Re: [PATCH 000 of 6] A few block-layer tidy-up patches.

[Jens Axboe <jens.axboe@oracle.com>]

On Fri, Aug 17 2007, Neil Brown wrote:
> On Friday August 17, jens.axboe@oracle.com wrote:
> > > > Please inspect the #block-2.6.24 branch to see the result.
> > > 
> > > I don't know where to look for this.  I checked
> > >   http://git.kernel.org/?p=linux/kernel/git/axboe/linux-2.6-block.git
> > > but they don't seem to be there.
> > > ??
> > 
> > That's where it is, but the kernel.org mirroring is just horrible slow.
> > Just checked now and it's there.
> 
> I discovered I was looking in the wrong place - not being very
> familiar with git terminology.
> I found them and it looks right.

Ah, good.

> I had a bit of a look at removing bio_data and ->buffer ...  the
> usages outside of drivers/ide are fairly easy to deal with - I might
> send a patch for that.  The drivers/ide stuff looks like a fairly
> substantial rewrite is in order.
> e.g.  idefloppy_packet_command_s seems to duplicate a lot of 
>   fields from 'struct request', and it should probably use the request
>   struct directly.

We can do it bits at the time, the triviel ones first. I have some
experience with drivers/ide/ myself, so I can attempt to tackle some of
that myself.

> But a number of times ->buffer points to ->cmd, and there is no bio.
> I guess we should use bio_map_kern to make a bio?

It points to ->cmd?! But yes, generally things just need to be
converted to map the data to a bio with bio_map_kern() (or
bio_map_user(), where appropriate).

> I'll see if I can come up with something.... testing it might be
> awkward.  I have an ide cdrom I can test on.  Maybe an ide disk,, but
> not an ide floppy :-)

I can help with that as well :-)

-- 
Jens Axboe