diff mbox series

[v6,23/36] drm: vmwgfx: fix common struct sg_table related issues

Message ID 20200618153956.29558-24-m.szyprowski@samsung.com
State Superseded
Headers show
Series DRM: fix struct sg_table nents vs. orig_nents misuse | expand

Commit Message

Marek Szyprowski June 18, 2020, 3:39 p.m. UTC 
The Documentation/DMA-API-HOWTO.txt states that the dma_map_sg() function
returns the number of the created entries in the DMA address space.
However the subsequent calls to the dma_sync_sg_for_{device,cpu}() and
dma_unmap_sg must be called with the original number of the entries
passed to the dma_map_sg().

struct sg_table is a common structure used for describing a non-contiguous
memory buffer, used commonly in the DRM and graphics subsystems. It
consists of a scatterlist with memory pages and DMA addresses (sgl entry),
as well as the number of scatterlist entries: CPU pages (orig_nents entry)
and DMA mapped pages (nents entry).

It turned out that it was a common mistake to misuse nents and orig_nents
entries, calling DMA-mapping functions with a wrong number of entries or
ignoring the number of mapped entries returned by the dma_map_sg()
function.

To avoid such issues, lets use a common dma-mapping wrappers operating
directly on the struct sg_table objects and use scatterlist page
iterators where possible. This, almost always, hides references to the
nents and orig_nents entries, making the code robust, easier to follow
and copy/paste safe.

Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>

Acked-by: Roland Scheidegger <sroland@vmware.com>

---
 drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c | 17 ++++-------------
 1 file changed, 4 insertions(+), 13 deletions(-)

-- 
2.17.1

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Comments

kernel test robot June 19, 2020, 3:05 a.m. UTC | #1 
Hi Marek,

I love your patch! Perhaps something to improve:

[auto build test WARNING on next-20200618]
[also build test WARNING on v5.8-rc1]
[cannot apply to linuxtv-media/master staging/staging-testing drm-exynos/exynos-drm-next drm-intel/for-linux-next linus/master v5.8-rc1 v5.7 v5.7-rc7]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use  as documented in
https://git-scm.com/docs/git-format-patch]

url:    https://github.com/0day-ci/linux/commits/Marek-Szyprowski/DRM-fix-struct-sg_table-nents-vs-orig_nents-misuse/20200619-000417
base:    ce2cc8efd7a40cbd17841add878cb691d0ce0bba
config: x86_64-randconfig-s021-20200618 (attached as .config)
compiler: gcc-9 (Debian 9.3.0-13) 9.3.0
reproduce:
        # apt-get install sparse
        # sparse version: v0.6.2-rc1-10-gc17b1b06-dirty
        # save the attached .config to linux build tree
        make W=1 C=1 ARCH=x86_64 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__'

If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>


sparse warnings: (new ones prefixed by >>)

>> drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c:365:38: sparse: sparse: incorrect type in argument 2 (different base types) @@     expected struct sg_table *sgt @@     got struct sg_table sgt @@

>> drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c:365:38: sparse:     expected struct sg_table *sgt

>> drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c:365:38: sparse:     got struct sg_table sgt

   drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c:386:43: sparse: sparse: incorrect type in argument 2 (different base types) @@     expected struct sg_table *sgt @@     got struct sg_table sgt @@
   drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c:386:43: sparse:     expected struct sg_table *sgt
   drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c:386:43: sparse:     got struct sg_table sgt

vim +365 drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c

   352	
   353	/**
   354	 * vmw_ttm_unmap_from_dma - unmap  device addresses previsouly mapped for
   355	 * TTM pages
   356	 *
   357	 * @vmw_tt: Pointer to a struct vmw_ttm_backend
   358	 *
   359	 * Used to free dma mappings previously mapped by vmw_ttm_map_for_dma.
   360	 */
   361	static void vmw_ttm_unmap_from_dma(struct vmw_ttm_tt *vmw_tt)
   362	{
   363		struct device *dev = vmw_tt->dev_priv->dev->dev;
   364	
 > 365		dma_unmap_sgtable(dev, vmw_tt->sgt, DMA_BIDIRECTIONAL, 0);

   366		vmw_tt->sgt.nents = vmw_tt->sgt.orig_nents;
   367	}
   368	

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
kernel test robot June 19, 2020, 4:59 a.m. UTC | #2 
Hi Marek,

I love your patch! Yet something to improve:

[auto build test ERROR on next-20200618]
[also build test ERROR on v5.8-rc1]
[cannot apply to linuxtv-media/master staging/staging-testing drm-exynos/exynos-drm-next drm-intel/for-linux-next linus/master v5.8-rc1 v5.7 v5.7-rc7]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use  as documented in
https://git-scm.com/docs/git-format-patch]

url:    https://github.com/0day-ci/linux/commits/Marek-Szyprowski/DRM-fix-struct-sg_table-nents-vs-orig_nents-misuse/20200619-000417
base:    ce2cc8efd7a40cbd17841add878cb691d0ce0bba
config: x86_64-rhel-7.6 (attached as .config)
compiler: gcc-9 (Debian 9.3.0-13) 9.3.0
reproduce (this is a W=1 build):
        # save the attached .config to linux build tree
        make W=1 ARCH=x86_64 

If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>

All error/warnings (new ones prefixed by >>):

   drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c: In function 'vmw_ttm_unmap_from_dma':
>> drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c:365:31: error: incompatible type for argument 2 of 'dma_unmap_sgtable'

     365 |  dma_unmap_sgtable(dev, vmw_tt->sgt, DMA_BIDIRECTIONAL, 0);
         |                         ~~~~~~^~~~~
         |                               |
         |                               struct sg_table
   In file included from include/linux/dma-buf.h:20,
                    from drivers/gpu/drm/vmwgfx/ttm_object.h:40,
                    from drivers/gpu/drm/vmwgfx/ttm_lock.h:55,
                    from drivers/gpu/drm/vmwgfx/vmwgfx_drv.h:44,
                    from drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c:28:
   include/linux/dma-mapping.h:651:75: note: expected 'struct sg_table *' but argument is of type 'struct sg_table'
     651 | static inline void dma_unmap_sgtable(struct device *dev, struct sg_table *sgt,
         |                                                          ~~~~~~~~~~~~~~~~~^~~
   drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c: In function 'vmw_ttm_map_for_dma':
>> drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c:386:36: error: incompatible type for argument 2 of 'dma_map_sgtable'

     386 |  return dma_map_sgtable(dev, vmw_tt->sgt, DMA_BIDIRECTIONAL, 0);
         |                              ~~~~~~^~~~~
         |                                    |
         |                                    struct sg_table
   In file included from include/linux/dma-buf.h:20,
                    from drivers/gpu/drm/vmwgfx/ttm_object.h:40,
                    from drivers/gpu/drm/vmwgfx/ttm_lock.h:55,
                    from drivers/gpu/drm/vmwgfx/vmwgfx_drv.h:44,
                    from drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c:28:
   include/linux/dma-mapping.h:628:72: note: expected 'struct sg_table *' but argument is of type 'struct sg_table'
     628 | static inline int dma_map_sgtable(struct device *dev, struct sg_table *sgt,
         |                                                       ~~~~~~~~~~~~~~~~~^~~
>> drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c:387:1: warning: control reaches end of non-void function [-Wreturn-type]

     387 | }
         | ^

vim +/dma_unmap_sgtable +365 drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c

   352	
   353	/**
   354	 * vmw_ttm_unmap_from_dma - unmap  device addresses previsouly mapped for
   355	 * TTM pages
   356	 *
   357	 * @vmw_tt: Pointer to a struct vmw_ttm_backend
   358	 *
   359	 * Used to free dma mappings previously mapped by vmw_ttm_map_for_dma.
   360	 */
   361	static void vmw_ttm_unmap_from_dma(struct vmw_ttm_tt *vmw_tt)
   362	{
   363		struct device *dev = vmw_tt->dev_priv->dev->dev;
   364	
 > 365		dma_unmap_sgtable(dev, vmw_tt->sgt, DMA_BIDIRECTIONAL, 0);

   366		vmw_tt->sgt.nents = vmw_tt->sgt.orig_nents;
   367	}
   368	
   369	/**
   370	 * vmw_ttm_map_for_dma - map TTM pages to get device addresses
   371	 *
   372	 * @vmw_tt: Pointer to a struct vmw_ttm_backend
   373	 *
   374	 * This function is used to get device addresses from the kernel DMA layer.
   375	 * However, it's violating the DMA API in that when this operation has been
   376	 * performed, it's illegal for the CPU to write to the pages without first
   377	 * unmapping the DMA mappings, or calling dma_sync_sg_for_cpu(). It is
   378	 * therefore only legal to call this function if we know that the function
   379	 * dma_sync_sg_for_cpu() is a NOP, and dma_sync_sg_for_device() is at most
   380	 * a CPU write buffer flush.
   381	 */
   382	static int vmw_ttm_map_for_dma(struct vmw_ttm_tt *vmw_tt)
   383	{
   384		struct device *dev = vmw_tt->dev_priv->dev->dev;
   385	
 > 386		return dma_map_sgtable(dev, vmw_tt->sgt, DMA_BIDIRECTIONAL, 0);

 > 387	}

   388	

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
diff mbox series

Patch

diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
index bf0bc4697959..49ed6add6969 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
@@ -362,8 +362,7 @@  static void vmw_ttm_unmap_from_dma(struct vmw_ttm_tt *vmw_tt)
 {
 	struct device *dev = vmw_tt->dev_priv->dev->dev;
 
-	dma_unmap_sg(dev, vmw_tt->sgt.sgl, vmw_tt->sgt.nents,
-		DMA_BIDIRECTIONAL);
+	dma_unmap_sgtable(dev, vmw_tt->sgt, DMA_BIDIRECTIONAL, 0);
 	vmw_tt->sgt.nents = vmw_tt->sgt.orig_nents;
 }
 
@@ -383,16 +382,8 @@  static void vmw_ttm_unmap_from_dma(struct vmw_ttm_tt *vmw_tt)
 static int vmw_ttm_map_for_dma(struct vmw_ttm_tt *vmw_tt)
 {
 	struct device *dev = vmw_tt->dev_priv->dev->dev;
-	int ret;
-
-	ret = dma_map_sg(dev, vmw_tt->sgt.sgl, vmw_tt->sgt.orig_nents,
-			 DMA_BIDIRECTIONAL);
-	if (unlikely(ret == 0))
-		return -ENOMEM;
 
-	vmw_tt->sgt.nents = ret;
-
-	return 0;
+	return dma_map_sgtable(dev, vmw_tt->sgt, DMA_BIDIRECTIONAL, 0);
 }
 
 /**
@@ -449,10 +440,10 @@  static int vmw_ttm_map_dma(struct vmw_ttm_tt *vmw_tt)
 		if (unlikely(ret != 0))
 			goto out_sg_alloc_fail;
 
-		if (vsgt->num_pages > vmw_tt->sgt.nents) {
+		if (vsgt->num_pages > vmw_tt->sgt.orig_nents) {
 			uint64_t over_alloc =
 				sgl_size * (vsgt->num_pages -
-					    vmw_tt->sgt.nents);
+					    vmw_tt->sgt.orig_nents);
 
 			ttm_mem_global_free(glob, over_alloc);
 			vmw_tt->sg_alloc_size -= over_alloc;