Re: [linux-dvb] XC4000 patches for kernel 2.6.37.2

[Mauro Carvalho Chehab <mchehab@redhat.com>]

Em 02-06-2011 07:52, Devin Heitmueller escreveu:
> 2011/5/31 Dmitri Belimov <d.belimov@gmail.com>:
>> Is it possible make some patches and add support xc4000 into kernel?
>>
>> With my best regards, Dmitry.
> 
> What needs to happen here is somebody needs to prepare a patch series
> which contains all the relevant patches, including the SOBs.  This is
> entirely an janitorial task which can be done by anyone and frankly I
> don't have time for this sort of crap anymore.
> 
> Any volunteers?
> 
> All my patches have my SOB attached.  I explicitly got Davide's
> permission to add his SOB to his original patch, but it's not in the
> HG tree since I got the permission after I committed his change to my
> repo.  I can forward the email with his SOB so the person constructing
> the tree can add it on (as well as my SOB to preserve the chain of
> custody).
> 
> Secondly, we need to build a firmware image which is based off of the
> *actual* xceive firmware sources, so that we can be confident that all
> the blobs are from the same firmware revision and so that we can
> maintain them going forward.  I can provide them off-list to someone
> willing to do this work and testing.  Istann_v's firmware image is
> based off of i2c dumps and extracted by hand from disassembled
> firmware, which is less than ideal from an ongoing maintenance
> perspective.
> 
> And of course it's worth mentioning that the driver itself still needs
> a ton of cleanup, doesn't meet the coding standards, and wouldn't be
> accepted upstream in its current state.  Somebody will need to do the
> work to clean up the driver, as well as testing to make sure he/she
> didn't cause any regressions.
> 
> In summary, here are the four things that need to happen:
> 
> 1.  Assemble tree with current patches

It is probably easier for me to do this step, as I have my hg import
scripts. However, as I don't have the PCTV devices added at dib0700,
I can't test.

OK, I did this work, as it just took me a few minutes to rebase patches
1 and 2. I didn't apply the patches that started with "djh" since they
seemed to be a few hacks during the development time.

The tree is at:

git://linuxtv.org/mchehab/experimental.git branch xc4000

There are two warnings there that needs to be fixed:

drivers/media/common/tuners/xc4000.c:1293: warning: ‘xc4000_is_firmware_loaded’ defined but not used
drivers/media/common/tuners/xc4000.c: In function ‘check_firmware.clone.0’:
drivers/media/common/tuners/xc4000.c:1107: warning: ‘version’ may be used uninitialized in this function

Both seems to be trivial.

A disclaimer notice here: I didn't make any cleanup at the code,
(except by running a whitespace cleanup script) nor I've reviewed it. 

IMO, the next step is to test the rebases against a real hardware, 
and adding a few patches fixing it, if the rebases broke.

The next step would be fix the CodingStyle, and run checkpatch.pl.
There aren't many CodingStyle warnings/errors (13 errors, 28 warnings).
Most of the errors are due to the excess usage of printk's for debug,
and due to some obsolete code commented with //.

> 2.  Construct valid firmware image off of current sources
> 3.  Cleanup/coding style
> 4.  Testing
> 
> Now that we've got a bunch of people who are interested in seeing this
> upstream, who is going to volunteer to do which items in the above
> list?
> 
> Devin
>