Re: Problem accessing Sandisk CompactFlash Cards (Connected to the IDE bus)

[Neil Horman <nhorman@redhat.com>]

Marc Strämke wrote:
> Marcelo Tosatti wrote:
> 
>> Indeed this is a typo but has been fixed on 2.4.26:
>>
>>         if (drive->removable && id != NULL) {
> 
> It never gets past this check because drive->removable is not set!
Theres actually a more interesting difference to notice:

In the working dmesg output I see this:
 > hda: SanDisk SDCFB-128, CFA DISK drive

While in the non working dmesg output we have:
 > hdb: SanDisk SDCFB-128, ATA DISK drive

Tracing back through the code it looks to me like we get the ATA disk 
print in the event that this test in do_identify:
/*
  * Check for an ATAPI device
  */
  if (cmd == WIN_PIDENTIFY) {

that would explain why the drive_is_flashcard test is getting skipped, 
why setting removable is making no difference, and why your card is 
being identified as an ATA device.  It looks as though the WIN_PIDENTIFY 
command is sent down to this routine from ide_probe_for_drive in this 
snip of code:

/* if !(success||timed-out) */
                 if (do_probe(drive, WIN_IDENTIFY) >= 2) {
                         /* look for ATAPI device */
                         (void) do_probe(drive, WIN_PIDENTIFY);
                 }

So it would seem that WIN_PIDENTIFY is issued only if a WIN_IDENTIFY 
command fails with an rc greater than 2.  I would suggest instrumenting 
this area of the code to see what the WIN_IDENTIFY command is returning 
on the working and non-working systems.  I'm betting you will find a 
difference, and we'll be able to track down the problem from there.

HTH
Neil
-- 
/***************************************************
  *Neil Horman
  *Software Engineer
  *Red Hat, Inc.
  *nhorman@redhat.com
  *gpg keyid: 1024D / 0x92A74FA1
  *http://pgp.mit.edu
  ***************************************************/