Recent ATA changes break Thinkpad T40p

[Søren Schmidt]

Vaidas Damosevicius wrote:
> 
> On Sun, 22 Aug 2004, Søren Schmidt wrote:
> 
>>> This is *not* a controller problem, its because the CDROM drive has 
>>> buggy firmware that responds to both master and slave selection :(
>>
>>
>> OK, I have committed a workaround for those buggers to -current...
>>
>> (for those that uses 5.3 just take your ATA driver upto -current)
>>
> 
> Well, after cvsup and kernel rebuild I have the same problems:
> 
> ad0: WARNING - WRITE_DMA interrupt was seen but timeout fired LBA=52738515
> ad0: WARNING - WRITE_DMA interrupt was seen but timeout fired LBA=52738515
> ad0: WARNING - WRITE_DMA interrupt was seen but taskqueue stalled 
> LBA=52738515
> 
> ... and my system "hangs" - I can't login, I can't shutdown it - only 
> turn off :(

So you suffer from that "other" possibly non-ATA problem and not the 
"fake slave" one?

If you had provided a dmesg from the system, preferably from the kernel 
that fails, I could maybe tell you more...

Does backstepping only /sys/dev/ata to say aug 1st help ?(leave the rest 
of the kernel uptodate).

-Søren