From: Grant Grundler <grundler@parisc-linux.org>
To: parisc-linux@lists.parisc-linux.org
Subject: [parisc-linux] ext3 perf patch#2
Date: Sat, 26 Mar 2005 03:35:40 -0700 [thread overview]
Message-ID: <20050326103540.GA31557@colo.lackof.org> (raw)
James Bottomley in a private conversation suggested our
ext2_test_bit() and ext2_find_next_zero_bit() might not
be counting bits the same way in the bitmap.
ext2_test_bit() is quite straight forward looks right to me.
But my brain wasn't grokking ext2_find_next_zero_bit().
Cloning it from sparc64 was easy to try...
Since the first attempted worked, I also applli
The result was rsync was able to clone from one
ext3 (whole disk) to another ext3 using rsync
at 5-10MB/s (both disks on the same SCSI bus).
This is *alot* better than the previous experience
of < 1MB/s. Since my ext3 test filesystems hadn't aged (been used)
very much, bitmap search is not much overhead.
If someone else has a disposable, mature ext3 file system
they can test the following patch on, that would be great.
BTW, this patch assumes a 64-bit kernel and is certainly
not intended to be committed as is. This is just
an experiment so far.
Patch is against 2.6.11-pa2 (NOT 2.6.12 - sorry)
grant
Index: include/asm-parisc/bitops.h
===================================================================
RCS file: /var/cvs/linux-2.6/include/asm-parisc/bitops.h,v
retrieving revision 1.14
diff -u -p -r1.14 bitops.h
--- include/asm-parisc/bitops.h 18 Feb 2005 14:22:09 -0000 1.14
+++ include/asm-parisc/bitops.h 26 Mar 2005 09:58:58 -0000
@@ -466,49 +466,47 @@ static __inline__ int ext2_test_bit(int
return (ADDR[nr >> 3] >> (nr & 7)) & 1;
}
-/*
- * This implementation of ext2_find_{first,next}_zero_bit was stolen from
- * Linus' asm-alpha/bitops.h and modified for a big-endian machine.
- */
-#define ext2_find_first_zero_bit(addr, size) \
- ext2_find_next_zero_bit((addr), (size), 0)
-
-extern __inline__ unsigned long ext2_find_next_zero_bit(void *addr,
- unsigned long size, unsigned long offset)
+static __inline__ unsigned long ext2_find_next_zero_bit(unsigned long *addr, unsigned long size, unsigned long offset)
{
- unsigned int *p = ((unsigned int *) addr) + (offset >> 5);
- unsigned int result = offset & ~31UL;
- unsigned int tmp;
+ unsigned long *p = addr + (offset >> 6);
+ unsigned long result = offset & ~63UL;
+ unsigned long tmp;
if (offset >= size)
return size;
size -= result;
- offset &= 31UL;
- if (offset) {
- tmp = cpu_to_le32p(p++);
- tmp |= ~0UL >> (32-offset);
- if (size < 32)
+ offset &= 63UL;
+ if(offset) {
+ tmp = __swab64p((u64 *)p++);
+ tmp |= (~0UL >> (64-offset));
+ if(size < 64)
goto found_first;
- if (tmp != ~0U)
+ if(~tmp)
goto found_middle;
- size -= 32;
- result += 32;
+ size -= 64;
+ result += 64;
}
- while (size >= 32) {
- if ((tmp = cpu_to_le32p(p++)) != ~0U)
- goto found_middle;
- result += 32;
- size -= 32;
+ while(size & ~63) {
+ if(~(tmp = *(p++)))
+ goto found_middle_swap;
+ result += 64;
+ size -= 64;
}
- if (!size)
+ if(!size)
return result;
- tmp = cpu_to_le32p(p);
+ tmp = __swab64p((u64 *) p);
found_first:
- tmp |= ~0U << size;
+ tmp |= (~0UL << size);
+ if (tmp == ~0UL) /* Are any bits zero? */
+ return result + size; /* Nope. */
found_middle:
return result + ffz(tmp);
+
+found_middle_swap:
+ return result + ffz(__swab64((u64)tmp));
}
+
/* Bitmap functions for the minix filesystem. */
#define minix_test_and_set_bit(nr,addr) ext2_set_bit(nr,addr)
_______________________________________________
parisc-linux mailing list
parisc-linux@lists.parisc-linux.org
http://lists.parisc-linux.org/mailman/listinfo/parisc-linux
next reply other threads:[~2005-03-26 10:35 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-03-26 10:35 Grant Grundler [this message]
[not found] ` <42454E7C.3050509@tiscali.be>
2005-03-26 21:39 ` [parisc-linux] ext3 perf patch#2 Grant Grundler
2005-03-27 1:18 ` [parisc-linux] ext3 perf patch#4 Grant Grundler
2005-03-27 13:49 ` Joel Soete
2005-03-28 3:48 ` Grant Grundler
2005-03-28 12:51 ` Joel Soete
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20050326103540.GA31557@colo.lackof.org \
--to=grundler@parisc-linux.org \
--cc=parisc-linux@lists.parisc-linux.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox