lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Mon, 17 Aug 2020 20:47:21 +0800
From:   Guo Ren <guoren@...nel.org>
To:     Palmer Dabbelt <palmerdabbelt@...gle.com>
Cc:     Paul Walmsley <paul.walmsley@...ive.com>,
        Masami Hiramatsu <mhiramat@...nel.org>,
        Oleg Nesterov <oleg@...hat.com>,
        linux-riscv <linux-riscv@...ts.infradead.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Anup Patel <anup@...infault.org>, linux-csky@...r.kernel.org,
        Greentime Hu <greentime.hu@...ive.com>,
        Zong Li <zong.li@...ive.com>,
        Patrick Stählin <me@...ki.ch>,
        Bjorn Topel <bjorn.topel@...il.com>,
        Guo Ren <guoren@...ux.alibaba.com>
Subject: Re: [PATCH v3 3/7] riscv: Fixup kprobes handler couldn't change pc

On Sat, Aug 15, 2020 at 6:36 AM Palmer Dabbelt <palmerdabbelt@...gle.com> wrote:
>
> On Mon, 13 Jul 2020 16:39:18 PDT (-0700), guoren@...nel.org wrote:
> > From: Guo Ren <guoren@...ux.alibaba.com>
> >
> > The "Changing Execution Path" section in the Documentation/kprobes.txt
> > said:
> >
> > Since kprobes can probe into a running kernel code, it can change the
> > register set, including instruction pointer.
> >
> > Signed-off-by: Guo Ren <guoren@...ux.alibaba.com>
> > Cc: Masami Hiramatsu <mhiramat@...nel.org>
> > Cc: Palmer Dabbelt <palmerdabbelt@...gle.com>
> > ---
> >  arch/riscv/kernel/mcount-dyn.S | 3 ++-
> >  1 file changed, 2 insertions(+), 1 deletion(-)
> >
> > diff --git a/arch/riscv/kernel/mcount-dyn.S b/arch/riscv/kernel/mcount-dyn.S
> > index 35a6ed7..4b58b54 100644
> > --- a/arch/riscv/kernel/mcount-dyn.S
> > +++ b/arch/riscv/kernel/mcount-dyn.S
> > @@ -123,6 +123,7 @@ ENDPROC(ftrace_caller)
> >       sd      ra, (PT_SIZE_ON_STACK+8)(sp)
> >       addi    s0, sp, (PT_SIZE_ON_STACK+16)
> >
> > +     sd ra,  PT_EPC(sp)
> >       sd x1,  PT_RA(sp)
> >       sd x2,  PT_SP(sp)
> >       sd x3,  PT_GP(sp)
>
> So that's definately not going to be EPC any more.  I'm not sure that field is
> sanely named, though, as it's really just the PC when it comes to other ptrace
> stuff.
>
> > @@ -157,6 +158,7 @@ ENDPROC(ftrace_caller)
> >       .endm
> >
> >       .macro RESTORE_ALL
> > +     ld ra,  PT_EPC(sp)
> >       ld x1,  PT_RA(sp)
>
> x1 is ra, so loading it twice doesn't seem reasonable.
>
> >       ld x2,  PT_SP(sp)
> >       ld x3,  PT_GP(sp)
> > @@ -190,7 +192,6 @@ ENDPROC(ftrace_caller)
> >       ld x31, PT_T6(sp)
> >
> >       ld      s0, (PT_SIZE_ON_STACK)(sp)
> > -     ld      ra, (PT_SIZE_ON_STACK+8)(sp)
> >       addi    sp, sp, (PT_SIZE_ON_STACK+16)
> >       .endm
>
> If you're dropping the load you should drop the store above as well.  In
> general this seems kind of mixed up, both before and after this patch.

This is a wrong patch, it should be:

diff --git a/arch/riscv/kernel/mcount-dyn.S b/arch/riscv/kernel/mcount-dyn.S
index 35a6ed7..d82b8f0 100644
--- a/arch/riscv/kernel/mcount-dyn.S
+++ b/arch/riscv/kernel/mcount-dyn.S
@@ -120,10 +120,10 @@ ENDPROC(ftrace_caller)
        .macro SAVE_ALL
        addi    sp, sp, -(PT_SIZE_ON_STACK+16)
        sd      s0, (PT_SIZE_ON_STACK)(sp)
-       sd      ra, (PT_SIZE_ON_STACK+8)(sp)
        addi    s0, sp, (PT_SIZE_ON_STACK+16)

-       sd x1,  PT_RA(sp)
+       sd ra,  PT_EPC(sp)
+       sd ra,  PT_RA(sp)
        sd x2,  PT_SP(sp)
        sd x3,  PT_GP(sp)
        sd x4,  PT_TP(sp)
@@ -157,7 +157,7 @@ ENDPROC(ftrace_caller)
        .endm

        .macro RESTORE_ALL
-       ld x1,  PT_RA(sp)
+       ld ra,  PT_EPC(sp)
        ld x2,  PT_SP(sp)
        ld x3,  PT_GP(sp)
        ld x4,  PT_TP(sp)
@@ -190,7 +190,6 @@ ENDPROC(ftrace_caller)
        ld x31, PT_T6(sp)

        ld      s0, (PT_SIZE_ON_STACK)(sp)
-       ld      ra, (PT_SIZE_ON_STACK+8)(sp)
        addi    sp, sp, (PT_SIZE_ON_STACK+16)
        .endm

Now, I'm developing livepatch and they are so mixed features (kprobe,
livepatch, ftrace, optprobes, STACK_WALK, -fpatchable-function-entry
'no -pg'). I'll test this patch in the next version of the patchset.

Thx for the review.

-- 
Best Regards
 Guo Ren

ML: https://lore.kernel.org/linux-csky/

Powered by blists - more mailing lists