[pull] http://linuxtv.org/hg/~tap/v4l-dvb
Mauro, Please pull from http://linuxtv.org/hg/~tap/v4l-dvb for the following 4 changesets: 01/04: compat: Add DMA_BIT_MASK() macro http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=486add0e3f1f 02/04: zoran: fix bug when enumerating format -1 http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=f144948f22c8 03/04: v4l2-ioctl: Check buffer types using g_fmt instead of try_fmt http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=b54857195e2c 04/04: v4l2-ioctl: Clear buffer type specific trailing fields/padding http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=7cbf08aeb840 linux/drivers/media/video/v4l2-ioctl.c | 45 - linux/drivers/media/video/zoran/zoran_driver.c | 30 +++- v4l/compat.h |7 ++- 3 files changed, 55 insertions(+), 27 deletions(-) Thanks, Trent -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
[pull] http://linuxtv.org/hg/~tap/v4l-dvb
Mauro, Please pull from http://linuxtv.org/hg/~tap/v4l-dvb for the following 14 changesets: 01/14: build: Fix kernel output directory support http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=346bab8698ea 02/14: v4l2-ioctl: Check format for S_PARM and G_PARM http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=44632c5fe5b2 03/14: saa7146: Remove buffer type check from vidioc_g_parm http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=c879e8f8c32b 04/14: bttv: Remove buffer type check from vidioc_g_parm http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=935d095cbc31 05/14: gspca: Stop setting buffer type, and avoid memset in querycap http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=089aaa41d473 06/14: omap24xxcam: Remove buffer type check from vidioc_s/g_parm http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=c545d5d1afcc 07/14: stkwebcam: Remove buffer type check from g_parm and q/dq/reqbufs http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=fb30bf07ca40 08/14: vino: Remove code for things already done by video_ioctl2 http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=900a6e99e7b0 09/14: cafe_ccic: Remove buffer type check from XXXbuf http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=e698f5e2e1db 10/14: cx23885-417: Don't need to zero ioctl parameter fields http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=76e7154479e4 11/14: cx88-blackbird: Stop setting buffer type in XXX_fmt_vid_cap http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=ea402dd306a6 12/14: meye: Remove buffer type checks from XXX_fmt_vid_cap, XXXbuf http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=1d2ae100255d 13/14: usbvision: Remove buffer type checks from enum_fmt_vid_cap, XXXbuf http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=5dd6945acfc3 14/14: zr364xx: Remove code for things already done by video_ioctl2 http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=406c54139ea6 linux/drivers/media/common/saa7146_video.c|4 linux/drivers/media/video/bt8xx/bttv-driver.c |3 linux/drivers/media/video/cafe_ccic.c | 14 linux/drivers/media/video/cx23885/cx23885-417.c | 29 - linux/drivers/media/video/cx88/cx88-blackbird.c |5 linux/drivers/media/video/gspca/gspca.c |5 linux/drivers/media/video/meye.c | 32 - linux/drivers/media/video/omap24xxcam.c |7 linux/drivers/media/video/stk-webcam.c| 11 linux/drivers/media/video/usbvision/usbvision-video.c | 10 linux/drivers/media/video/v4l2-ioctl.c|7 linux/drivers/media/video/vino.c | 335 +++--- linux/drivers/media/video/zr364xx.c | 16 v4l/Makefile |3 14 files changed, 158 insertions(+), 323 deletions(-) Thanks, Trent -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
[pull] http://linuxtv.org/hg/~tap/v4l-dvb
Mauro, Please pull from http://linuxtv.org/hg/~tap/v4l-dvb for the following changeset: 01/01: build: have make_kconfig.pl ignore comments http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=9debb0a2ec70 make_kconfig.pl | 13 + 1 files changed, 9 insertions(+), 4 deletions(-) Thanks, Trent -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
[pull] http://linuxtv.org/hg/~tap/v4l-dvb-zoran
Mauro, Please pull from http://linuxtv.org/hg/~tap/v4l-dvb-zoran for the following 6 changesets: 01/06: build: Clean up FM801-TEA575x Kconfig http://linuxtv.org/hg/~tap/v4l-dvb-zoran?cmd=changeset;node=a9792eb3e828 02/06: zoran: Unify buffer descriptors http://linuxtv.org/hg/~tap/v4l-dvb-zoran?cmd=changeset;node=41592ce6cba3 03/06: zoran: Drop the lock_norm module parameter http://linuxtv.org/hg/~tap/v4l-dvb-zoran?cmd=changeset;node=da9f9c766f07 04/06: zoran: Don't frighten users with failed buffer allocation http://linuxtv.org/hg/~tap/v4l-dvb-zoran?cmd=changeset;node=f0a275bde923 05/06: zoran: Pass zoran_fh pointers instead of file pointers http://linuxtv.org/hg/~tap/v4l-dvb-zoran?cmd=changeset;node=9eb7b94aa6b6 06/06: zoran: replace functions names in strings with __func__ http://linuxtv.org/hg/~tap/v4l-dvb-zoran?cmd=changeset;node=60b0989f6c7a linux/Documentation/video4linux/Zoran |3 linux/drivers/media/video/zoran/zoran.h| 59 - linux/drivers/media/video/zoran/zoran_card.c | 115 -- linux/drivers/media/video/zoran/zoran_device.c | 20 linux/drivers/media/video/zoran/zoran_device.h |2 linux/drivers/media/video/zoran/zoran_driver.c | 1357 ++--- v4l/Kconfig.sound |9 7 files changed, 669 insertions(+), 896 deletions(-) Thanks, Trent -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
[PULL] http://linuxtv.org/hg/~tap/v4l-dvb
Mauro, Please pull from http://linuxtv.org/hg/~tap/v4l-dvb for the following 2 changesets: 01/02: compat: Add INIT_DELAYED_WORK http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=f6382daee1df 02/02: zoran: Change first argument to zoran_v4l2_buffer_status http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=784ecbdb42a5 linux/drivers/media/video/zoran/zoran_driver.c |9 - v4l/compat.h |1 + 2 files changed, 5 insertions(+), 5 deletions(-) Thanks, Trent -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
[PULL] http://linuxtv.org/hg/~tap/v4l-dvb
Mauro, Please pull from http://linuxtv.org/hg/~tap/v4l-dvb for the following 4 changesets: 01/04: videodev: not possible to register NULL video_device http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=3c87763acae5 02/04: cx88: remove unnecessary forward declaration of cx88_core http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=ffc649bf14a6 03/04: compat: rename struct delayed_work to work_struct on pre-2.6.20 http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=62d29b435438 04/04: v4l2: Move code to zero querybuf output struct to v4l2_ioctl http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=d31e6e7a8241 linux/drivers/media/dvb/b2c2/flexcop-pci.c|4 linux/drivers/media/dvb/dvb-usb/dvb-usb.h |4 linux/drivers/media/radio/radio-si470x.c |4 linux/drivers/media/video/cx88/cx88.h |1 - linux/drivers/media/video/em28xx/em28xx.h |4 linux/drivers/media/video/meye.c |4 linux/drivers/media/video/stk-webcam.c|8 +--- linux/drivers/media/video/usbvision/usbvision-video.c |3 --- linux/drivers/media/video/v4l2-dev.c |7 +++ linux/drivers/media/video/v4l2-ioctl.c|5 + linux/drivers/media/video/zoran/zoran_driver.c|6 -- v4l/compat.h |8 12 files changed, 17 insertions(+), 41 deletions(-) Thanks, Trent -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PULL] http://linuxtv.org/hg/~tap/v4l-dvb
On Tue, 3 Mar 2009 13:59:27 -0800 (PST) Trent Piepho xy...@speakeasy.org wrote: 04/04: v4l2: Move code to zero querybuf output struct to v4l2_ioctl http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=d31e6e7a8241 Hmm.. The idea of using this memset seemed interesting: f02233e4 (Trent Piepho 2009-03-03 20:44:45 -0300 974) memset(0, p-bytesused, f02233e4 (Trent Piepho 2009-03-03 20:44:45 -0300 975) sizeof(*p) - offsetof(typeof(*p), bytesused)); But the end result is that, when compiling it on my -git tree, to have those errors with i386: drivers/media/video/v4l2-ioctl.c: In function ‘__video_do_ioctl’: drivers/media/video/v4l2-ioctl.c:974: warning: cast from pointer to integer of different size drivers/media/video/v4l2-ioctl.c:974: warning: cast from pointer to integer of different size drivers/media/video/v4l2-ioctl.c:974: warning: passing argument 2 of ‘__memset_generic’ makes integer from pointer without a cast drivers/media/video/v4l2-ioctl.c:974: warning: passing argument 2 of ‘__memset_generic’ makes integer from pointer without a cast Could you please provide us a fix? Cheers, Mauro -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PULL] http://linuxtv.org/hg/~tap/v4l-dvb
On Thu, 5 Feb 2009 10:57:21 -0800 (PST) Trent Piepho xy...@speakeasy.org wrote: Mauro, Please pull from http://linuxtv.org/hg/~tap/v4l-dvb for the following 3 changesets: 01/03: cx88: Fix MPEG kconfig dependencies http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=2e05fd69e6b6 Hmm... This patch converts a depends on into a select. This is not nice, since select is known to not be properly handled by kernel build system. Also, the kbuild maintainer recommended us to use select only for options that doesn't have any dependency. This is reflected on Kbuild docs: Note: select should be used with care. select will force a symbol to a value without visiting the dependencies. By abusing select you are able to select a symbol FOO even if FOO depends on BAR that is not set. In general use select only for non-visible symbols (no prompts anywhere) and for symbols with no dependencies. That will limit the usefulness but on the other hand avoid the illegal configurations all over. kconfig should one day warn about such things. So, IMO, we should solve the issue using a different logic that relies only on depends on. Cheers, Mauro -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PULL] http://linuxtv.org/hg/~tap/v4l-dvb
On Mon, 9 Feb 2009, Mauro Carvalho Chehab wrote: On Thu, 5 Feb 2009 10:57:21 -0800 (PST) Trent Piepho xy...@speakeasy.org wrote: 01/03: cx88: Fix MPEG kconfig dependencies http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=2e05fd69e6b6 Hmm... This patch converts a depends on into a select. This is not nice, since select is known to not be properly handled by kernel build system. Also, the kbuild maintainer recommended us to use select only for options that doesn't have any dependency. This is reflected on Kbuild docs: Note: select should be used with care. select will force a symbol to a value without visiting the dependencies. By abusing select you are able to select a symbol FOO even if FOO depends on BAR that is not set. In general use select I'm aware of the problem with select. One way to deal with is to have whatever is selecting FOO also depend on BAR. In this case CX88_MPEG only has one dependency, VIDEO_CX88. Both CX88_BLACKBIRD and CX88_DVB already depend on VIDEO_CX88 so there is no problem. Since CX88_MPEG is a hidden option and it's highly unlikely it will gain a dependency that CX88_BLACKBIRD and CX88_DVB do not also also have, I think this use of select is not so bad. So, IMO, we should solve the issue using a different logic that relies only on depends on. This can be done, it was the other option I mentioned. MPEG needs to be made a visible option, which BLACKBIRD and DVB will then depend on instead of selecting. I didn't do this because of complaints about kernel configuration getting too complex with too many options. So I didn't think a patch exposing another option would be better. It would look something like this: config VIDEO_CX88_MPEG tristate MPEG and digital TV interface depends on VIDEO_CX88 help Support for the MPEG port on Conexant CX2388x chips, which shows up as function 2 of the PCI device. This is used by DVB/ATSC digital TV cards as well as cards with the Conexant Blackbird CX23416 MPEG encoder. config VIDEO_CX88_MPEG tristate DVB/ATSC Support for cx2388x based TV cards depends on VIDEO_CX88_MPEG ... config VIDEO_CX88_BLACKBIRD tristate Blackbird MPEG encoder support (cx2388x + cx23416) depends on VIDEO_CX88_MPEG ... Would that be better? No select, but there is another option for the user to pick when configuring their cards. -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PULL] http://linuxtv.org/hg/~tap/v4l-dvb
On Mon, 9 Feb 2009 16:48:27 -0800 (PST) Trent Piepho xy...@speakeasy.org wrote: On Mon, 9 Feb 2009, Mauro Carvalho Chehab wrote: On Thu, 5 Feb 2009 10:57:21 -0800 (PST) Trent Piepho xy...@speakeasy.org wrote: 01/03: cx88: Fix MPEG kconfig dependencies http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=2e05fd69e6b6 Hmm... This patch converts a depends on into a select. This is not nice, since select is known to not be properly handled by kernel build system. Also, the kbuild maintainer recommended us to use select only for options that doesn't have any dependency. This is reflected on Kbuild docs: Note: select should be used with care. select will force a symbol to a value without visiting the dependencies. By abusing select you are able to select a symbol FOO even if FOO depends on BAR that is not set. In general use select I'm aware of the problem with select. One way to deal with is to have whatever is selecting FOO also depend on BAR. In this case CX88_MPEG only has one dependency, VIDEO_CX88. Both CX88_BLACKBIRD and CX88_DVB already depend on VIDEO_CX88 so there is no problem. Since CX88_MPEG is a hidden option and it's highly unlikely it will gain a dependency that CX88_BLACKBIRD and CX88_DVB do not also also have, I think this use of select is not so bad. I took some time during this kernel cycle trying to fix mpeg dependencies. One issue that I had to deal with is that some possible configurations with CX88, CX88_BLACKBIRD and CX88_DVB causes breakages (weird configs like having one as M and others as Y). Those weird configurations don't have any practical usage, but causes lot of headaches at the Kconfig logic. I suspect that the current way compiles fine in-kernel, but maybe the out-of-tree logic may have some troubles on handling it (at least, Ingo and others that likes to play with all possible configurations didn't complain yet with the current cx88 Kbuild). So, IMO, we should solve the issue using a different logic that relies only on depends on. This can be done, it was the other option I mentioned. MPEG needs to be made a visible option, which BLACKBIRD and DVB will then depend on instead of selecting. I didn't do this because of complaints about kernel configuration getting too complex with too many options. So I didn't think a patch exposing another option would be better. It would look something like this: config VIDEO_CX88_MPEG tristate MPEG and digital TV interface depends on VIDEO_CX88 help Support for the MPEG port on Conexant CX2388x chips, which shows up as function 2 of the PCI device. This is used by DVB/ATSC digital TV cards as well as cards with the Conexant Blackbird CX23416 MPEG encoder. config VIDEO_CX88_MPEG I'm assuming CX88_DVB here. tristate DVB/ATSC Support for cx2388x based TV cards depends on VIDEO_CX88_MPEG ... config VIDEO_CX88_BLACKBIRD tristate Blackbird MPEG encoder support (cx2388x + cx23416) depends on VIDEO_CX88_MPEG ... Would that be better? No select, but there is another option for the user to pick when configuring their cards. I almost used a similar logic like above (just replacing tristate by boolean on CX88_DVB and CX88_BLACKBIRD to avoid the weird and unpractical kconfigs). I agree that adding more questions to the poor end-user is evil. So, I decided for a hidden var for CX88_MPEG, than adding a meaningless option (on the user POV), on the latest CX88 Kbuild changeset. Cheers, Mauro -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PULL] http://linuxtv.org/hg/~tap/v4l-dvb
On Tue, 10 Feb 2009, Mauro Carvalho Chehab wrote: On Mon, 9 Feb 2009, Mauro Carvalho Chehab wrote: On Thu, 5 Feb 2009 10:57:21 -0800 (PST) Trent Piepho xy...@speakeasy.org wrote: 01/03: cx88: Fix MPEG kconfig dependencies http://linuxtv.org/hg/~tap/v4l-dvb?cmd=changeset;node=2e05fd69e6b6 Hmm... This patch converts a depends on into a select. This is not nice, since select is known to not be properly handled by kernel build system. Also, the kbuild maintainer recommended us to use select only for options that doesn't have any dependency. This is reflected on Kbuild docs: Note: select should be used with care. select will force a symbol to a value without visiting the dependencies. By abusing select you are able to select a symbol FOO even if FOO depends on BAR that is not set. In general use select I'm aware of the problem with select. One way to deal with is to have whatever is selecting FOO also depend on BAR. In this case CX88_MPEG only has one dependency, VIDEO_CX88. Both CX88_BLACKBIRD and CX88_DVB already depend on VIDEO_CX88 so there is no problem. Since CX88_MPEG is a hidden option and it's highly unlikely it will gain a dependency that CX88_BLACKBIRD and CX88_DVB do not also also have, I think this use of select is not so bad. I took some time during this kernel cycle trying to fix mpeg dependencies. One issue that I had to deal with is that some possible configurations with CX88, CX88_BLACKBIRD and CX88_DVB causes breakages (weird configs like having one as M and others as Y). Those weird configurations don't have any practical usage, but causes lot of headaches at the Kconfig logic. Both of the methods I've show will work 100% correctly with any combination of Y, M, and so on. I suspect that the current way compiles fine in-kernel, but maybe the I don't think it will. It's going to set MPEG to 'y' when CX88_VIDEO is 'm' because the default line will override the dependency. Trying to avoid select by using default X if ... lines has the same problem as select: dependencies are ignored. It's worse than select since it's more complex and fails on various odd cases like one 'y' and one 'm'. It would look something like this: config VIDEO_CX88_MPEG tristate MPEG and digital TV interface depends on VIDEO_CX88 help Support for the MPEG port on Conexant CX2388x chips, which shows up as function 2 of the PCI device. This is used by DVB/ATSC digital TV cards as well as cards with the Conexant Blackbird CX23416 MPEG encoder. config VIDEO_CX88_MPEG I'm assuming CX88_DVB here. tristate DVB/ATSC Support for cx2388x based TV cards depends on VIDEO_CX88_MPEG ... config VIDEO_CX88_BLACKBIRD tristate Blackbird MPEG encoder support (cx2388x + cx23416) depends on VIDEO_CX88_MPEG ... Would that be better? No select, but there is another option for the user to pick when configuring their cards. I almost used a similar logic like above (just replacing tristate by boolean on CX88_DVB and CX88_BLACKBIRD to avoid the weird and unpractical kconfigs). I agree that adding more questions to the poor end-user is evil. So, I decided for a hidden var for CX88_MPEG, than adding a meaningless option (on the user POV), on the latest CX88 Kbuild changeset. Is the latest CX88 kbuild changeset different than what's in Hg now? I think my patch with select is the best way. There is no reason that MPEG would depend on something that DVB and BLACKBIRD don't, so there is no problem with select. There is also no problem with MPEG getting selected from many places throughout the kernel which might not have the right dependencies, since nothing besides DVB and BLACKBIRD are going to select it. -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PULL] http://linuxtv.org/hg/~tap/v4l-dvb
On Mon, 9 Feb 2009 20:01:05 -0800 (PST) Trent Piepho xy...@speakeasy.org wrote: ... I took some time during this kernel cycle trying to fix mpeg dependencies. One issue that I had to deal with is that some possible configurations with CX88, CX88_BLACKBIRD and CX88_DVB causes breakages (weird configs like having one as M and others as Y). Those weird configurations don't have any practical usage, but causes lot of headaches at the Kconfig logic. Both of the methods I've show will work 100% correctly with any combination of Y, M, and so on. I suspect that the current way compiles fine in-kernel, but maybe the I don't think it will. It's going to set MPEG to 'y' when CX88_VIDEO is 'm' because the default line will override the dependency. Trying to avoid select by using default X if ... lines has the same problem as select: dependencies are ignored. It's worse than select since it's more complex and fails on various odd cases like one 'y' and one 'm'. ... I almost used a similar logic like above (just replacing tristate by boolean on CX88_DVB and CX88_BLACKBIRD to avoid the weird and unpractical kconfigs). I agree that adding more questions to the poor end-user is evil. So, I decided for a hidden var for CX88_MPEG, than adding a meaningless option (on the user POV), on the latest CX88 Kbuild changeset. Is the latest CX88 kbuild changeset different than what's in Hg now? I think my patch with select is the best way. There is no reason that MPEG would depend on something that DVB and BLACKBIRD don't, so there is no problem with select. There is also no problem with MPEG getting selected from many places throughout the kernel which might not have the right dependencies, since nothing besides DVB and BLACKBIRD are going to select it. Take a look on the changesets I'm referring: changeset: 10190:eb1848dcb073 user:Mauro Carvalho Chehab mche...@redhat.com date:Tue Jan 06 17:06:07 2009 -0200 summary: cx88: Fix some Kbuild troubles changeset: 10191:7c430437c967 user:Mauro Carvalho Chehab mche...@redhat.com date:Wed Jan 07 00:53:22 2009 -0200 summary: cx88: Fix Kconfig MPEG dependencies There is a large descriptionon 10190 explaining what were the issues. I had even to touch inside the code to avoid breakages, since, otherwise, things like this would break compilation: CONFIG_VIDEO_CX88=y CONFIG_VIDEO_CX88_BLACKBIRD=y CONFIG_VIDEO_CX88_DVB=m CONFIG_DVB_CORE=m Btw, at the in-kernel compilation, if you do something like: config FOO depends on BAR default y if BAR is m, then FOO is also m; if BAR is n then FOO is n. I've checked it, and also, this is what Roman (Kbuild maintainer) recommended us as a way to avoid the usage of select: http://lkml.org/lkml/2008/5/25/220 I suspect that the issue is on the parsing rules on our build system. I would like to have the offended .config, in order to test against 2.6.29-rc and see if we have such troubles upstream. Cheers, Mauro -- To unsubscribe from this list: send the line unsubscribe linux-media in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html