Commit 14bac5acfdb6a40be64acc042c6db73f1a68f6a4

Authored by Nick Piggin
Committed by Linus Torvalds
1 parent 538f8ea6c8

mm: xip/ext2 fix block allocation race

XIP can call into get_xip_mem concurrently with the same file,offset with
create=1.  This usually maps down to get_block, which expects the page
lock to prevent such a situation.  This causes ext2 to explode for one
reason or another.

Serialise those calls for the moment.  For common usages today, I suspect
get_xip_mem rarely is called to create new blocks.  In future as XIP
technologies evolve we might need to look at which operations require
scalability, and rework the locking to suit.

Signed-off-by: Nick Piggin <npiggin@suse.de>
Cc: Jared Hulbert <jaredeh@gmail.com>
Acked-by: Carsten Otte <cotte@freenet.de>
Cc: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Showing 1 changed file with 4 additions and 1 deletions Side-by-side Diff

... ... @@ -248,15 +248,16 @@
248 248 int err;
249 249  
250 250 /* maybe shared writable, allocate new block */
  251 + mutex_lock(&xip_sparse_mutex);
251 252 error = mapping->a_ops->get_xip_mem(mapping, vmf->pgoff, 1,
252 253 &xip_mem, &xip_pfn);
  254 + mutex_unlock(&xip_sparse_mutex);
253 255 if (error)
254 256 return VM_FAULT_SIGBUS;
255 257 /* unmap sparse mappings at pgoff from all other vmas */
256 258 __xip_unmap(mapping, vmf->pgoff);
257 259  
258 260 found:
259   - printk("%s insert %lx@%lx\n", current->comm, (unsigned long)vmf->virtual_address, xip_pfn);
260 261 err = vm_insert_mixed(vma, (unsigned long)vmf->virtual_address,
261 262 xip_pfn);
262 263 if (err == -ENOMEM)
263 264  
... ... @@ -340,8 +341,10 @@
340 341 &xip_mem, &xip_pfn);
341 342 if (status == -ENODATA) {
342 343 /* we allocate a new page unmap it */
  344 + mutex_lock(&xip_sparse_mutex);
343 345 status = a_ops->get_xip_mem(mapping, index, 1,
344 346 &xip_mem, &xip_pfn);
  347 + mutex_unlock(&xip_sparse_mutex);
345 348 if (!status)
346 349 /* unmap page at pgoff from all other vmas */
347 350 __xip_unmap(mapping, index);