Re: 2.6.23-rc9 boot failure (megaraid?)

[James Bottomley <James.Bottomley@SteelEye.com>]

On Tue, 2007-10-02 at 20:15 +0200, Adrian Bunk wrote:
> Cc's added, the complete bug report is at
>   http://lkml.org/lkml/2007/10/2/243
> 
> On Tue, Oct 02, 2007 at 12:48:26PM -0400, Burton Windle wrote:
> > 2.6.23-rc9 fails to boot for me; 2.6.22.9 works fine.
> >
> > System is a Dell Poweredge with PERC 2/DC with RAID1 volume.
> >...
> 
> Thanks for your report.
> 
> Diff'ing the dmesg's shows:
> 
> <--  snip  -->
> 
>  scsi0: scanning scsi channel 4 [P0] for physical devices.
>  scsi0: scanning scsi channel 5 [P1] for physical devices.
>  st: Version 20070203, fixed bufsize 32768, s/g segs 256
> -sd 0:0:0:0: [sda] 17547264 512-byte hardware sectors (8984 MB)
> +sd 0:0:0:0: [sda] Sector size 0 reported, assuming 512.
> +sd 0:0:0:0: [sda] 1 512-byte hardware sectors (0 MB)
>  sd 0:0:0:0: [sda] Write Protect is off
>  sd 0:0:0:0: [sda] Asking for cache data failed
>  sd 0:0:0:0: [sda] Assuming drive cache: write through
> -sd 0:0:0:0: [sda] 17547264 512-byte hardware sectors (8984 MB)
> +sd 0:0:0:0: [sda] Sector size 0 reported, assuming 512.
> +sd 0:0:0:0: [sda] 1 512-byte hardware sectors (0 MB)
>  sd 0:0:0:0: [sda] Write Protect is off
>  sd 0:0:0:0: [sda] Asking for cache data failed
>  sd 0:0:0:0: [sda] Assuming drive cache: write through
>   sda: sda1
> + sda: p1 exceeds device capacity
> 
> <--  snip  -->
> 
> -	case MEGA_BULK_DATA:
> -		if (scb->cmd->use_sg == 0)
> -			length = scb->cmd->request_bufflen;
> -		else {
> -			struct scatterlist *sgl =
> -				(struct scatterlist *)scb->cmd->request_buffer;
> -			length = sgl->length;
> -		}
> -		pci_unmap_page(adapter->dev, scb->dma_h_bulkdata,
> -			       length, scb->dma_direction);
> -		break;
> -

This is the problem piece I think.  We've reintroduced a very old bug:

commit 51c928c34fa7cff38df584ad01de988805877dba
Author: James Bottomley <James.Bottomley@SteelEye.com>
Date:   Sat Oct 1 09:38:05 2005 -0500

    [SCSI] Legacy MegaRAID: Fix READ CAPACITY
    
    Some Legacy megaraid cards can't actually cope with the scatter/gather
    version of the READ CAPACITY command (which is what we now send them
    since altering all SCSI internal I/O to go via the block layer).  Fix
    this (and a few other broken megaraid driver assumptions) by sending
    the non-sg version of the command if the sg list only has a single
    element.
    
    Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>

So what we have to do is put back the check for use_sg == 1 and send
that as a bulk transfer command.

James