I have to apologize, I just now read the part where find_next_zero is
using 42% of CPU.
Jeff, this has the "feel" of a bug (affecting performance not
correctness). I am skeptical that we are searching only bitmap blocks
in which we successfully find and use a free block. Could you look at
the code and his results with some care? This profiling result is what
I would have expected to see BEFORE your optimizations occurred, and I
would not expect it now. Thanks Mike for bringing this to our attention.
Hans
-------------------------
From: Jeff Mahoney <[EMAIL PROTECTED]>
Subject: [PATCH 5/5] reiserfs: make bitmap use cached first zero bit
Currently, the bitmap code uses half of the hinting data gathered and
cached
and wastes the other half. We'll skip completely full bitmaps, but start
scanning in bitmaps at locations where if we consulted the zero bit hint,
we'd know there aren't any free bits available.
This patch uses the first zero hint to bump the beginning of the search
window to where we know there is at least one zero bit.
fs/reiserfs/bitmap.c | 5 ++++-
1 files changed, 4 insertions(+), 1 deletion(-)
Signed-off-by: Jeff Mahoney <[EMAIL PROTECTED]>
diff -ruNpX ../dontdiff linux-2.6.15.orig.staging1/fs/reiserfs/bitmap.c
linux-2.6.15.orig.staging2/fs/reiserfs/bitmap.c
--- linux-2.6.15.orig.staging1/fs/reiserfs/bitmap.c 2006-01-16
16:53:35.663319136 -0500
+++ linux-2.6.15.orig.staging2/fs/reiserfs/bitmap.c 2006-01-16
16:53:35.673317616 -0500
@@ -187,7 +187,10 @@ static int scan_bitmap_block(struct reis
return 0; // No free blocks in this bitmap
}
- /* search for a first zero bit -- beggining of a window */
+ if (*beg < bi->first_zero_hint)
+ *beg = bi->first_zero_hint;
+
+ /* search for a first zero bit -- beginning of a window */
*beg = reiserfs_find_next_zero_le_bit
((unsigned long *)(bh->b_data), boundary, *beg);
