Re: Linux 2.4.30-rc3 md/ext3 problems (ext3 gurus : please check)

From: Stephen C. Tweedie
Date: Thu Apr 28 2005 - 06:17:28 EST


On Thu, 2005-04-28 at 09:54, Hifumi Hisashi wrote:
> At 07:22 05/04/28, Stephen C. Tweedie wrote:
> >In checking out the patch one last time, though, I found one anomaly.
> >The patch that was submitted to 2.6 for this problem also changed
> >write_inode_now() to return an error value. The patch that got
> >submitted to 2.4 had no such change. Was there a reason for this
> >difference between the two versions?
> Right. Also in ver 2.4, I know write_inode_now() has to be changed for perfect
> I/O error detection during synchronous writing.
> In do_generic_file_write(mm/filemap.c), does the current return value handling is
> unchanged?

At present I'm using your changed version, but that is wrong --- in most
cases, we want to return short writes on errors. It's only in the
synchronous case that we want to do otherwise, and even then only on EIO
on the inode. For example, if we run out of quota halfway through an
O_SYNC write beyond EOF, then the bulk of the IO *has* completed
successfully, and we want to indicate that in the error. So I'll be
re-jigging that bit of the code.


To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at