Skip to contentSkip to conversation

isofs compress: Remove VLA usage

PATCH Superseded Kyle Spiers 4 participants 6 messages Apr 10, 2018 – Apr 11, 2018 Isofs filesystem

v3 of 4 Superseded Latest is v4 →

lore
Share & tools
lore
#0 lore raw

PATCH v3 1/1 isofs compress: Remove VLA usage

As part of the effort to remove VLAs from the kernel[1], this changes the allocation of the bhs and pages arrays from being on the stack to being kcalloc()ed. This also allows for the removal of the explicit zeroing of bhs.

https://lkml.org/lkml/2018/3/7/621

---
Fixed build for sparc64
Fixed error recovery mistakes

---
 fs/isofs/compress.c | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)
fs/isofs/compress.c +12 -3
index 9bb2fe35799d..4eba16bf173c 100644
@@ -20,6 +20,7 @@
 #include <linux/init.h>
 #include <linux/bio.h>
 
+#include <linux/slab.h>
 #include <linux/vmalloc.h>
 #include <linux/zlib.h>
 
@@ -59,7 +60,7 @@ static loff_t zisofs_uncompress_block(struct inode *inode, loff_t block_start,
 				>> bufshift;
 	int haveblocks;
 	blkcnt_t blocknum;
-	struct buffer_head *bhs[needblocks + 1];
+	struct buffer_head **bhs;
 	int curbh, curpage;
 
 	if (block_size > deflateBound(1UL << zisofs_block_shift)) {
@@ -80,7 +81,9 @@ static loff_t zisofs_uncompress_block(struct inode *inode, loff_t block_start,
 
 	/* Because zlib is not thread-safe, do all the I/O at the top. */
 	blocknum = block_start >> bufshift;
-	memset(bhs, 0, (needblocks + 1) * sizeof(struct buffer_head *));
+	bhs = kcalloc(needblocks + 1, sizeof(*bhs), GFP_KERNEL);
+	if (!bhs)
+		return -ENOMEM;
 	haveblocks = isofs_get_blocks(inode, blocknum, bhs, needblocks);
 	ll_rw_block(REQ_OP_READ, 0, haveblocks, bhs);
 
@@ -190,6 +193,7 @@ static loff_t zisofs_uncompress_block(struct inode *inode, loff_t block_start,
 b_eio:
 	for (i = 0; i < haveblocks; i++)
 		brelse(bhs[i]);
+	kfree(bhs);
 	return stream.total_out;
 }
 
@@ -305,7 +309,7 @@ static int zisofs_readpage(struct file *file, struct page *page)
 	unsigned int zisofs_pages_per_cblock =
 		PAGE_SHIFT <= zisofs_block_shift ?
 		(1 << (zisofs_block_shift - PAGE_SHIFT)) : 0;
-	struct page *pages[max_t(unsigned, zisofs_pages_per_cblock, 1)];
+	struct page **pages;
 	pgoff_t index = page->index, end_index;
 
 	end_index = (inode->i_size + PAGE_SIZE - 1) >> PAGE_SHIFT;
@@ -330,6 +334,10 @@ static int zisofs_readpage(struct file *file, struct page *page)
 		full_page = 0;
 		pcount = 1;
 	}
+	pages = kcalloc(max_t(unsigned int, zisofs_pages_per_cblock, 1),
+					sizeof(*pages), GFP_KERNEL);
+	if (!pages)
+		return -ENOMEM;
 	pages[full_page] = page;
 
 	for (i = 0; i < pcount; i++, index++) {
@@ -357,6 +365,7 @@ static int zisofs_readpage(struct file *file, struct page *page)
 	}			
 
 	/* At this point, err contains 0 or -EIO depending on the "critical" page */
+	kfree(pages);
 	return err;
 }
 
Signature · 1 line
2.17.0.484.g0c8726318c-goog
#1 lore raw
Quoted 6 lines · Kyle Spiers

As part of the effort to remove VLAs from the kernel[1], this changes the allocation of the bhs and pages arrays from being on the stack to being kcalloc()ed. This also allows for the removal of the explicit zeroing of bhs.

https://lkml.org/lkml/2018/3/7/621

Do you even bother reading the feedback given to such patches? I'm just curious,

fs/isofs/compress.c static loff_t zisofs_uncompress_block(struct inode *inode, loff_t block_start, · line 86

Quoted 8 lines
fs/isofs/compress.c +3 -1
@@ -80,7 +81,9 @@ static loff_t zisofs_uncompress_block(struct inode *inode, loff_t block_start,
 
 	/* Because zlib is not thread-safe, do all the I/O at the top. */
 	blocknum = block_start >> bufshift;
-	memset(bhs, 0, (needblocks + 1) * sizeof(struct buffer_head *));
+	bhs = kcalloc(needblocks + 1, sizeof(*bhs), GFP_KERNEL);
+	if (!bhs)
+		return -ENOMEM;

... because I distinctly remember comments along the lines of "check what other failure exits look like in there, such as

                *errp = -EIO;
                return 0;

nearby".

fs/isofs/compress.c static int zisofs_readpage(struct file *file, struct page *page) · line 341

Quoted 9 lines
fs/isofs/compress.c +4 -0
@@ -330,6 +334,10 @@ static int zisofs_readpage(struct file *file, struct page *page)
 		full_page = 0;
 		pcount = 1;
 	}
+	pages = kcalloc(max_t(unsigned int, zisofs_pages_per_cblock, 1),
+					sizeof(*pages), GFP_KERNEL);
+	if (!pages)
+		return -ENOMEM;
 	pages[full_page] = page;

What, in your opinion, is going to unlock that page after that failure exit?

#2 lore raw
Quoted 10 lines · Al Viro
Quoted 6 lines · Kyle Spiers

As part of the effort to remove VLAs from the kernel[1], this changes the allocation of the bhs and pages arrays from being on the stack to being kcalloc()ed. This also allows for the removal of the explicit zeroing of bhs.

https://lkml.org/lkml/2018/3/7/621

Do you even bother reading the feedback given to such patches? I'm just curious,

He made a tool mistake (sent v2 again) -- please see v4.

-Kees

Signature · 2 lines
Kees Cook
Pixel Security
#3 lore raw
Quoted 7 lines · Kees Cook
Quoted 6 lines · Al Viro
Quoted 5 lines · Kyle Spiers

As part of the effort to remove VLAs from the kernel[1], this changes the allocation of the bhs and pages arrays from being on the stack to being kcalloc()ed. This also allows for the removal of the explicit zeroing of bhs.

Quoted 6 lines
Quoted 4 lines
Quoted 1 line

https://lkml.org/lkml/2018/3/7/621

Do you even bother reading the feedback given to such patches? I'm just curious,

He made a tool mistake (sent v2 again) -- please see v4.

No worries but Kyle didn't just mistakenly resend v2. It was a v3 with a different below the dashes changelog.

Kyle, it would be good if next time you were to add a listing of changes in each revision below the dashes.

#4 lore raw
Quoted 13 lines · Joe Perches
Quoted 12 lines · Kees Cook
Quoted 10 lines · Al Viro
Quoted 6 lines · Kyle Spiers

As part of the effort to remove VLAs from the kernel[1], this changes the allocation of the bhs and pages arrays from being on the stack to being kcalloc()ed. This also allows for the removal of the explicit zeroing of bhs.

https://lkml.org/lkml/2018/3/7/621

Do you even bother reading the feedback given to such patches? I'm just curious,

Quoted 7 lines
Quoted 1 line

He made a tool mistake (sent v2 again) -- please see v4.

No worries but Kyle didn't just mistakenly resend v2. It was a v3 with a different below the dashes changelog.

Kyle, it would be good if next time you were to add a listing of changes in each revision below the dashes.

Shit happens...

Al, fully expecting to discover having made something spectacularly bonehead 2 minutes after sending a patch out - happens often enough.

#5 lore raw
Quoted 4 lines · Al Viro

Shit happens...

Al, fully expecting to discover having made something spectacularly bonehead 2 minutes after sending a patch out - happens often enough.

I've done that. Again, no worries.