Skip to contentSkip to conversation

balloc free bit already set and byte minus one

DISCUSSION Pat LaVarre 1 participant 2 messages Dec 5, 2003 – Dec 10, 2003 VFS

lore
Share & tools
lore
#0 lore raw

-----Forwarded Message-----

To: [email protected]
Subject: balloc free bit already set and byte minus one
Date: 04 Dec 2003 17:37:54 -0700

Quoted 5 lines

To: [email protected]
Date: 06 Oct 2003 18:31:31 -0600

Often I see dmesg like:

Quoted 7 lines

UDF-fs DEBUG fs/udf/balloc.c:192:udf_bitmap_free_blocks: bit 16383 already set UDF-fs DEBUG fs/udf/balloc.c:193:udf_bitmap_free_blocks: byte=ffffffff

Specifically I see these dmesg if ...

Do these messages mean trouble, or should I ignore them?
...

YES these mean trouble i.e. "trying to free a block twice".

Offline kindly just now I received that answer.

Pat LaVarre

#1 lore raw
Quoted 3 lines
Quoted 2 lines

UDF-fs DEBUG fs/udf/balloc.c:192:udf_bitmap_free_blocks: bit 16383 already set UDF-fs DEBUG fs/udf/balloc.c:193:udf_bitmap_free_blocks: byte=ffffffff

... mean trouble ...trying to free a block twice...

Reading fs/udf/balloc.c now, ...

1)

CONFIG_SMP and CONFIG_PREEMPT may break less often since much work occurs inside { lock_super(sb); ... unlock_super(sb); }.

2)

--- fs/udf/balloc.c:

#define udf_clear_bit(nr,addr) ext2_clear_bit(nr,addr) #define udf_set_bit(nr,addr) ext2_set_bit(nr,addr) #define udf_test_bit(nr, addr) ext2_test_bit(nr, addr)

---

Possibly these mean mess with the bitmask (1 << nr) at byte-aligned (addr) and return whether that bitmask was nonzero.

By contrast, dropping the udf_ or ext2_ prefixes may contract for a long-aligned address e.g. via include/asm-ppc/bitops.h

Pat LaVarre