Skip to contentSkip to conversation

2.6.20-rc6-mm3 and NTFS: BUG: at arch/i386/mm/highmem.c:52 kmap_atomic()

DISCUSSION Robert Hancock 2 participants 2 messages Feb 6, 2007 NTFS filesystem

lore
Share & tools
lore
#0 lore raw

Seeing these BUGs on 2.6.20-rc6-mm3 when mounting an NTFS partition. I saw some reports of something like this on -mm1, was this supposed to be patched already?

BUG: at arch/i386/mm/highmem.c:52 kmap_atomic()
 [<c041f810>] kmap_atomic+0xb4/0x1cd
 [<e0de8276>] ntfs_end_buffer_async_read+0x276/0x2db [ntfs]
 [<c048f92b>] end_bio_bh_io_sync+0x0/0x39
 [<c048f959>] end_bio_bh_io_sync+0x2e/0x39
 [<c049138f>] bio_endio+0x5b/0x63
 [<c05ffe97>] _spin_lock_irqsave+0x9/0xd
 [<c042f3f6>] lock_timer_base+0x15/0x2f
 [<c04d9f03>] __end_that_request_first+0x176/0x421
 [<e08c45a6>] scsi_delete_timer+0xf/0x50 [scsi_mod]
 [<e08c6817>] scsi_request_fn+0x2d4/0x324 [scsi_mod]
 [<e092ee7b>] ata_scsi_qc_complete+0x344/0x354 [libata]
 [<c04dba71>] __blk_run_queue+0x17/0x35
 [<e08c5601>] scsi_end_request+0x1a/0xa8 [scsi_mod]
 [<e08c5826>] scsi_io_completion+0x15a/0x32b [scsi_mod]
 [<c0409f81>] pit_next_event+0x2b/0x2f
 [<e08505c9>] sd_rw_intr+0x21b/0x245 [sd_mod]
 [<e08c13ff>] scsi_finish_command+0x84/0x8b [scsi_mod]
 [<c04dc19b>] blk_done_softirq+0x49/0x54
 [<c042b645>] __do_softirq+0x5d/0xba
 [<c042b6d4>] do_softirq+0x32/0x36
 [<c042b926>] irq_exit+0x38/0x6b
 [<c0406eb7>] do_IRQ+0x87/0x9c
 [<c0405837>] common_interrupt+0x23/0x28
 [<c043007b>] __dequeue_signal+0x97/0x14e
 [<c052709f>] acpi_processor_idle+0x209/0x3ca
 [<c0526e96>] acpi_processor_idle+0x0/0x3ca
 [<c0403410>] cpu_idle+0xaa/0xd0
 [<c071ca1f>] start_kernel+0x435/0x43d
 [<c071c1ae>] unknown_bootoption+0x0/0x202
 =======================
Signature · 3 lines
Robert Hancock      Saskatoon, SK, Canada
To email, remove "nospam" from [email protected]
Home Page: http://www.roberthancock.com/
#1 lore raw
Quoted 26 lines · On Mon, 05 Feb 2007 20:55:35 -0600 Robert Hancock <[email protected]> wrote:

Seeing these BUGs on 2.6.20-rc6-mm3 when mounting an NTFS partition. I saw some reports of something like this on -mm1, was this supposed to be patched already?

BUG: at arch/i386/mm/highmem.c:52 kmap_atomic()
  [<c041f810>] kmap_atomic+0xb4/0x1cd
  [<e0de8276>] ntfs_end_buffer_async_read+0x276/0x2db [ntfs]
  [<c048f92b>] end_bio_bh_io_sync+0x0/0x39
  [<c048f959>] end_bio_bh_io_sync+0x2e/0x39
  [<c049138f>] bio_endio+0x5b/0x63
  [<c05ffe97>] _spin_lock_irqsave+0x9/0xd
  [<c042f3f6>] lock_timer_base+0x15/0x2f
  [<c04d9f03>] __end_that_request_first+0x176/0x421
  [<e08c45a6>] scsi_delete_timer+0xf/0x50 [scsi_mod]
  [<e08c6817>] scsi_request_fn+0x2d4/0x324 [scsi_mod]
  [<e092ee7b>] ata_scsi_qc_complete+0x344/0x354 [libata]
  [<c04dba71>] __blk_run_queue+0x17/0x35
  [<e08c5601>] scsi_end_request+0x1a/0xa8 [scsi_mod]
  [<e08c5826>] scsi_io_completion+0x15a/0x32b [scsi_mod]
  [<c0409f81>] pit_next_event+0x2b/0x2f
  [<e08505c9>] sd_rw_intr+0x21b/0x245 [sd_mod]
  [<e08c13ff>] scsi_finish_command+0x84/0x8b [scsi_mod]
  [<c04dc19b>] blk_done_softirq+0x49/0x54
  [<c042b645>] __do_softirq+0x5d/0xba
  [<c042b6d4>] do_softirq+0x32/0x36
  [<c042b926>] irq_exit+0x38/0x6b
Quoted 8 lines
  [<c0406eb7>] do_IRQ+0x87/0x9c
  [<c0405837>] common_interrupt+0x23/0x28
  [<c043007b>] __dequeue_signal+0x97/0x14e
  [<c052709f>] acpi_processor_idle+0x209/0x3ca
  [<c0526e96>] acpi_processor_idle+0x0/0x3ca
  [<c0403410>] cpu_idle+0xaa/0xd0
  [<c071ca1f>] start_kernel+0x435/0x43d
  [<c071c1ae>] unknown_bootoption+0x0/0x202

Yes, that was a real bug. It got fixed in 2.6.20.

Below is the fix, which I have just uploaded to the hot-fixes directory, thanks.

fs/ntfs/aops.c renamed from fs/ntfs/aops.c~ntfs-kmap_atomic-atomicity-fix +4 -0
@@ -92,10 +92,12 @@ static void ntfs_end_buffer_async_read(s
 			ofs = 0;
 			if (file_ofs < init_size)
 				ofs = init_size - file_ofs;
+			local_irq_save(flags);
 			kaddr = kmap_atomic(page, KM_BIO_SRC_IRQ);
 			memset(kaddr + bh_offset(bh) + ofs, 0,
 					bh->b_size - ofs);
 			kunmap_atomic(kaddr, KM_BIO_SRC_IRQ);
+			local_irq_restore(flags);
 			flush_dcache_page(page);
 		}
 	} else {
@@ -143,11 +145,13 @@ static void ntfs_end_buffer_async_read(s
 		recs = PAGE_CACHE_SIZE / rec_size;
 		/* Should have been verified before we got here... */
 		BUG_ON(!recs);
+		local_irq_save(flags);
 		kaddr = kmap_atomic(page, KM_BIO_SRC_IRQ);
 		for (i = 0; i < recs; i++)
 			post_read_mst_fixup((NTFS_RECORD*)(kaddr +
 					i * rec_size), rec_size);
 		kunmap_atomic(kaddr, KM_BIO_SRC_IRQ);
+		local_irq_restore(flags);
 		flush_dcache_page(page);
 		if (likely(page_uptodate && !PageError(page)))
 			SetPageUptodate(page);

_