Messages in this thread | | | Date | Tue, 23 Aug 2011 22:17:44 +0100 | From | Al Viro <> | Subject | Re: [uml-devel] SYSCALL, ptrace and syscall restart breakages (Re: [RFC] weird crap with vdso on uml/i386) |
| |
On Tue, Aug 23, 2011 at 12:43:28PM -0700, Linus Torvalds wrote: > On Tue, Aug 23, 2011 at 12:41 PM, Al Viro <viro@zeniv.linux.org.uk> wrote: > > > > And it's not cheap - doing that on each syscall will be unpleasant... > > Frankly, I'd rather stopped telling the uml userland about vdso in such > > setups. ?And anything that plays with SYSCALL outside of vdso... <shrug> > > we already have a "don't run it native on 32bit", adding "don't run > > it on 32bit uml on amd64 host" is not too serious. ?At least for now... > > I do agree that the solution might well be to just stop using the > non-int80 vdsos for UML. That may just solve everything in practice.
SYSENTER works fine, actually... And we can easily check if we have an affected SYSCALL, simply by forking a child, tracing it into a syscall and doing POKEUSER to ebp on the second stop (i.e. on the way out). If the value ends up in ecx after __kernel_vsyscall(), we have SYSCALL-based variant on amd64 host, if it's lost completely - it's SYSENTER, if it shows up in ebp - int 0x80.
| |