[Wilfried.Weissmann@gmx.at: Bug#277534: hal: harddisk partition scan does not check harddisk size]

[Wilfried Weissmann]

Kay Sievers wrote:
> On Fri, 2004-10-22 at 20:21 +0200, Wilfried Weissmann wrote:
>>The partition size is correct but the partition is physically not on 
>>that device. The first raid device contains the partition table at the 
>>same position where the kernel expects the table of the standalone 
>>harddisk (bad, bad policy from highpoint-tech). The kernel just does not 
>>know that this disk is part of a raid volume and it provides block 
>>devices for hdg1-3.
>>The partition table is valid for the raid volume but not for hdg.
> 
> 
> Do I get this right? The kernel creates devices for your disk, which are
> not usable? Also the sysfs repreentation is invalid then? Then the
> kernel should be fixed, right?

You are right! Not creating hdg3 in the kernel when the location of the 
partition is obviously beyond the end of the disk is a good idea. 
However, as David Zeuthen pointed out in his second mail, this does not 
completely fix any issues with hdg1.

> 
> Or is there any signature (any single byte or magic string) on the
> volume that indicates, that this is part of a raid set (similar to the
> linux-raid, which writes a magic at the end of the volume)?

There is a raid signature at sector 9 which indicates the volume 
configuration. See: http://people.redhat.com/~heinzm/sw/dmraid/
There is also a single disk configuration mode. In that case the 
partition table is still valid.
I think checking this signature would create just duplicate code. 
Interworking between hal and dmraid (or its ataraid-lib) prevents 
reimplementing existing functionality. I also plan to use ataraid-lib 
for an evms-plugin. I just did not have time to do so (blame on me).

> 
> Kay

Greetings,
Wilfried
_______________________________________________
hal mailing list
hal at freedesktop.org
http://freedesktop.org/mailman/listinfo/hal