Please do not reply to this email: if you want to comment on the bug, go to the URL shown below and enter yourcomments there. https://bugs.freedesktop.org/show_bug.cgi?id=6242
------- Additional Comments From [EMAIL PROTECTED] 2006-04-01 06:39 ------- (In reply to comment #6) > (In reply to comment #5) > > Created an attachment (id=5123) --> (https://bugs.freedesktop.org/attachment.cgi?id=5123&action=view) [edit] [edit] > > mach64-drm: DMA buffers for vertices - try 2 > > > > enable test of register values in the vertex buffer. > > It's good to see this driver getting some attention again, I'm sorry I don't > have time to help out with the code right now. > > I wanted to point out that if you go this route (using client DMA buffers for > vertices), you'll need to unmap the buffers from userspace before verifying > and > submitting them to make this secure. If you look at the CVS history, you'll > see > that we actually used to use DMA buffers, but changed to the current method > when > beginning the work to make the driver secure. We decided that copying the > data > from client memory into a pool of unmapped DMA buffers in the kernel would be > faster than unmapping and re-mapping the buffers for each submit. Copying the > userspace data into DMA buffers allocated from the mapped pool was a stopgap > until we had a way to allocate unmapped buffers in the kernel for this > purpose. > Grrh, I knew I was missing something basic when I tried to touch this stuff ... also a bit stupid of me to lurk and search mesa-dev for mach64 DMA instead of dri-devel. I'll probably do my "homework" over the weekend. Just a question to make sure I am not completely mixed up: As it stands now, is drm_blit secure ? I mean a malicious client can still change the contents of the DMA buffer after submission to the drm in the case of a blit also, esp. the first MACH64_HOSTDATA_BLIT_OFFSET bytes. > It's been a while since I looked at the code in detail, but the problem at the > time was a limitation of the DRM buffer management infrastructure. As I > recall, > there wasn't an easy way to create two sets of DMA buffers, one mapped to > userspace, and one which wasn't mapped. You may want to look into the mailing > list archives for more discussion on this. I think that this was the last > remaining issue in securing the driver. > > -Leif Thanks for your comments and your work on open source drivers, george. -- Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. ------------------------------------------------------- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642 -- _______________________________________________ Dri-devel mailing list Dri-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dri-devel