On Thu, Jul 13, 2017 at 10:11 PM, Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx> wrote: > This one actually seems to imply that vmw_cmd_invalid() is broken: > > drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c:522:34: warning: the omitted > middle operand in ?: will always be ‘true’, suggest explicit middle > operand [-Wparentheses] > return capable(CAP_SYS_ADMIN) ? : -EINVAL; > ^ > > gcc is very right. I think the *intent* of that function is to return > 0 if CAP_SYS_ADMIN, but it always returns nonzero. > > If CAP_SYS_ADMIN is true, it returns 1, and if it's false it returns -EINVAL. > > Since nobody has noticed in many many years, maybe it should just > always return -EINVAL unconditionally? Fun one. It's the security parser afaict, and it looks like no one wrote a testcase to make sure it actually catches the bad guys. Or it's not needed, since there's 3 more security checks above it in vmw_cmd_check. Worst case it's become uapi and we can't fix it :-) Anyway, not my area, I'll leave this to Sinclair to figure out. -Daniel -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch _______________________________________________ dri-devel mailing list dri-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/dri-devel
