all the rest patches!

GIT PULL Antti Palosaari 2 participants 5 messages Sep 27, 2012 – Oct 1, 2012 1 message not in the archive Media (V4L/DVB)

lore

Share & tools

lore

Antti Palosaari

author

2012-09-27 22:15 UTC #0 lore raw

Mauro,
New attempt. I really want that "PCTV 520e workaround for DRX-K fw loading" in too or find out other fix quickly. I have answered too many bug reports according to it currently. Will take debugs now...

regards
Antti

The following changes since commit 8928b6d1568eb9104cc9e2e6627d7086437b2fb3:

[media] media: mx2_camera: use managed functions to clean up code
(2012-09-27 15:56:47 -0300)

are available in the git repository at:

git://linuxtv.org/anttip/media_tree.git for_v3.7_mauro-2

for you to fetch changes up to 2baf1e9dd547402b8a5748e66f894af9c6a2789a:

em28xx: PCTV 520e workaround for DRX-K fw loading (2012-09-28
01:06:38 +0300)

----------------------------------------------------------------
Antti Palosaari (4):
em28xx: implement FE set_lna() callback
cxd2820r: use static GPIO config when GPIOLIB is undefined
em28xx: do not set PCTV 290e LNA handler if fe attach fail
em28xx: PCTV 520e workaround for DRX-K fw loading

 drivers/media/dvb-frontends/cxd2820r_core.c | 29 
++++++++++++++++++++---------
 drivers/media/usb/em28xx/em28xx-dvb.c       | 61 
++++++++++++++++++++++++++++++++++++++++++++++++++-----------
 2 files changed, 70 insertions(+), 20 deletions(-)

Signature · 1 line

http://palosaari.fi/

A message that is not in the archive (sent off-list or not yet received): [email protected]

Antti Palosaari

author

2012-09-29 18:42 UTC #2 lore raw

Updated, one new USB ID for RTL2832U.

The following changes since commit 8928b6d1568eb9104cc9e2e6627d7086437b2fb3:

[media] media: mx2_camera: use managed functions to clean up code
(2012-09-27 15:56:47 -0300)

are available in the git repository at:

git://linuxtv.org/anttip/media_tree.git for_v3.7_mauro-3

for you to fetch changes up to bf342b50ac6c5801a95d6a089086587446c8d6cf:

rtl28xxu: [0ccd:00d3] TerraTec Cinergy T Stick RC (Rev. 3)
(2012-09-29 21:39:26 +0300)

----------------------------------------------------------------
Antti Palosaari (5):
em28xx: implement FE set_lna() callback
cxd2820r: use static GPIO config when GPIOLIB is undefined
em28xx: do not set PCTV 290e LNA handler if fe attach fail
em28xx: PCTV 520e workaround for DRX-K fw loading
rtl28xxu: [0ccd:00d3] TerraTec Cinergy T Stick RC (Rev. 3)

Gianluca Gennari (3):
fc2580: define const as UL to silence a warning
fc2580: silence uninitialized variable warning
fc2580: use macro for 64 bit division and reminder

 drivers/media/dvb-frontends/cxd2820r_core.c | 29 
++++++++++++++++++++---------
 drivers/media/tuners/fc2580.c               |  7 +++----
 drivers/media/usb/dvb-usb-v2/rtl28xxu.c     |  2 ++
 drivers/media/usb/em28xx/em28xx-dvb.c       | 61 
++++++++++++++++++++++++++++++++++++++++++++++++++-----------
 4 files changed, 75 insertions(+), 24 deletions(-)

Quoted 79 lines · Antti Palosaari

PULL request updated, contains now 3 fc2580 driver fixes from Gianluca Gennari. Could you handle these quickly :)
The following changes since commit
8928b6d1568eb9104cc9e2e6627d7086437b2fb3:
[media] media: mx2_camera: use managed functions to clean up code
(2012-09-27 15:56:47 -0300)
are available in the git repository at:
   git://linuxtv.org/anttip/media_tree.git for_v3.7_mauro-3
for you to fetch changes up to 5c3cc06e4922d2d6a050948181a8601281227c66:
fc2580: use macro for 64 bit division and reminder (2012-09-28
13:18:56 +0300)
----------------------------------------------------------------
Antti Palosaari (4):
em28xx: implement FE set_lna() callback
cxd2820r: use static GPIO config when GPIOLIB is undefined
em28xx: do not set PCTV 290e LNA handler if fe attach fail
em28xx: PCTV 520e workaround for DRX-K fw loading
Gianluca Gennari (3):
fc2580: define const as UL to silence a warning
fc2580: silence uninitialized variable warning
fc2580: use macro for 64 bit division and reminder
  drivers/media/dvb-frontends/cxd2820r_core.c | 29
++++++++++++++++++++---------
  drivers/media/tuners/fc2580.c               |  7 +++----
  drivers/media/usb/em28xx/em28xx-dvb.c       | 61
++++++++++++++++++++++++++++++++++++++++++++++++++-----------
  3 files changed, 73 insertions(+), 24 deletions(-)
Quoted 38 lines · Antti Palosaari
Mauro,
New attempt. I really want that "PCTV 520e workaround for DRX-K fw loading" in too or find out other fix quickly. I have answered too many bug reports according to it currently. Will take debugs now...
regards
Antti
The following changes since commit
8928b6d1568eb9104cc9e2e6627d7086437b2fb3:
[media] media: mx2_camera: use managed functions to clean up code
(2012-09-27 15:56:47 -0300)
are available in the git repository at:
   git://linuxtv.org/anttip/media_tree.git for_v3.7_mauro-2
for you to fetch changes up to 2baf1e9dd547402b8a5748e66f894af9c6a2789a:
   em28xx: PCTV 520e workaround for DRX-K fw loading (2012-09-28
01:06:38 +0300)
----------------------------------------------------------------
Antti Palosaari (4):
em28xx: implement FE set_lna() callback
cxd2820r: use static GPIO config when GPIOLIB is undefined
em28xx: do not set PCTV 290e LNA handler if fe attach fail
em28xx: PCTV 520e workaround for DRX-K fw loading
  drivers/media/dvb-frontends/cxd2820r_core.c | 29
++++++++++++++++++++---------
  drivers/media/usb/em28xx/em28xx-dvb.c       | 61
++++++++++++++++++++++++++++++++++++++++++++++++++-----------
  2 files changed, 70 insertions(+), 20 deletions(-)

Signature · 1 line

http://palosaari.fi/

Mauro Carvalho Chehab

2012-10-01 19:32 UTC #3 lore raw

Em Sat, 29 Sep 2012 21:41:41 +0300
Antti Palosaari <[email protected]> escreveu:

Quoted 16 lines

Updated, one new USB ID for RTL2832U.
The following changes since commit 8928b6d1568eb9104cc9e2e6627d7086437b2fb3:
[media] media: mx2_camera: use managed functions to clean up code
(2012-09-27 15:56:47 -0300)
are available in the git repository at:
   git://linuxtv.org/anttip/media_tree.git for_v3.7_mauro-3
for you to fetch changes up to bf342b50ac6c5801a95d6a089086587446c8d6cf:
rtl28xxu: [0ccd:00d3] TerraTec Cinergy T Stick RC (Rev. 3)
(2012-09-29 21:39:26 +0300)

Quoted 6 lines

----------------------------------------------------------------
Antti Palosaari (5):
em28xx: implement FE set_lna() callback
cxd2820r: use static GPIO config when GPIOLIB is undefined
em28xx: do not set PCTV 290e LNA handler if fe attach fail
em28xx: PCTV 520e workaround for DRX-K fw loading

All applied except for the above.

As I said before: sleeping for 2 seconds doesn't give any warranty that the firmware got loaded (and I know one system where firmware load generally takes more than 2 seconds to start - probably because the root fs is using nfs, and the machine uses an Atom single core processor).

As I've explained, if the driver needs to wait for a firmware load, it should use something that will actually wait for firmware load to complete, instead of just sleeping in the hope that the amount of sleeping time would be enough.

Regards,
Mauro

Antti Palosaari

author

2012-10-01 19:38 UTC #4 lore raw

Quoted 33 lines · Mauro Carvalho Chehab

Em Sat, 29 Sep 2012 21:41:41 +0300
Antti Palosaari <[email protected]> escreveu:
Quoted 16 lines
Updated, one new USB ID for RTL2832U.
The following changes since commit 8928b6d1568eb9104cc9e2e6627d7086437b2fb3:
[media] media: mx2_camera: use managed functions to clean up code
(2012-09-27 15:56:47 -0300)
are available in the git repository at:
    git://linuxtv.org/anttip/media_tree.git for_v3.7_mauro-3
for you to fetch changes up to bf342b50ac6c5801a95d6a089086587446c8d6cf:
rtl28xxu: [0ccd:00d3] TerraTec Cinergy T Stick RC (Rev. 3)
(2012-09-29 21:39:26 +0300)
Quoted 6 lines
----------------------------------------------------------------
Antti Palosaari (5):
em28xx: implement FE set_lna() callback
cxd2820r: use static GPIO config when GPIOLIB is undefined
em28xx: do not set PCTV 290e LNA handler if fe attach fail
em28xx: PCTV 520e workaround for DRX-K fw loading
All applied except for the above.
As I said before: sleeping for 2 seconds doesn't give any warranty that the firmware got loaded (and I know one system where firmware load generally takes more than 2 seconds to start - probably because the root fs is using nfs, and the machine uses an Atom single core processor).

Quoted 4 lines

As I've explained, if the driver needs to wait for a firmware load, it should use something that will actually wait for firmware load to complete, instead of just sleeping in the hope that the amount of sleeping time would be enough.

Yes I certainly know. My aim was to provide working driver around 99% of users even it is not correct. I don't understand why you all the time reject this kind of workarounds leaving devices 100% non-working state. And as it is the device what I am mostly responsible, I will get all the angry feedback from the unhappy users as their device does not work anymore after the Kernel upgrade.

regards
Antti

Signature · 1 line

http://palosaari.fi/

Mauro Carvalho Chehab

2012-10-01 19:54 UTC #5 lore raw

Em Mon, 01 Oct 2012 22:38:07 +0300
Antti Palosaari <[email protected]> escreveu:

Quoted 34 lines · Mauro Carvalho Chehab

Quoted 33 lines
Em Sat, 29 Sep 2012 21:41:41 +0300
Antti Palosaari <[email protected]> escreveu:
Quoted 16 lines
Updated, one new USB ID for RTL2832U.
The following changes since commit 8928b6d1568eb9104cc9e2e6627d7086437b2fb3:
[media] media: mx2_camera: use managed functions to clean up code
(2012-09-27 15:56:47 -0300)
are available in the git repository at:
    git://linuxtv.org/anttip/media_tree.git for_v3.7_mauro-3
for you to fetch changes up to bf342b50ac6c5801a95d6a089086587446c8d6cf:
rtl28xxu: [0ccd:00d3] TerraTec Cinergy T Stick RC (Rev. 3)
(2012-09-29 21:39:26 +0300)
Quoted 6 lines
----------------------------------------------------------------
Antti Palosaari (5):
em28xx: implement FE set_lna() callback
cxd2820r: use static GPIO config when GPIOLIB is undefined
em28xx: do not set PCTV 290e LNA handler if fe attach fail
em28xx: PCTV 520e workaround for DRX-K fw loading
All applied except for the above.
As I said before: sleeping for 2 seconds doesn't give any warranty that the firmware got loaded (and I know one system where firmware load generally takes more than 2 seconds to start - probably because the root fs is using nfs, and the machine uses an Atom single core processor).

Quoted 8 lines

Quoted 4 lines
As I've explained, if the driver needs to wait for a firmware load, it should use something that will actually wait for firmware load to complete, instead of just sleeping in the hope that the amount of sleeping time would be enough.
Yes I certainly know. My aim was to provide working driver around 99% of users even it is not correct. I don't understand why you all the time reject this kind of workarounds leaving devices 100% non-working state.

Because:

1) this is a hack;

2) there's absolutely no warranties that it will fix the issue[1];

3) the real fix is not complex; I've pointed you what should be changed in order to use a wait queue, that will wait for the firmware load to complete[2].

[1] the system I'm aware of where this hack won't work is a reference kit from a major CPU vendor, to be used on media centers. It doesn't sound a corner case for me.

[2] if you don't agree with my proposal, you're free, of course, to come with some other alternative, like:
- adding some callback to report when a firmware load task is completed;
- defer the tuner attach work to happen after firmware load.

Adding a code there that we know it will cause us future headaches because it is broken by design doesn't sound nice.

Quoted 3 lines

And as it is the device what I am mostly responsible, I will get all the angry feedback from the unhappy users as their device does not work anymore after the Kernel upgrade.

If you don't have time to work on that, I can try to help you after the end of the merge window.

Regards,
Mauro