Re: [Ext2-devel] [PATCH 1/2] ext2/3: Support 2^32-1 blocks(Kernel)

["Theodore Ts'o" <tytso@mit.edu>]

On Wed, Mar 15, 2006 at 09:39:20PM +0900, Takashi Sato wrote:
>  - Modify to call C functions(ext2fs_set_bit(),ext2fs_test_bit())
>    defined in lib/ex2fs/bitops.c on x86 and mc68000 architecture.

I just did some quick tests, and using the C functions instead of the
asm instructions increases the CPU user time by 7.8% on a test
filesystem, and increased the wall clock time of the e2fsck regression
test suite by 4.5%.  That's not huge, and the test suite was cached so
the percentage difference will probably be less in real-life
situation, but I'd still like to avoid it if I can.

I've just checked my i386 assembly language reference, and I don't see
any indication that the btsl, btrl, and btl instructions don't work if
the high bit is set on the bit number.  Have you done tests showing
that these instructions do not work correctly for filesystem sizes >
2**31 blocks, or have references showing that these instructions
interpret the bit number as a signed integer?

Thanks regards,

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