Hi Andrew,
With the following patch I'm able to get good interactive response
even when running a 256-thread iozone session over NFSv3 + X +... (as
opposed to simply crashing due to OOM). It contains the following
elements:
- Add missing unstable write accounting in sync_inodes_sb()
- Add missing unstable write accounting in wakeup_bdflush().
- Decrement nr_unstable only when the COMMIT RPC call is done
(rather then doing so when we schedule the RPC call). This ensures
that we do try to wait for completion.
- It is better to do too many rather than too *few* writes. Don't
overestimate how many pages we wrote out in nfs_writepages().
Comments and/or objections before I send it on to Linus?
Cheers,
Trond
--- linux-2.5.68-up/fs/nfs/write.c.orig 2003-04-15 04:46:02.000000000 +0200
+++ linux-2.5.68-up/fs/nfs/write.c 2003-05-04 16:15:50.000000000 +0200
@@ -280,8 +280,6 @@
err = nfs_wb_all(inode);
} else
nfs_commit_file(inode, NULL, 0, 0, 0);
- /* Avoid races. Tell upstream we've done all we were told to do */
- wbc->nr_to_write = 0;
out:
return err;
}
@@ -490,7 +488,6 @@
int res;
res = nfs_scan_list(&nfsi->commit, dst, file, idx_start, npages);
nfsi->ncommit -= res;
- sub_page_state(nr_unstable,res);
if ((nfsi->ncommit == 0) != list_empty(&nfsi->commit))
printk(KERN_ERR "NFS: desynchronized value of nfs_i.ncommit.\n");
return res;
@@ -1009,6 +1006,7 @@
{
struct nfs_write_data *data = (struct nfs_write_data *)task->tk_calldata;
struct nfs_page *req;
+ int res = 0;
dprintk("NFS: %4d nfs_commit_done (status %d)\n",
task->tk_pid, task->tk_status);
@@ -1043,7 +1041,9 @@
nfs_mark_request_dirty(req);
next:
nfs_unlock_request(req);
+ res++;
}
+ sub_page_state(nr_unstable,res);
}
#endif
--- linux-2.5.68-up/fs/fs-writeback.c.orig 2003-04-08 12:16:30.000000000 +0200
+++ linux-2.5.68-up/fs/fs-writeback.c 2003-05-04 15:09:32.000000000 +0200
@@ -367,7 +367,8 @@
};
get_page_state(&ps);
- wbc.nr_to_write = ps.nr_dirty + ps.nr_dirty / 4;
+ wbc.nr_to_write = ps.nr_dirty + ps.nr_unstable +
+ (ps.nr_dirty + ps.nr_unstable) / 4;
spin_lock(&inode_lock);
sync_sb_inodes(sb, &wbc);
spin_unlock(&inode_lock);
--- linux-2.5.68-up/mm/page-writeback.c.orig 2003-04-20 08:22:32.000000000 +0200
+++ linux-2.5.68-up/mm/page-writeback.c 2003-05-04 14:39:40.000000000 +0200
@@ -272,7 +272,7 @@
struct page_state ps;
get_page_state(&ps);
- nr_pages = ps.nr_dirty;
+ nr_pages = ps.nr_dirty + ps.nr_unstable;
}
return pdflush_operation(background_writeout, nr_pages);
}
-
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/
This archive was generated by hypermail 2b29 : Wed May 07 2003 - 22:00:20 EST