How I must report that a driver has been broken?

How I must report that a driver has been broken?

[Alfredo Jesús Delaiti]

Hi

New features of the driver has left a card does not work.
How I must report that a driver has been broken?

Thanks

-- 
Dona tu voz
http://www.voxforge.org/es

Re: How I must report that a driver has been broken?

[Hans de Goede]

Hi

On 05/12/2012 06:26 AM, Alfredo Jesús Delaiti wrote:
> Hi
>
> New features of the driver has left a card does not work.
> How I must report that a driver has been broken?

Well this list would be a good place for starters, please
send a *detailed* bug report to this list, including
things like what is the last (kernel) version it worked
with, what is the first version it is broken.

What did it do before breaking what know now longer works,
etc.

Regards,

Hans

Re: How I must report that a driver has been broken?

[Alfredo Jesús Delaiti]

Hi

Thanks for your response Hans and Patrick

Maybe I doing wrong this, because it reports twice:

http://www.mail-archive.com/linux-media@vger.kernel.org/msg45199.html
http://www.mail-archive.com/linux-media@vger.kernel.org/msg44846.html

and I have not had any response.


Thanks again,

Alfredo





El 12/05/12 05:17, Hans de Goede escribió:
> Hi
>
> On 05/12/2012 06:26 AM, Alfredo Jesús Delaiti wrote:
>> Hi
>>
>> New features of the driver has left a card does not work.
>> How I must report that a driver has been broken?
>
> Well this list would be a good place for starters, please
> send a *detailed* bug report to this list, including
> things like what is the last (kernel) version it worked
> with, what is the first version it is broken.
>
> What did it do before breaking what know now longer works,
> etc.
>
> Regards,
>
> Hans


-- 
Dona tu voz
http://www.voxforge.org/es

Re: How I must report that a driver has been broken?

[Patrick Dickey]

I'm not an expert (or anywhere near one), but I'm guessing that the
reason you didn't get a response to your two previous (actually three or
four previous) emails is because you "submitted" a patch that was
already in the kernel.

You were intending to show that the patch breaks something, yet instead
of submitting the fix for the problem, you resubmitted the patch that
was already applied.  Either that, or for some other reason, the patch
that you submitted was rejected.

My suggestion is to take what you've submitted (if it was the fix for
the problem) and resubmit it. If you were just pointing out the broken
code, then submit your fix for the issue (by creating a patch that fixes
or removes the broken code).

Also one other thing I found is that you submitted the original patches
for the card.  Or at least you submitted some patches for it around
November of last year. So my question is, do the patches that you
submitted back then work, or are they the broken code now?

Have a great day:)
Patrick.

P.S. Again, I'm not an expert here, nor do I claim to be. So if someone
else gives you an answer, I'd put more weight on theirs--as they
probably have more experience than I do.


On Sat, 2012-05-12 at 16:26 -0300, Alfredo Jesús Delaiti wrote: 
> Hi
> 
> Thanks for your response Hans and Patrick
> 
> Maybe I doing wrong this, because it reports twice:
> 
> http://www.mail-archive.com/linux-media@vger.kernel.org/msg45199.html
> http://www.mail-archive.com/linux-media@vger.kernel.org/msg44846.html
> 
> and I have not had any response.
> 
> 
> Thanks again,
> 
> Alfredo
> 
> 
> 
> 
> 
> El 12/05/12 05:17, Hans de Goede escribió:
> > Hi
> >
> > On 05/12/2012 06:26 AM, Alfredo Jesús Delaiti wrote:
> >> Hi
> >>
> >> New features of the driver has left a card does not work.
> >> How I must report that a driver has been broken?
> >
> > Well this list would be a good place for starters, please
> > send a *detailed* bug report to this list, including
> > things like what is the last (kernel) version it worked
> > with, what is the first version it is broken.
> >
> > What did it do before breaking what know now longer works,
> > etc.
> >
> > Regards,
> >
> > Hans
> 
> 

Re: How I must report that a driver has been broken?

[Hans de Goede]

Hi,

On 05/12/2012 09:26 PM, Alfredo Jesús Delaiti wrote:
> Hi
>
> Thanks for your response Hans and Patrick
>
> Maybe I doing wrong this, because it reports twice:
>
> http://www.mail-archive.com/linux-media@vger.kernel.org/msg45199.html
> http://www.mail-archive.com/linux-media@vger.kernel.org/msg44846.html

In your last message you indicate that you've found the patch causing it,
and that you were looking into figuring which bit of the patch actually
breaks things, so I guess people reading the thread were / are
waiting for you to follow up on it with the results of your attempts
to further isolate the cause.

What I were do if I were you is send a mail directly to the author
of the patch causing the problems, with what you've discovered
about the problem sofar in there, and put the list in the CC.

Regards,

Hans

Re: How I must report that a driver has been broken?

[Devin Heitmueller]

On Sun, May 13, 2012 at 1:50 AM, Hans de Goede <hdegoede@redhat.com> wrote:
> Hi,
>
>
> On 05/12/2012 09:26 PM, Alfredo Jesús Delaiti wrote:
>>
>> Hi
>>
>> Thanks for your response Hans and Patrick
>>
>> Maybe I doing wrong this, because it reports twice:
>>
>> http://www.mail-archive.com/linux-media@vger.kernel.org/msg45199.html
>> http://www.mail-archive.com/linux-media@vger.kernel.org/msg44846.html
>
>
> In your last message you indicate that you've found the patch causing it,
> and that you were looking into figuring which bit of the patch actually
> breaks things, so I guess people reading the thread were / are
> waiting for you to follow up on it with the results of your attempts
> to further isolate the cause.
>
> What I were do if I were you is send a mail directly to the author
> of the patch causing the problems, with what you've discovered
> about the problem sofar in there, and put the list in the CC.
>
> Regards,
>
> Hans

Steven loaned me his HVR-1850 board last week, and I'm hoping to debug
the regression this week (I have an HVR-1800 that is also effected).
I suspect the problem is related to a codepath for the cx23888's
onboard DIF being executed for 885 based boards.  Steven did a whole
series of patches to make the cx23888 work properly and I think a
regression snuck in there.  Simply backing out the change isn't the
correct fix.

I've got all the boards and the datasheets - I just need to find a bit
of time to get the current tree installed onto a machine and plug in
the various boards...

In short, it's in my queue so please be patient.

Devin

-- 
Devin J. Heitmueller - Kernel Labs
http://www.kernellabs.com

Re: How I must report that a driver has been broken?

[Alfredo Jesús Delaiti]

Hi

Thank you all for your responses.

Devin, I appreciate the time and labor you do to revise the code.

My previous letters maybe I can help you see where the problem and the 
date you began.
I thought of a patch of this type:

if (card != mycard) {

"bad code for my card"}

but unfortunately not so easy for me.


I will be patient.

Thank you.


Alfredo


El 14/05/12 01:51, Devin Heitmueller escribió:
> On Sun, May 13, 2012 at 1:50 AM, Hans de Goede<hdegoede@redhat.com>  wrote:
>> Hi,
>>
>>
>> On 05/12/2012 09:26 PM, Alfredo Jesús Delaiti wrote:
>>> Hi
>>>
>>> Thanks for your response Hans and Patrick
>>>
>>> Maybe I doing wrong this, because it reports twice:
>>>
>>> http://www.mail-archive.com/linux-media@vger.kernel.org/msg45199.html
>>> http://www.mail-archive.com/linux-media@vger.kernel.org/msg44846.html
>>
>> In your last message you indicate that you've found the patch causing it,
>> and that you were looking into figuring which bit of the patch actually
>> breaks things, so I guess people reading the thread were / are
>> waiting for you to follow up on it with the results of your attempts
>> to further isolate the cause.
>>
>> What I were do if I were you is send a mail directly to the author
>> of the patch causing the problems, with what you've discovered
>> about the problem sofar in there, and put the list in the CC.
>>
>> Regards,
>>
>> Hans
> Steven loaned me his HVR-1850 board last week, and I'm hoping to debug
> the regression this week (I have an HVR-1800 that is also effected).
> I suspect the problem is related to a codepath for the cx23888's
> onboard DIF being executed for 885 based boards.  Steven did a whole
> series of patches to make the cx23888 work properly and I think a
> regression snuck in there.  Simply backing out the change isn't the
> correct fix.
>
> I've got all the boards and the datasheets - I just need to find a bit
> of time to get the current tree installed onto a machine and plug in
> the various boards...
>
> In short, it's in my queue so please be patient.
>
> Devin
>


-- 
Dona tu voz
http://www.voxforge.org/es

Re: How I must report that a driver has been broken?

[Devin Heitmueller]

On Fri, May 18, 2012 at 10:15 AM, Alfredo Jesús Delaiti
<alfredodelaiti@netscape.net> wrote:
> Hi
>
> Thank you all for your responses.
>
> Devin, I appreciate the time and labor you do to revise the code.
>
> My previous letters maybe I can help you see where the problem and the date
> you began.
> I thought of a patch of this type:
>
> if (card != mycard) {
>
> "bad code for my card"}
>
> but unfortunately not so easy for me.

Some initial analysis of the driver code I did last night suggests
it's much more complicated than that (in addition to the HVR-1850
support there was a bunch of refactoring done to the both the cx23885
and cx25840 drivers).

You can keep an eye on http://www.kernellabs.com/blog for updates.

Devin

-- 
Devin J. Heitmueller - Kernel Labs
http://www.kernellabs.com