Re: libata errors

[Robert Hancock <hancockrwd@gmail.com>]

Jeff Garzik wrote:
> Tejun Heo wrote:
>> Hello,
>>
>> Robert Hancock wrote:
>>> I tend to doubt that the driver is doing the wrong thing.
>>
>> I've tested everything including recording and cd ripping and most of
>> my devices seem happy but you never know.
>>
>>> I suspect it's more that ATAPI devices on SATA are about as
>>> standards-compliant in general as ATAPI devices on PATA (which is to
>>> say not very good) and some of them break if the controller does
>>> something slightly differently or with different timing, which is
>>> quite possible in AHCI mode.
>>>
>>> Tejun was looking at some SATA analyzer traces of a few drives that were
>>> apparently failing consistently on AHCI. Tejun, did you get anywhere
>>> with this?
>>
>> I need to take a deeper look at that but in general I couldn't find
>> anything standing out.  After a series of ATAPI changes (IIRC, the DMA
>> padding change was the last thing), most ATAPI bug reports got
>> resolved and I'm not getting many ATAPI failure bug reports other than
>> recent ahci ones anymore.  I'm not sure whethre it indicates that we
>> finally got it right or something else tho.
>>
>> Jeff, what kind of failure are you seeing?
> 
> Hardware:  SATA DVD-ROM/CD-RW devices, AHCI, ICH$foo
> 
> Reading and writing data CDs and DVDs works fine.  Ripping audio CDs 
> using gRip application leads to timeouts almost immediately.

Well, it's a potentially unusual transfer length compared to other 
commands, which might be part of it. Do you have some error output? It 
would be useful to look at the reported failed CDB.