Re: [Dri-devel] Re: [Dri-patches] CVS Update: xc (branch: bsd-2-0-0-branch)
It helps with some hangs related to unbinding of AGP textures. It would happen the second time you started X, if I recall correctly. We were letting areas get unbound twice. What problems are people having on BSD DRI at this point? P.S. I can't seem to send mail to you ( [EMAIL PROTECTED] ), unrouteable mail domain apply Eric's fixes for AGP support I went looking through dri-devel a bit to see if I could find more information, not much luck. What was fixed? (here's hoping for a very good answer.. hehe) -- Eric Anholt [EMAIL PROTECTED] ___ Dri-devel mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [Dri-devel] Re: [Dri-patches] CVS Update: xc (branch: bsd-2-0-0-branch)
On Fri, Apr 27, 2001 at 11:43:45AM +0100, Doug Rabson wrote: This is great Alan. When this gets closer to working, I would like to try it out. I have a few ideas on how to simplify the authentication code in the drivers. The linux driver relies on being able to track each individual open call to the driver which isn't the way BSD drivers work so I try to look up the auth information using the current PID. I recently realised an alternative way of writing drivers in BSD which would make it possible to track each open call which would make life easier for DRM. Definately could use your help when the compilation issues are out of the way. One thing though. We use the copy_from_user and copy_to_user in linux, and I've noticed you didn't do that in the original BSD drivers. You just passed a pointer. Was there a reason for this ? From what I can see - we should be using copyin/copyout ? Is this right (BSD kernel knowledge is very new!). I guess we are talking about ioctls. In BSD (and I think in most ATT derived unix kernels), the kernel handles the copyin/copyout of ioctl arguments. It requires that the data direction and size fields of the ioctl number are correct and it uses that information to copy to/from a kernel buffer. The driver's ioctl routing is actually passed a pointer to a kernel buffer, not the user's pointer. Of course, if the structure passed to ioctl contains pointers to other user data structures, the driver must perform its own copyin. It would be pretty easy to do the copy_from_user/copy_to_user for the linux drivers in common code and I think it would simplify the drivers. ___ Dri-devel mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/dri-devel
Re: [Dri-devel] Re: [Dri-patches] CVS Update: xc (branch: bsd-2-0-0-branch)
On Fri, Apr 27, 2001 at 12:06:51PM +0100, Doug Rabson wrote: On Fri, Apr 27, 2001 at 11:43:45AM +0100, Doug Rabson wrote: This is great Alan. When this gets closer to working, I would like to try it out. I have a few ideas on how to simplify the authentication code in the drivers. The linux driver relies on being able to track each individual open call to the driver which isn't the way BSD drivers work so I try to look up the auth information using the current PID. I recently realised an alternative way of writing drivers in BSD which would make it possible to track each open call which would make life easier for DRM. Definately could use your help when the compilation issues are out of the way. One thing though. We use the copy_from_user and copy_to_user in linux, and I've noticed you didn't do that in the original BSD drivers. You just passed a pointer. Was there a reason for this ? From what I can see - we should be using copyin/copyout ? Is this right (BSD kernel knowledge is very new!). I guess we are talking about ioctls. In BSD (and I think in most ATT derived unix kernels), the kernel handles the copyin/copyout of ioctl arguments. It requires that the data direction and size fields of the ioctl number are correct and it uses that information to copy to/from a kernel buffer. The driver's ioctl routing is actually passed a pointer to a kernel buffer, not the user's pointer. Of course, if the structure passed to ioctl contains pointers to other user data structures, the driver must perform its own copyin. It would be pretty easy to do the copy_from_user/copy_to_user for the linux drivers in common code and I think it would simplify the drivers. Sure. Thanks for that. Alan. ___ Dri-devel mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/dri-devel
