Re: non aligned DMA transfer attempted

Alexander Motin wrote:
> Ryan Rogers wrote:
>> Alexander Motin wrote:
>>> Ryan Rogers wrote:
>>>> I'm having a bit of a problem getting my DVD drive(s) to work 
>>>> correctly.   I'm trying to transfer my DVD collection to my media 
>>>> server, but whenever I run vobcopy, /var/log/messages gets spammed 
>>>> with:
>>>>
>>>> acd0: FAILURE - non aligned DMA transfer attempted
>>>> acd0: setting up DMA failed
>>>>
>>>> I added a bit more information to the first message to see if I 
>>>> could figure out what was actually going on.  request->data was 
>>>> 0xd40e0c37, ch->dma.alignment was 2, and request->bytecount was 2048.
>>>
>>> Actually I don't understand what for this check was made there. It is 
>>> busdma infrastructure business to implement buffer bouncing to manage 
>>> requested alignment. But this check enforces application level to 
>>> bother with this. Usually it works fine, as memory often allocated 
>>> aligned. But probably here is some specifics in your application. 
>>> Could you try to just to comment-out that request->data check?
>>
>> I commented out that part of the check, and it worked in the sense 
>> that it didn't spit out any errors at me, but I'm not 100% certain 
>> that the data isn't getting corrupted.  I tried ripping two discs.  
>> The first is completely corrupted, the second appears to be fine on 
>> first glance. I'm going to try a couple more discs to see if maybe the 
>> first one was just bad.
> 
> Please.
> 
>> Also, one other thing that I noticed that seems odd is this:
>>
>> acd0: DVDR <TSSTcorp CDDVDW SH-S223L/SB02> at ata6-master SATA150
>> ....
>> cd0: <TSSTcorp CDDVDW SH-S223L SB02> Removable CD-ROM SCSI-0 device
>> cd0: 3.300MB/s transfers
>>
>> I timed how long the first disc took to rip, and it seems to match 
>> that reported speed.  "Working but slow" is certainly better than "not 
>> working", but my speeds in Windows using DVD Decrypter on the same 
>> disk are at least 3 times faster, bursting up to about 5 times faster.
> 
> It's just a small atapicam misfeature. It doesn't translates SATA modes 
> to CAM speeds. It's only cosmetics, it shouldn't be important.
> 

I've ripped a handful of DVDs since my last post, and I haven't had any 
error messages, and the output appears to be correct.  I even re-ripped 
the one that had originally been corrupted, and it appears to be fine as 
well.  So, it appears removing that part of the check fixed the issue 
for me.

Ryan