From: Toshiyuki Okajima <toshi.okajima@xxxxxxxxxxxxxx> wb_kupdate() function has a bug on linux-2.6.30-rc5. This bug causes generic_sync_sb_inodes() to start to write inodes back much earlier than our expectations because it miscalculates oldest_jif in wb_kupdate(). Signed-off-by: Toshiyuki Okajima <toshi.okajima@xxxxxxxxxxxxxx> --- page-writeback.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff -Nurp linux-2.6.30-rc5.orig/mm/page-writeback.c linux-2.6.30-rc5/mm/page-writeback.c --- linux-2.6.30-rc5.orig/mm/page-writeback.c 2009-05-09 09:14:14.000000000 +0900 +++ linux-2.6.30-rc5/mm/page-writeback.c 2009-05-15 20:19:03.000000000 +0900 @@ -94,12 +94,12 @@ unsigned long vm_dirty_bytes; /* * The interval between `kupdate'-style writebacks */ -unsigned int dirty_writeback_interval = 5 * 100; /* sentiseconds */ +unsigned int dirty_writeback_interval = 5 * 100; /* centiseconds */ /* * The longest time for which data is allowed to remain dirty */ -unsigned int dirty_expire_interval = 30 * 100; /* sentiseconds */ +unsigned int dirty_expire_interval = 30 * 100; /* centiseconds */ /* * Flag that makes the machine dump writes/reads and block dirtyings. @@ -770,7 +770,7 @@ static void wb_kupdate(unsigned long arg sync_supers(); - oldest_jif = jiffies - msecs_to_jiffies(dirty_expire_interval); + oldest_jif = jiffies - msecs_to_jiffies(dirty_expire_interval * 10); start_jif = jiffies; next_jif = start_jif + msecs_to_jiffies(dirty_writeback_interval * 10); nr_to_write = global_page_state(NR_FILE_DIRTY) + -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
