Hrm... looks like a bug... srb->cmnd[2] & 0x3F == 0x3F, not 0x01. This is
probably a leftover from when sd.c used to request page 1, instead of "all
pages" -- I changed this when I discovered devices that didn't implement
page 1, causing the write-protect test to fail. Requesting all pages is
guaranteed to be good, tho.
Matt
On Mon, Mar 26, 2001 at 11:17:00PM -0600, William Robison wrote:
> Hey Rob:
> More information...
>
> Here is a clip of the log with debugging enabled:
>
> usb-storage: *** thread sleeping.
> usb-storage: queuecommand() called
> usb-storage: *** thread awakened.
> usb-storage: Command MODE_SENSE (6 bytes)
> usb-storage: 1a 00 3f 00 08 00 00 00 00 00 1c c0
> usb-storage: CB_reset() called
> usb-storage: CB_reset: clearing endpoint halt
>
> And the clip out of sddr09.c:
>
> if (srb->cmnd[0] == MODE_SENSE) {
> // Read-write error recovery page: there needs
> to
> // be a check for write-protect here
> if ( (srb->cmnd[2] & 0x3F) == 0x01 ) {
> if (ptr==NULL || srb->request_bufflen<4)
> return USB_STOR_TRANSPORT_ERROR;
> memcpy(ptr, mode_page_01, sizeof(mode_page_01));
>
> return USB_STOR_TRANSPORT_GOOD;
> }
> // FIXME: sense buffer?
>
> return USB_STOR_TRANSPORT_ERROR;
>
> If I understand things correctly, srb->cmnd[2] is 0x3F so sddr09
> returns an error status. sd.c seems to ibe snooping for is the
> write protect bit (buffer[2] & 0x80).
>
> I'll give the next step a try and let you know in the morning (return
> mode page
> 1 and a good status).
>
> regards
> -Willy
>
> . I'll probably have to investigate the ALLOW MEDIUM REMOVAL
> command,
> but
> . the MODE SENSE command is a little puzzling... Can you provide a
> kernel
> . log of the command that gets sent to the driver? The driver is
> supposed
> . to handle one MODE SENSE command, but if there are more it should
>
> . handle, I should probably build those in also
> .
> . Thanks,
> .
> . --Rob
> .
> . Willy wrote:
> .
> . > Greetings:
> . >
> . > I hope this is of interest to Matthew Dharm...
> . >
> . > Playing with a recently arrived SanDisk SDDR09
> . > (box marking SDDR-09-01).
> . >
> . > It appears that this device does not want to see
> . > "ALLOW MEDIUM REMOVAL" or "MODE SENSE".
> . >
> . > With these commands delivered to the device, a timeout
> . > occurs. A modification to "usb.c" to eliminate these commands,
>
> . > in a manner similar to processing the START_STOP command,
> . > seems to eliminate the timeout. (adding flags to
> unusual_devs.h
> . > and usb.h in a manner similar to US_FL_START_STOP).
> . >
> . > Note that I've not attempted to provide any data to the
> MODE_SENSE
> . > request, just skipped it like START_STOP does...
> . >
> . > Kernal 2.4.2 pasted into Redhat "fisher"
> . >
> . > regards
> . > -Willy
> . > William Robison
>
>
>
>
>
> _______________________________________________
> [EMAIL PROTECTED]
> To unsubscribe, use the last form field at:
> http://lists.sourceforge.net/lists/listinfo/linux-usb-devel
--
Matthew Dharm Home: [EMAIL PROTECTED]
Maintainer, Linux USB Mass Storage Driver
Hi. I have my back hairs caught in my computer fan.
-- Customer
User Friendly, 8/20/1998
PGP signature
