Eek, that does not bear good. Somebody else mailed me this patch, so
would you mind sending me your orriginal patch for comparison?
And I assume that the scsi_unregister call should in fact not be in
there? If I understand it correctly there is either some function that
first calls adpt_release (from the address given in the info struct) and
then scsi_unregister, or scsi_unregister itself calls us.
Btw, none of this code was yet incorporated into the 2.6.3 kernel which
is why I started looking around for patches and actually looking at
making the changes myself. Btw, I also note that a couple of free's has
moved around - this gives me the idea that either a memory corruption
bug has been fixed (which leaves at least two of our machines in danger
of crashing) or a memory leak has been introduced. Or is it simply that
memory handling has been optimized a bit?
Jaco
Salyzyn, Mark wrote:
>My code fragment has (original multiversion and version in 2.4.24):
>
>/*
> * scsi_unregister will be called AFTER we return.
> */
>static int adpt_release(struct Scsi_Host *host)
>{
> adpt_hba* pHba = (adpt_hba*) host->hostdata[0];
>// adpt_i2o_quiesce_hba(pHba);
> adpt_i2o_delete_hba(pHba);
> return 0;
>}
>
>So I am unsure as to how *that* (the scsi_unregister call) got in there
>in the 2.6.3 stream. adpt_i2o_queisce was commented out in version 2.4.5
>of the driver (Prior to the `historical documents' and at least the
>in-box driver for RH7.3). It was not part of the 2.6.2 tree I based my
>patch on and not part of the submitted 2.6 patch for the dpt_i2o driver.
>
>adpt_i2o_quiesce tells the adapter to stop all activity, including
>builds, and is a blocking command with a 4 minute timeout.
>
>Sincerely -- Mark Salyzyn
>
>-----Original Message-----
>From: David Lang [mailto:[email protected]]
>Sent: Thursday, February 26, 2004 5:24 AM
>To: Salyzyn, Mark; Jaco Kroon
>Cc: Adrian Bunk; Linux Kernel Mailing List
>Subject: Re: 2.6.3 adaptec I2O will not compile
>
>Mark, do you have any comments on this?
>
>On Thu, 26 Feb 2004, Jaco Kroon wrote:
>
>
>
>>Date: Thu, 26 Feb 2004 01:28:14 -0800
>>From: Jaco Kroon <[email protected]>
>>To: David Lang <[email protected]>
>>Cc: Adrian Bunk <[email protected]>,
>> Linux Kernel Mailing List <[email protected]>
>>Subject: Re: 2.6.3 adaptec I2O will not compile
>>
>>David Lang wrote:
>>
>>
>>
>>>I received a post from Mark Salyzyn with a new driver that does
>>>
>>>
>>compile,
>>
>>
>>>he said that he has submitted the patches (I didn't get a chance to
>>>
>>>
>try
>
>
>>>the new kernel yet, I expect to do that shortly and will report any
>>>problems)
>>>
>>>D
>>>
>>>
>>>
>>Somebody just mailed me a patch too, but I found the following
>>discrepancy:
>>
>>/*
>> * scsi_unregister will be called AFTER we return.
>> */
>>static int adpt_release(struct Scsi_Host *host)
>>{
>> adpt_hba* pHba = (adpt_hba*) host->hostdata[0];
>>// adpt_i2o_quiesce_hba(pHba);
>> adpt_i2o_delete_hba(pHba);
>> scsi_unregister(host);
>> return 0;
>>}
>>
>>This is used to release the host, now read the comment, and then the
>>line just before the return. This line was added by the patch, which
>>also commented out the quiesce line. Is it possible to get any
>>confirmation on how this is supposed to function?
>>
>>Jaco
>>
>>===========================================
>>This message and attachments are subject to a disclaimer. Please refer
>>to http://www.it.up.ac.za/documentation/governance/disclaimer/ for full
>>details.
>>Hierdie boodskap en aanhangsels is aan 'n vrywaringsklousule
>>
>>
>onderhewig.
>
>
>>Volledige besonderhede is by
>>http://www.it.up.ac.za/documentation/governance/disclaimer/ beskikbaar.
>>===========================================
>>
>>
>>
>>
>
>
>
--
That that is is that that is not is not.
===========================================
This message and attachments are subject to a disclaimer. Please refer to http://www.it.up.ac.za/documentation/governance/disclaimer/ for full details.
Hierdie boodskap en aanhangsels is aan 'n vrywaringsklousule onderhewig. Volledige besonderhede is by http://www.it.up.ac.za/documentation/governance/disclaimer/ beskikbaar.
===========================================