Sync_page_io() call hangs when called from make_request (2.6.25)

From: Unmil Tambe
Date: Wed Jul 16 2008 - 09:57:50 EST


Hello,

I am writing a stacking block device driver which writes some metadata
on a block device before redirecting the data bio to underlying disk.
For this I have been using sync_page_io from raid/md.h to do the
metadata write synchronously.

This worked fine in 2.6.18.2 kernel. I am now trying to port the
driver to 2.6.25 kernel, but even in a basic stacking driver in this
kernel, the sync_page_io call does not return. The process (dd in this
case) never returns. The machine does not freeze, nor is there an OOPS
generated. It is as if waiting on a lock.

All I could find in my search was this msg at:
http://repo.ugm.ac.id/linux/ChangeLog-2.6.22.12

"Author: Milan Broz <mbroz@xxxxxxxxxx>
Date: Thu Jul 12 17:28:00 2007 +0100

dm snapshot: fix invalidation deadlock

patch fcac03abd325e4f7a4cc8fe05fea2793b1c8eb75 in mainline

Process persistent exception store metadata IOs in a separate thread.

A snapshot may become invalid while inside generic_make_request().
A synchronous write is then needed to update the metadata while still
inside that function. Since the introduction of
md-dm-reduce-stack-usage-with-stacked-block-devices.patch this has to
be performed by a separate thread to avoid deadlock. "


I dont really know whether this is related to my problem or not. Any ideas?

Thanks,
Unmil
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/