[PULL] request for 2.6.38-rc1

[PULL] request for 2.6.38-rc1

[Patrick Boettcher]

Hi Mauro,

if it is not too late, here is a pull request for some new devices from 
DiBcom. It would be nice to have it in 2.6.38-rc1.

Pull from

git://linuxtv.org/pb/media_tree.git staging/for_2.6.38-rc1.dibcom

for

DiBxxxx: Codingstype updates
DiB0700: add support for several board-layouts
DiB7090: add support for the dib7090 based
DIB9000: initial support added
DiB0090: misc improvements
DiBx000: add addition i2c-interface names
DiB8000: add diversity support
DiB0700: add function to change I2C-speed

  drivers/media/dvb/dvb-usb/dib0700.h          |    2 +
  drivers/media/dvb/dvb-usb/dib0700_core.c     |   47 +-
  drivers/media/dvb/dvb-usb/dib0700_devices.c  | 1374 ++++++++++++++--
  drivers/media/dvb/dvb-usb/dvb-usb-ids.h      |    5 +
  drivers/media/dvb/frontends/Kconfig          |    8 +
  drivers/media/dvb/frontends/Makefile         |    1 +
  drivers/media/dvb/frontends/dib0090.c        | 1583 ++++++++++++++----
  drivers/media/dvb/frontends/dib0090.h        |   31 +
  drivers/media/dvb/frontends/dib7000p.c       | 1943 ++++++++++++++++------
  drivers/media/dvb/frontends/dib7000p.h       |   96 +-
  drivers/media/dvb/frontends/dib8000.c        |  833 ++++++----
  drivers/media/dvb/frontends/dib8000.h        |   20 +
  drivers/media/dvb/frontends/dib9000.c        | 2350 ++++++++++++++++++++++++++
  drivers/media/dvb/frontends/dib9000.h        |  131 ++
  drivers/media/dvb/frontends/dibx000_common.c |  279 +++-
  drivers/media/dvb/frontends/dibx000_common.h |  152 ++-
  16 files changed, 7487 insertions(+), 1368 deletions(-)


best regards,
--

Patrick Boettcher - Kernel Labs
http://www.kernellabs.com/

Re: [PULL] request for 2.6.38-rc1

[Mauro Carvalho Chehab]

Em 14-01-2011 12:51, Patrick Boettcher escreveu:
> Hi Mauro,
> 
> if it is not too late, here is a pull request for some new devices from DiBcom. It would be nice to have it in 2.6.38-rc1.
> 
> Pull from
> 
> git://linuxtv.org/pb/media_tree.git staging/for_2.6.38-rc1.dibcom
> 
> for
> 
> DiBxxxx: Codingstype updates


Not sure if this is by purpose, but you're changing all
msleep(10) into msleep(20). This sounds very weird for a
CodingStyle fix:

-	msleep(10);
+	msleep(20);

> DiB0700: add support for several board-layouts

Hmm...

+	if (request_firmware(&state->frontend_firmware, "dib9090.fw", &adap->dev->udev->dev)) {

Where's dib9090.fw firmware is available? The better is to submit a patch to linux-firmware
with the firmware binary, with some license that allows end-users to use it with your device
and distros/distro partners to re-distribute it. While here, please add also the other
dibcom firmwares.

Vendors are free to use their own legal text for it. There are several examples for it
at:

http://git.kernel.org/?p=linux/kernel/git/dwmw2/linux-firmware.git;a=blob_plain;f=WHENCE;hb=HEAD


Btw, there are two alignment errors (one at dib7000p, for some cases, aligned with 4 chars),
and another at dib8000, where all statements after an if are aligned with 3 tabs plus one space.
I'm fixing those issues, c/c you at the fix patches.

Cheers,
Mauro

Re: [PULL] request for 2.6.38-rc1

[Patrick Boettcher]

Hi Mauro,

On Sun, 16 Jan 2011, Mauro Carvalho Chehab wrote:

> Em 14-01-2011 12:51, Patrick Boettcher escreveu:
>> Hi Mauro,
>>
>> if it is not too late, here is a pull request for some new devices from DiBcom. It would be nice to have it in 2.6.38-rc1.
>>
>> Pull from
>>
>> git://linuxtv.org/pb/media_tree.git staging/for_2.6.38-rc1.dibcom
>>
>> for
>>
>> DiBxxxx: Codingstype updates
>
>
> Not sure if this is by purpose, but you're changing all
> msleep(10) into msleep(20). This sounds very weird for a
> CodingStyle fix:
>
> -	msleep(10);
> +	msleep(20);

I was as surprised as you when I saw that changed, but in fact it is a 
checkpatch-fix: it seems that checkpatch is warning about msleep or less 
than 20ms.

Maybe it is not the right fix to put them to msleep(20), but I think this 
is better than to do udelay(10000).

What do you think?


> +	if (request_firmware(&state->frontend_firmware, "dib9090.fw", &adap->dev->udev->dev)) {
>
> Where's dib9090.fw firmware is available? The better is to submit a patch to linux-firmware
> with the firmware binary, with some license that allows end-users to use it with your device
> and distros/distro partners to re-distribute it. While here, please add also the other
> dibcom firmwares.

The dib0700-firmware is already available through a license. The 
dib9090-firmware will come later. It'll take a moment before everything is 
ready.


> Vendors are free to use their own legal text for it. There are several examples for it
> at:
>
> http://git.kernel.org/?p=linux/kernel/git/dwmw2/linux-firmware.git;a=blob_plain;f=WHENCE;hb=HEAD
>
>
> Btw, there are two alignment errors (one at dib7000p, for some cases, aligned with 4 chars),
> and another at dib8000, where all statements after an if are aligned with 3 tabs plus one space.
> I'm fixing those issues, c/c you at the fix patches.

Nice, thank you.

best regards,
--

Patrick

Re: [PULL] request for 2.6.38-rc1

[Mauro Carvalho Chehab]

Em 19-01-2011 09:34, Patrick Boettcher escreveu:
> Hi Mauro,
> 
> On Sun, 16 Jan 2011, Mauro Carvalho Chehab wrote:
> 
>> Em 14-01-2011 12:51, Patrick Boettcher escreveu:
>>> Hi Mauro,
>>>
>>> if it is not too late, here is a pull request for some new devices from DiBcom. It would be nice to have it in 2.6.38-rc1.
>>>
>>> Pull from
>>>
>>> git://linuxtv.org/pb/media_tree.git staging/for_2.6.38-rc1.dibcom
>>>
>>> for
>>>
>>> DiBxxxx: Codingstype updates
>>
>>
>> Not sure if this is by purpose, but you're changing all
>> msleep(10) into msleep(20). This sounds very weird for a
>> CodingStyle fix:
>>
>> -    msleep(10);
>> +    msleep(20);
> 
> I was as surprised as you when I saw that changed, but in fact it is a checkpatch-fix: it seems that checkpatch is warning about msleep or less than 20ms.
> 
> Maybe it is not the right fix to put them to msleep(20), but I think this is better than to do udelay(10000).
> 
> What do you think?

Well, that is more a warning for the developer that the actual sleep time may/will vary, depending
on how CONFIG_HZ is set. See: http://lkml.org/lkml/2007/8/3/250

There's actually a new function that can be used to disable such warnings (also at linux/delay.h):
	usleep_range(unsigned long min, unsigned long max);

If the device is very sensible if sleeping for a long time, using usleep_range() is a good idea,
as it will rely on a different wake up implementation. Otherwise, if anything between 10-20ms is OK,
I would just use msleep(10).

>> +    if (request_firmware(&state->frontend_firmware, "dib9090.fw", &adap->dev->udev->dev)) {
>>
>> Where's dib9090.fw firmware is available? The better is to submit a patch to linux-firmware
>> with the firmware binary, with some license that allows end-users to use it with your device
>> and distros/distro partners to re-distribute it. While here, please add also the other
>> dibcom firmwares.
> 
> The dib0700-firmware is already available through a license. The dib9090-firmware will come later. It'll take a moment before everything is ready.

Hmm... on a quick search at the web, I found this:
	http://www.kernellabs.com/firmware/dib0700/README.dib0700

If those are the latest license terms, they are OK for linux-firmware tree submission. After your
OK, I'll be adding the dibcom firmwares to my linux-firmware tree and ask the upstream maintainer
to pull from it. It would be nice if dib9090 firmware could be licensed under similar terms, and
also be submitted to linux-firmware tree.

>> Vendors are free to use their own legal text for it. There are several examples for it
>> at:
>>
>> http://git.kernel.org/?p=linux/kernel/git/dwmw2/linux-firmware.git;a=blob_plain;f=WHENCE;hb=HEAD
>>
>>
>> Btw, there are two alignment errors (one at dib7000p, for some cases, aligned with 4 chars),
>> and another at dib8000, where all statements after an if are aligned with 3 tabs plus one space.
>> I'm fixing those issues, c/c you at the fix patches.
> 
> Nice, thank you.

Unfortunately, I did one error with the last patch. Accidentally, I folded a checkpatch.pl
"shut up" patch. Due to that, I missed the merge window. I'll be sending an upstream request
for pulling it on a separate branch, together with two other improvement patches (vb2 and ngene).
Let's hope that those will be accepted for .38.

Cheers,
Mauro