On Mon, Jul 6, 2020 at 9:21 AM Kees Cook wrote:
>
> On Mon, Jul 06, 2020 at 09:01:51AM +0800, Guo Ren wrote:
> > Yeah! :) I just want to show you, how about the format: use tp in gpr
> > to do that. The format is similar to arm64.
> >
> > tp is the task_struct point in riscv.
>
> Sounds good to me
On Mon, Jul 06, 2020 at 09:01:51AM +0800, Guo Ren wrote:
> Yeah! :) I just want to show you, how about the format: use tp in gpr
> to do that. The format is similar to arm64.
>
> tp is the task_struct point in riscv.
Sounds good to me, yes. Thanks! Is there anyone looking at the GCC and
Clang sid
On Mon, Jul 6, 2020 at 4:40 AM Kees Cook wrote:
>
> On Sun, Jul 05, 2020 at 02:13:17PM +, guo...@kernel.org wrote:
> > From: Guo Ren
> >
> > After compare arm64 and x86 implementations, seems arm64's is more
> > flexible and readable. The key point is how gcc get the offset of
> > stack_canar
On Sun, Jul 05, 2020 at 02:13:17PM +, guo...@kernel.org wrote:
> From: Guo Ren
>
> After compare arm64 and x86 implementations, seems arm64's is more
> flexible and readable. The key point is how gcc get the offset of
> stack_canary from gs/el0_sp.
>
> x86: Use a fix offset from gs, not flex
From: Guo Ren
After compare arm64 and x86 implementations, seems arm64's is more
flexible and readable. The key point is how gcc get the offset of
stack_canary from gs/el0_sp.
x86: Use a fix offset from gs, not flexible.
struct fixed_percpu_data {
/*
* GCC hardcodes the stack c
5 matches
Mail list logo
