x86/signal: Detect and prevent an alternate signal stack overflow
authorChang S. Bae <chang.seok.bae@intel.com>
Tue, 18 May 2021 20:03:19 +0000 (13:03 -0700)
committerBorislav Petkov <bp@suse.de>
Wed, 19 May 2021 10:40:30 +0000 (12:40 +0200)
commit2beb4a53fc3f1081cedc1c1a198c7f56cc4fc60c
tree06304bf99dde06fa21c6631e428048a93783ef8d
parentbdf6c8b84a4fa726c382ef6d3518f3ae123a7ebd
x86/signal: Detect and prevent an alternate signal stack overflow

The kernel pushes context on to the userspace stack to prepare for the
user's signal handler. When the user has supplied an alternate signal
stack, via sigaltstack(2), it is easy for the kernel to verify that the
stack size is sufficient for the current hardware context.

Check if writing the hardware context to the alternate stack will exceed
it's size. If yes, then instead of corrupting user-data and proceeding with
the original signal handler, an immediate SIGSEGV signal is delivered.

Refactor the stack pointer check code from on_sig_stack() and use the new
helper.

While the kernel allows new source code to discover and use a sufficient
alternate signal stack size, this check is still necessary to protect
binaries with insufficient alternate signal stack size from data
corruption.

Fixes: c2bc11f10a39 ("x86, AVX-512: Enable AVX-512 States Context Switch")
Reported-by: Florian Weimer <fweimer@redhat.com>
Suggested-by: Jann Horn <jannh@google.com>
Suggested-by: Andy Lutomirski <luto@kernel.org>
Signed-off-by: Chang S. Bae <chang.seok.bae@intel.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Len Brown <len.brown@intel.com>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lkml.kernel.org/r/20210518200320.17239-6-chang.seok.bae@intel.com
Link: https://bugzilla.kernel.org/show_bug.cgi?id=153531
arch/x86/kernel/signal.c
include/linux/sched/signal.h