[V4L] PATCH 2.4.0.10.6: bttv mmap rewrite

2001-02-20 Thread Jeff Garzik
update (new member: mmap_vma). * KNOWN BUG: If your gbufsize causes a frame to cross a page boundary, you lose. The code doesn't support this, and doesn't check for it either. Boom. (this is fixable though) * Totally untested. I don't own any bttv hardware. -- Jeff Garzik | "

Re: [V4L] races in v4l read + write in 2.4

2000-10-30 Thread Jeff Garzik
Alan Cox wrote: If this was to be done what good is videodev.c ? videodev.c just handles minor number registration. Everything else it tried to help with was an error *pounce* What do you think about the attached patch, against 2.4.0-test10-pre7? -- Jeff Garzik | "Mind

Re: [V4L] races in v4l read + write in 2.4

2000-10-30 Thread Jeff Garzik
Erm, nevermind :) totally in the wrong place.. -- Jeff Garzik | "Mind if I drive?" -Sam Building 1024 | "Not if you don't mind me clawing at the MandrakeSoft| dash and shrieking like a cheerleader." |

[V4L] PATCH: file_operations for videodev

2000-10-30 Thread Jeff Garzik
Ok, much better. devfs wants "void *ops" not a struct file_operations. Attached is a better patch, against 2.4.0-test10-pre7, which allows video devices to handle all file ops themselves, giving themselves full power to do whatever they wish. -- Jeff Garzik | "

v4l multi-channel handling?

2000-05-23 Thread Jeff Garzik
applications accomplish this by multiple open, or should the driver register multiple v4l intefaces? The spec seems to assume that people will switch between channels, not view them all at once. Thanks, Jeff -- Jeff Garzik | Liberty is always dangerous, but Building 1024