>> Matt, is it ok with you for me to add this patch to the tree?
> I'd like to hold off a few more days while I try to find out what the
> 'secret sauce' that the other OSes use for a device like this.
Hmm. You do not confuse two situations, do you?
In the past few days I made two devices work.
One was a Feiya 5-in-1 CF / SM / SD card reader
(Vendor Id: 090c, Product Id: 1132, Revision 1.00).
It returned a capacity that is one too large, and becomes
very unhappy if one tries to read a sector past the end.
So, a flag was needed to tell that the result of READ CAPACITY
needs fixing.
The other was a Travelmate CF / SM / SD card reader
(Vendor Id: 3538, Product Id: 0001, Revision 2.05).
It became unhappy when MODE_SENSE asked for too much data.
A patch on sd.c solved this.
Andries
On Thu, Sep 05, 2002 at 12:56:00AM +0200, [email protected] wrote:
> >> Matt, is it ok with you for me to add this patch to the tree?
>
> > I'd like to hold off a few more days while I try to find out what the
> > 'secret sauce' that the other OSes use for a device like this.
>
> Hmm. You do not confuse two situations, do you?
> In the past few days I made two devices work.
>
> One was a Feiya 5-in-1 CF / SM / SD card reader
> (Vendor Id: 090c, Product Id: 1132, Revision 1.00).
> It returned a capacity that is one too large, and becomes
> very unhappy if one tries to read a sector past the end.
> So, a flag was needed to tell that the result of READ CAPACITY
> needs fixing.
Seems reasonble, Matt, any objection?
> The other was a Travelmate CF / SM / SD card reader
> (Vendor Id: 3538, Product Id: 0001, Revision 2.05).
> It became unhappy when MODE_SENSE asked for too much data.
> A patch on sd.c solved this.
Linus already added this patch to his tree :)
thanks,
greg k-h
Nope, not confused.
I'm trying to find out why Windows doesn't choke on the strange
READ_CAPACITY value. My guess is that there is code somewhere to
universally fixup READ_CAPACITY, and I'd like to know what the magic
formula is.
The MODE_SENSE bit is fine by me. But it's a SCSI change, so I'm not the
person to ask. The fix to sddr09.c for that seems reasonable, but you put
it all together and I haven't had time to split it up. Of course, I fully
expect the changes to MODE_SENSE in the SCSI layer to break other USB
devices, but there is only one way to find out....
Matt
On Thu, Sep 05, 2002 at 12:56:00AM +0200, [email protected] wrote:
> >> Matt, is it ok with you for me to add this patch to the tree?
>
> > I'd like to hold off a few more days while I try to find out what the
> > 'secret sauce' that the other OSes use for a device like this.
>
> Hmm. You do not confuse two situations, do you?
> In the past few days I made two devices work.
>
> One was a Feiya 5-in-1 CF / SM / SD card reader
> (Vendor Id: 090c, Product Id: 1132, Revision 1.00).
> It returned a capacity that is one too large, and becomes
> very unhappy if one tries to read a sector past the end.
> So, a flag was needed to tell that the result of READ CAPACITY
> needs fixing.
>
> The other was a Travelmate CF / SM / SD card reader
> (Vendor Id: 3538, Product Id: 0001, Revision 2.05).
> It became unhappy when MODE_SENSE asked for too much data.
> A patch on sd.c solved this.
>
> Andries
--
Matthew Dharm Home: [email protected]
Maintainer, Linux USB Mass Storage Driver
It's not that hard. No matter what the problem is, tell the customer
to reinstall Windows.
-- Nurse
User Friendly, 3/22/1998
On Wed, Sep 04, 2002 at 04:10:42PM -0700, Matthew Dharm wrote:
> I'm trying to find out why Windows doesn't choke on the strange
> READ_CAPACITY value.
That is an easy one.
It belongs to the recent partitioning discussion on l-k.
Windows knows the type of partition table, so reads the
partition table and the boot sector and the FAT and is happy.
Linux tries various things, depending on how you compiled your kernel,
and among other things also needs to examine the last sector.
So, only Linux will do bad things in case the capacity is off by one,
and only when your config includes partitioning types that use this
last sector.
Andries
Well, that's the best answer I've gotten so far.
I guess the patch can go in.
Matt
On Thu, Sep 05, 2002 at 01:46:53AM +0200, Andries Brouwer wrote:
> On Wed, Sep 04, 2002 at 04:10:42PM -0700, Matthew Dharm wrote:
>
> > I'm trying to find out why Windows doesn't choke on the strange
> > READ_CAPACITY value.
>
> That is an easy one.
> It belongs to the recent partitioning discussion on l-k.
>
> Windows knows the type of partition table, so reads the
> partition table and the boot sector and the FAT and is happy.
>
> Linux tries various things, depending on how you compiled your kernel,
> and among other things also needs to examine the last sector.
> So, only Linux will do bad things in case the capacity is off by one,
> and only when your config includes partitioning types that use this
> last sector.
--
Matthew Dharm Home: [email protected]
Maintainer, Linux USB Mass Storage Driver
I don't have a left mouse button. I only have one mouse and it's on my right.
-- Customer
User Friendly, 2/13/1999
On Wed, Sep 04, 2002 at 04:10:42PM -0700, Matthew Dharm wrote:
> The MODE_SENSE bit is fine by me. But it's a SCSI change, so I'm not the
> person to ask. The fix to sddr09.c for that seems reasonable, but you put
> it all together and I haven't had time to split it up. Of course, I fully
> expect the changes to MODE_SENSE in the SCSI layer to break other USB
> devices, but there is only one way to find out....
Speaking of sddr09 devices (of which I possess one), what's the current
state of support for the sddr09? Last I knew, at the time when I finally
got mine working, the sddr09 was only supported read-only, since there
were not-fully-understood issues that resulted in filesystem corruption
if you tried to write to it. Currently, the driver will still allow me
to mount it only as a read-only filesystem. Can write support for the
sddr09 be expected any time soon?
--
********* Fight Back! It may not be just YOUR life at risk. *********
phil stracchino :: [email protected] :: [email protected]
unix ronin :::: renaissance man :::: mystic zen biker geek
2000 CBR929RR, 1991 VFR750F3 (foully murdered), 1986 VF500F (sold)
Linux Now! ...because friends don't let friends use Microsoft.