Bug#376001: partman-base: humandev chokes on /dev/cciss/c0d0

["Alex.Owen" <r.a.owen@qmul.ac.uk>]

Will try and remember to test this on Monday.

By the way having a quick look at the diff I notice the comment:
+           # It would be a lot easier to parse the /sys/block/*/device
+           # symlink. Unfortunately, unlike other block devices, this
+           # doesn't seem to exist in this case, so we just have to live
+           # with parsing the device name.

I've also found this LKML post:
http://www.uwsg.iu.edu/hypermail/linux/kernel/0606.2/1982.html
Date: Wed Jun 21 2006
Subject: [PATCH 3/22] [PATCH] CCISS: add device symlink to the block cciss 
block devices in sysfs

containing the one line kernel patch:
--
drivers/block/cciss.c | 1 +
1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/drivers/block/cciss.c b/drivers/block/cciss.c
index 1319d8f..25c3c4a 100644
--- a/drivers/block/cciss.c
+++ b/drivers/block/cciss.c
@@ -3237,6 +3237,7 @@ #endif /* CCISS_DEBUG */
disk->fops = &cciss_fops;
disk->queue = q;
disk->private_data = drv;
+ disk->driverfs_dev = &pdev->dev;
/* we must register the controller even if no disks exist */
/* this is for the online array utilities */
if(!drv->heads && j)
--

I don't know if that would be usefull to have in the debian kernel... or 
we just wait till it hits the mainline kernel.

Thanks for the speedy response!
Alex Owen