On 4 Jul, [EMAIL PROTECTED] wrote:
>>From: Patrick Ohly <[EMAIL PROTECTED]>
>
>>> >From: [EMAIL PROTECTED]
>>>
>>> >> As you see the March 99 version if sg.c implemented
>>> SG_SET_RESERVED_SIZE as NO-OP.
>>>
>>> >I don't know which sg version this file corresponds to. I only have the
>>>
>>> I told you: the march 99 version!
>
>>As I haven't followed the complete history (especially including
>>publishing dates) I am not able to directly relate this date to
>>a interface definition.
>
>
> It seems that you forgot what your original question has been.
> It was similar to:
>
> "Why did not libscg support this earlier"
No, it wasn't. Sorry if you perceived it like this. I didn't mean to
accuse you, but rater tried to explore the technical background and
limitations of libscg and/or sg.
> The version from March 99 _is_ not usable for libscg; this is the important
> comclusion. The interface from March 99 would force me to send a command
> with a high DMA count to the target to check if it would work.
I don't see it this way. It's probably true that this version didn't
actually support raising the DMA buffer size, but it provided a
guarantee that commands with less than a certain amount of data would
execute without problems due to insufficient buffer memory. You haven't
answered the question about this in my last mail, so I still don't know
if you disagree with me and Douglas here.
Actually setting the buffer was added later without requiring interface
changes (as far as I can tell), so apart from getting rid of the
problematic and out-dated usage of SG_BIG_BUFF to determine the maximum
DMA size at compile time, cdrecord could also have been prepared to use
this feature as soon as it was implemented.
I really don't want to start the discussion again why it wasn't done,
I'd rather like to know what technical reasons made it necessary to
disabled the new code again for Linux kernels older than 2.2.10?
>>I'll use open_scsi() with a NULL scsidev to do a bus scan.
>>How can I build a scsidev string to open another SCSI pointer to
>>access just one drive in a portable way?
>
> Read the cdrecord man page. It has a paragraph about the dev= parameter.
I know this man page and the parameter, but I don't know if there is a
way to figure out the devicename part of "devicename:scsibus,target"
while I do a bus scan, or if it is safe to omit it once a bus scan via
bus/target/lun was successful.
Bye, Patrick Ohly
--
[EMAIL PROTECTED] (new permanent address)
[EMAIL PROTECTED] (MakeCD related mails)
http://home.pages.de/~Ohly/
http://makecd.core.de/ (MakeCD home page)
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
