios

 drivers/gpu/drm/i915/i915_debugfs.c      |   28 ++++---
 drivers/gpu/drm/i915/i915_drv.h          |   12 +--
 drivers/gpu/drm/i915/i915_gem.c          |   11 +--
 drivers/gpu/drm/i915/i915_ioc32.c        |    5 +-
 drivers/gpu/drm/i915/i915_trace_points.c |    2 +
 drivers/gpu/drm/i915/intel_acpi.c        |    1 +
 drivers/gpu/drm/i915/intel_bios.c        |   20 +++--
 drivers/gpu/drm/i915/intel_display.c     |    2 +-
 drivers/gpu/drm/i915/intel_fb.c          |    2 +-
 drivers/gpu/drm/i915/intel_opregion.c    |   67 ++++++++-------
 drivers/gpu/drm/i915/intel_overlay.c     |  130 +++++++++++++++++-------------
 drivers/gpu/drm/i915/intel_ringbuffer.c  |   28 ++++---
 drivers/gpu/drm/i915/intel_sdvo.c        |   26 ++++--
 13 files changed, 197 insertions(+), 137 deletions(-)
Signature · 6 lines
1.7.10

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 3 lines
Daniel Vetter
Mail: [email protected]
Mobile: +41 (0)79 365 57 48