Skip to contentSkip to conversation

ppc highmem fix

PATCH In review Paolo Galtieri 2 participants 3 messages Oct 10, 2005 – Oct 11, 2005

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ppc highmem fix

I've noticed that the calculations for seg_size and nr_segs in __dma_sync_page_highmem() (arch/ppc/kernel/dma-mapping.c) are wrong. The incorrect calculations can result in either an oops or a panic when running fsck depending on the size of the partition. The problem with the
seg_size calculation is that it can result in a negative number if size is offset > size. The problem with the nr_segs caculation is returns the
wrong number of segments, e.g. it returns 1 when size is 200 and offset is 4095, when it should return 2 or more.

Here is the patch to fix the problem.

--- linux-2.6.14-rc3-git8/arch/ppc/kernel/dma-mapping.c 2005-10-10 13:55:37.000000000 -0700
+++ linux-2.6.14/arch/ppc/kernel/dma-mapping.c 2005-10-10 13:57:51.000000000 -0700
@@ -401,10 +401,10 @@
static inline void __dma_sync_page_highmem(struct page *page,
unsigned long offset, size_t size, int direction)
{
- size_t seg_size = min((size_t)PAGE_SIZE, size) - offset;

+       size_t seg_size = min((size_t)(PAGE_SIZE - offset), size);
        size_t cur_size = seg_size;
        unsigned long flags, start, seg_offset = offset;
-       int nr_segs = PAGE_ALIGN(size + (PAGE_SIZE - offset))/PAGE_SIZE;
+       int nr_segs = 1 + ((size - seg_size) + PAGE_SIZE - 1)/PAGE_SIZE;
        int seg_nr = 0;

        local_irq_save(flags);

Paolo Galtieri ([email protected])

#1 lore raw
Quoted 32 lines · Paolo Galtieri

I've noticed that the calculations for seg_size and nr_segs in __dma_sync_page_highmem() (arch/ppc/kernel/dma-mapping.c) are wrong. The incorrect calculations can result in either an oops or a panic when running fsck depending on the size of the partition. The problem with the
seg_size calculation is that it can result in a negative number if size is offset > size. The problem with the nr_segs caculation is returns the
wrong number of segments, e.g. it returns 1 when size is 200 and offset is 4095, when it should return 2 or more.

Here is the patch to fix the problem.

--- linux-2.6.14-rc3-git8/arch/ppc/kernel/dma-mapping.c 2005-10-10 13:55:37.000000000 -0700
+++ linux-2.6.14/arch/ppc/kernel/dma-mapping.c 2005-10-10 13:57:51.000000000 -0700

(quoted diff) +2 -2
@@ -401,10 +401,10 @@
 static inline void __dma_sync_page_highmem(struct page *page,
                unsigned long offset, size_t size, int direction)
 {
-       size_t seg_size = min((size_t)PAGE_SIZE, size) - offset;
+       size_t seg_size = min((size_t)(PAGE_SIZE - offset), size);
        size_t cur_size = seg_size;
        unsigned long flags, start, seg_offset = offset;
-       int nr_segs = PAGE_ALIGN(size + (PAGE_SIZE - offset))/PAGE_SIZE;
+       int nr_segs = 1 + ((size - seg_size) + PAGE_SIZE - 1)/PAGE_SIZE;
        int seg_nr = 0;
 
        local_irq_save(flags);
Quoted 2 lines

Paolo Galtieri ([email protected])

I want to apologize in advance for the previous post. When I included the patch inline I dropped the tabs :-(. Here is the correct patch with tabs:

--- linux-2.6.14-rc3-git8/arch/ppc/kernel/dma-mapping.c 2005-10-10
13:55:37.000000000 -0700
+++ linux-2.6.14/arch/ppc/kernel/dma-mapping.c 2005-10-10
13:57:51.000000000 -0700
@@ -401,10 +401,10 @@
static inline void __dma_sync_page_highmem(struct page *page,
unsigned long offset, size_t size, int direction)
{
- size_t seg_size = min((size_t)PAGE_SIZE, size) - offset;
+ size_t seg_size = min((size_t)(PAGE_SIZE - offset), size);
size_t cur_size = seg_size;
unsigned long flags, start, seg_offset = offset;
- int nr_segs = PAGE_ALIGN(size + (PAGE_SIZE - offset))/PAGE_SIZE;
+ int nr_segs = 1 + ((size - seg_size) + PAGE_SIZE - 1)/PAGE_SIZE;
int seg_nr = 0;

local_irq_save(flags);

Paolo

#2 lore raw
Quoted 12 lines · Paolo Galtieri
Quoted 11 lines · Paolo Galtieri

I've noticed that the calculations for seg_size and nr_segs in __dma_sync_page_highmem() (arch/ppc/kernel/dma-mapping.c) are wrong. The incorrect calculations can result in either an oops or a panic when running fsck depending on the size of the partition. The problem with the
seg_size calculation is that it can result in a negative number if size is offset > size. The problem with the nr_segs caculation is returns the
wrong number of segments, e.g. it returns 1 when size is 200 and offset is 4095, when it should return 2 or more.

Quoted 1 line
Quoted 1 line

Here is the patch to fix the problem.

Looks good, should probably still make it to 2.6.14, Andrew, Linus ?

(Paolo, can you send them a clean patch ?)

Quoted 51 lines
Quoted 19 lines

--- linux-2.6.14-rc3-git8/arch/ppc/kernel/dma-mapping.c 2005-10-10 13:55:37.000000000 -0700
+++ linux-2.6.14/arch/ppc/kernel/dma-mapping.c 2005-10-10 13:57:51.000000000 -0700

(quoted diff) +2 -2
@@ -401,10 +401,10 @@
 static inline void __dma_sync_page_highmem(struct page *page,
                unsigned long offset, size_t size, int direction)
 {
-       size_t seg_size = min((size_t)PAGE_SIZE, size) - offset;
+       size_t seg_size = min((size_t)(PAGE_SIZE - offset), size);
        size_t cur_size = seg_size;
        unsigned long flags, start, seg_offset = offset;
-       int nr_segs = PAGE_ALIGN(size + (PAGE_SIZE - offset))/PAGE_SIZE;
+       int nr_segs = 1 + ((size - seg_size) + PAGE_SIZE - 1)/PAGE_SIZE;
        int seg_nr = 0;
 
        local_irq_save(flags);
Quoted 2 lines

Paolo Galtieri ([email protected])

I want to apologize in advance for the previous post. When I included the patch inline I dropped the tabs :-(. Here is the correct patch with tabs:

--- linux-2.6.14-rc3-git8/arch/ppc/kernel/dma-mapping.c 2005-10-10
13:55:37.000000000 -0700
+++ linux-2.6.14/arch/ppc/kernel/dma-mapping.c 2005-10-10
13:57:51.000000000 -0700

(quoted diff) +2 -2
@@ -401,10 +401,10 @@
 static inline void __dma_sync_page_highmem(struct page *page,
 		unsigned long offset, size_t size, int direction)
 {
-	size_t seg_size = min((size_t)PAGE_SIZE, size) - offset;
+	size_t seg_size = min((size_t)(PAGE_SIZE - offset), size);
 	size_t cur_size = seg_size;
 	unsigned long flags, start, seg_offset = offset;
-	int nr_segs = PAGE_ALIGN(size + (PAGE_SIZE - offset))/PAGE_SIZE;
+	int nr_segs = 1 + ((size - seg_size) + PAGE_SIZE - 1)/PAGE_SIZE;
 	int seg_nr = 0;
 
 	local_irq_save(flags);

Paolo

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/