linux-kernel - Re: [PATCH 2/2] block: convert bsg, scsi_ioctl and cdrom to new blk_rq_map

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20060913093243.GH23515@kernel.dk>
Date:	Wed, 13 Sep 2006 11:32:43 +0200
From:	Jens Axboe <axboe@...nel.dk>
To:	Mike Christie <michaelc@...wisc.edu>
Cc:	linux-kernel@...r.kernel.org, linux-scsi@...r.kernel.org
Subject: Re: [PATCH 2/2] block: convert bsg, scsi_ioctl and cdrom to new  blk_rq_map_user fns

On Sat, Sep 09 2006, Mike Christie wrote:
> On Sat, 2006-09-09 at 16:53 -0400, Mike Christie wrote:
> > This patch just converts bsg, scsi_ioctl and cdrom to
> > the new functions. There is not really much of a change
> > for scsi_ioctl and cdrom. The block layer functions
> > handle things like bounce buffers, and bio unmapping accounting
> > for them now.
> > 
> > The bsg change is larger since I based the blk layer functions
> > on it and the scsi lib code, so there are lots of deletions.
> > There were also a couple of bugs that got fixes as a result
> > of using the blk layer API:
> > - If the hdr iovec_count = 0, then the request was not getting
> > freed because blk_unmap_sghdr_rq thought blk_rq_unmap_user
> > would free the request which it does not.
> > - blk_unmap_sghdr_rq would always call bio_unmap_user, even if
> > bio_copy_user was used. This was causing all types of weird bugs
> > like slab corruption.
> > 
> > I have tested the copy and map code with scsi_ioctl.c with large
> > requests having multiple bios. I have not tested bsg or cdrom.
> > And High Mem is not tested well. I am not sure how to force high
> > mem pages to be used for the mapping. Was there a proc or module
> > param that can be used?
> > 
> > Patch was made over Jens's block tree and the bsg branch
> > 
> > Signed-off-by: Mike Christie <michaelc@...wisc.edu>
> 
> 
> That patch was bad. As you can probably tell, I was trying to use git
> and was hand editing the patch at the same time. The patch below fixes
> the compile errors found in the previous patch.

Thanks, I will rebase bsg on top of the block branch and apply once we
get the 1/2 patch sorted.

-- 
Jens Axboe

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