Re: [PATCH] Block layer: separate out queue-oriented ioctls Re: [PATCH] Block layer: separate out queue-oriented ioctls

Re: [PATCH] Block layer: separate out queue-oriented ioctls Re: [PATCH] Block layer: separate out queue-oriented ioctls

[Joerg Schilling]

Hi,

this is a repost as I like to know the current state of the problem...

The USB DMA size problem is known to exist on Linux since February 2004.
I am still in hope that there will be a fix soon.

/*--------------------------------------------------------------------------*/
Alan Stern <stern@rowland.harvard.edu> wrote:

> Well, if Doug wants to reduce the value returned by SG_GET_RESERVED_SIZE, 
> it's okay with me.  An advantage of doing this is that older versions of 
> cdrecord would then work correctly.
>
> However you don't seem to realize that people can use programs like
> cdrecord with devices whose drivers don't support SG_GET_RESERVED_SIZE --
> because that ioctl works only with sg.  Programs would have to try
> SG_GET_RESERVED_SIZE and if it faied, then try BLKSECTGET.

Is there any reason not to have one single ioctl for one basic feature?

> Remember also, the "reserved size" is _not_ the maximum allowed size of a
> DMA transfer.  Rather, it is the size of an internal buffer maintained by
> sg.  It's legal to do an I/O transfer larger than the "reserved size", but 
> it is not legal to do an I/O transfer larger than max_sectors.

At the time the call SG_GET_RESERVED_SIZE has been discussed/defined, we did 
originally agree that the max value should be limited to what the HW allows
as DMA size. This is why I did originally files a bug against 
SG_GET_RESERVED_SIZE.
/*--------------------------------------------------------------------------*/

Jörg

-- 
 EMail:joerg@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin
       js@cs.tu-berlin.de                (uni)  
       schilling@fokus.fraunhofer.de     (work) Blog: http://schily.blogspot.com/
 URL:  http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily

Re: [PATCH] Block layer: separate out queue-oriented ioctls Re: [PATCH] Block layer: separate out queue-oriented ioctls

[Alan Stern]

On Mon, 2 Apr 2007, Joerg Schilling wrote:

> Hi,
> 
> this is a repost as I like to know the current state of the problem...
> 
> The USB DMA size problem is known to exist on Linux since February 2004.
> I am still in hope that there will be a fix soon.

Me too.  I submitted the most recent version of the patch (labelled as857) 
over a month ago and have received essentially no feedback on it.

Douglas, James...  What's the story?

Alan Stern

Re: [PATCH] Block layer: separate out queue-oriented ioctls Re: [PATCH] Block layer: separate out queue-oriented ioctls

[Alan Stern]

On Mon, 2 Apr 2007, Joerg Schilling wrote:

> Hi,
> 
> this is a repost as I like to know the current state of the problem...
> 
> The USB DMA size problem is known to exist on Linux since February 2004.
> I am still in hope that there will be a fix soon.

Me too.  I submitted the most recent version of the patch (labelled as857) 
over a month ago and have received essentially no feedback on it.

Douglas, James...  What's the story?

Alan Stern

Re: [PATCH] Block layer: separate out queue-oriented ioctls Re: [PATCH] Block layer: separate out queue-oriented ioctls

[Jens Axboe]

On Mon, Apr 02 2007, Alan Stern wrote:
> On Mon, 2 Apr 2007, Joerg Schilling wrote:
> 
> > Hi,
> > 
> > this is a repost as I like to know the current state of the problem...
> > 
> > The USB DMA size problem is known to exist on Linux since February 2004.
> > I am still in hope that there will be a fix soon.
> 
> Me too.  I submitted the most recent version of the patch (labelled as857) 
> over a month ago and have received essentially no feedback on it.
> 
> Douglas, James...  What's the story?

FWIW, it had my ack, I think we were just waiting for Doug to ack the sg
bits.

-- 
Jens Axboe

Re: [PATCH] Block layer: separate out queue-oriented ioctls Re: [PATCH] Block layer: separate out queue-oriented ioctls

[James Bottomley]

On Mon, 2007-04-02 at 16:23 +0200, Jens Axboe wrote:
> FWIW, it had my ack, I think we were just waiting for Doug to ack the sg
> bits.

And there's really nothing I can do (well, except write the thing) since
the changes are not in any SCSI pieces I maintain directly ... they're
block and sg.

James

Re: [PATCH] Block layer: separate out queue-oriented ioctls Re: [PATCH] Block layer: separate out queue-oriented ioctls

[James Bottomley]

On Mon, 2007-04-02 at 16:23 +0200, Jens Axboe wrote:
> FWIW, it had my ack, I think we were just waiting for Doug to ack the sg
> bits.

And there's really nothing I can do (well, except write the thing) since
the changes are not in any SCSI pieces I maintain directly ... they're
block and sg.

James