Subject: Re: 2.6.4-mm2
From: Chris Mason <mason@suse.com>
To: Daniel McNeil <daniel@osdl.org>
Cc: Andrew Morton <akpm@osdl.org>,
       Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
       "linux-aio@kvack.org" <linux-aio@kvack.org>
In-Reply-To: <1079632431.6930.30.camel@ibm-c.pdx.osdl.net>
References: <20040314172809.31bd72f7.akpm@osdl.org>
	 <1079461971.23783.5.camel@ibm-c.pdx.osdl.net>
	 <1079474312.4186.927.camel@watt.suse.com>
	 <20040316152106.22053934.akpm@osdl.org>
	 <20040316152843.667a623d.akpm@osdl.org>
	 <20040316153900.1e845ba2.akpm@osdl.org>
	 <1079485055.4181.1115.camel@watt.suse.com>
	 <1079487710.3100.22.camel@ibm-c.pdx.osdl.net>
	 <20040316180043.441e8150.akpm@osdl.org>
	 <1079554288.4183.1938.camel@watt.suse.com>
	 <20040317123324.46411197.akpm@osdl.org>
	 <1079563568.4185.1947.camel@watt.suse.com>
	 <20040317150909.7fd121bd.akpm@osdl.org>
	 <1079566076.4186.1959.camel@watt.suse.com>
	 <20040317155111.49d09a87.akpm@osdl.org>
	 <1079568387.4186.1964.camel@watt.suse.com>
	 <20040317161338.28b21c35.akpm@osdl.org>
	 <1079569870.4186.1967.camel@watt.suse.com>
	 <20040317163332.0385d665.akpm@osdl.org>
	 <1079572511.6930.5.camel@ibm-c.pdx.osdl.net>
	 <1079632431.6930.30.camel@ibm-c.pdx.osdl.net>
Content-Type: text/plain
Message-Id: <1079635678.4185.2100.camel@watt.suse.com>
Mime-Version: 1.0
Date: Thu, 18 Mar 2004 13:47:58 -0500
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1414
Lines: 36

On Thu, 2004-03-18 at 12:53, Daniel McNeil wrote:
> I'm ran 2.6.4-mm2 plus the 2 wait_on_page_range() patches,
> the test_set_page_writeback() patch and clear_page_dirty_for_io patch
> overnight.
> 
> 6 copies of direct_read_under test on 8-cpu system on 1
> ext3 file system in 1 directory on a scsi disk.
> (http://developer.osdl.org/daniel/AIO/TESTS/direct_read_under.c)
> 
> 5 of the 6 tests saw uninitialized data within 2 hours.
> The sixth test ran overnight.

Do you still have the errors generated?  I wondering how big the range
of uninitialized data was.  When I was bug hunting yesterday, I saw
ranges from 64k to 32mb in size, which was why I decided writes weren't
getting to the disk at all.

It might be interesting to try with data=writeback, or on ext2.  Things
might be easier to track if we're not worried about ll_rw_block.

It might also be interesting to significantly lower the size of the
reads and writes done by direct_read_under, or anything else you can
think of to get the reproduce time down to something smaller than 2
hours...

It's probably a good idea to upgrade to 2.6.5-rc1-mm2, just so we're all
staring at the same code.

-chris


-
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/