Messages in this thread | | | Subject | Re: [PATCH v4] x86, selftests: Add sigreturn selftest | From | Michael Ellerman <> | Date | Wed, 08 Apr 2015 13:42:10 +1000 |
| |
On Mon, 2015-04-06 at 23:11 -0700, Andy Lutomirski wrote: > On Mon, Apr 6, 2015 at 8:39 PM, Michael Ellerman <mpe@ellerman.id.au> wrote: > > On Mon, 2015-04-06 at 19:01 -0700, Andy Lutomirski wrote: > >> This is my sigreturn test, added mostly unchanged from its old home. > >> It exercises the sigreturn(2) syscall, specifically focusing on its > >> interactions with various IRET corner cases. It tests for correct > >> behavior in several areas that were historically dangerously buggy. > >> For example, it exercises espfix on kernels of both bitnesses under > >> various conditions, and it contains exploits for several now-fixed > >> bugs in IRET error handling. > >> > >> If you run it on older kernels, your system will crash. It probably > >> won't eat your data in the process. > >> > >> There is no released kernel on which the sigreturn_64 test will > >> pass, but it passes on tip:x86/asm. > >> > >> IMO it's unfortunate that I need to provide a special script to run > >> tests. I'd rather just list my targets. > > > > If you use lib.mk you can. > > > > https://git.kernel.org/cgit/linux/kernel/git/shuah/linux-kselftest.git/log/?h=next > > > > See for example: > > > > https://git.kernel.org/cgit/linux/kernel/git/shuah/linux-kselftest.git/commit/?h=next&id=5744de542dd4b963c2975e6f70844ce2899864e4 > > Will do for 4.2. In the mean time, there's no base on which lib.mk > exists and the test works.
OK.
Shua seems to have already sent a pull request for 4.1, which is rather early to say the least, but suggests you've missed 4.1 anyway - unless you want it to go via some other tree.
cheers
| |