linux-kernel - Re: [PATCH v3] tracing: Function stack size and its name mismatch in arm64

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 linux-cve-announce 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]

Message-ID: <CAEXW_YSUgqt2RykAXH+1hbVDka4CipNatK-ktZ5+W1e5nWMQvA@mail.gmail.com>
Date:   Sat, 3 Aug 2019 04:32:19 -0400
From:   Joel Fernandes <joel@...lfernandes.org>
To:     Steven Rostedt <rostedt@...dmis.org>
Cc:     Jiping Ma <jiping.ma2@...driver.com>,
        Ingo Molnar <mingo@...hat.com>,
        Catalin Marinas <catalin.marinas@....com>,
        Will Deacon <will.deacon@....com>,
        LKML <linux-kernel@...r.kernel.org>,
        "moderated list:ARM/STM32 ARCHITECTURE" 
        <linux-arm-kernel@...ts.infradead.org>,
        kernel-team <kernel-team@...roid.com>
Subject: Re: [PATCH v3] tracing: Function stack size and its name mismatch in arm64

On Sat, Aug 3, 2019 at 4:26 AM Joel Fernandes <joel@...lfernandes.org> wrote:
>
> On Fri, Aug 02, 2019 at 11:22:59AM -0400, Steven Rostedt wrote:
> [snip]
> > > There is not PC in ARM64 stack, LR is used to for walk_stackframe in
> > > ARM64. Tere is no the issue in ARM32 because there is PC in ARM32 stack.
> > > PC is used to calculate the stack size in trace_stack.c, so the
> > > function name and its stack size appear to be off-by-one.
> > > ARM64 stack layout:
> > >     LR
> > >         FP
> > >         ......
> > >         LR
> > >         FP
> > >         ......
> >
> > I think you are not explaining the issue correctly. From looking at the
> > document, I think what you want to say is that the LR is saved *after*
> > the data for the function. Is that correct? If so, then yes, it would
> > cause the stack tracing algorithm to be incorrect.
> >
> > Most archs do this:
> >
> > On entry to a function:
> >
> >       save return address
> >       reserve local variables and such for current function
> >
> > I think you are saying that arm64 does this backwards.
> >
> >       reserve local variables and such for current function
> >       save return address (LR)
>
> Actually for arm64 it is like what you said about 'Most archs'. It saves FP
> and LR first onto the current stack frame, then assigns the top of the stack
> to FP (marking the new frame). Then executes branch-link, and then allocates
> space to variables on stack in the callee.

Just to add to that, when the branch-link (BL) instruction is
executed, the Link Register (LR) will contain the return address. This
why the existing LR needs to be saved first before the call. The
existing LR will contain the return address of the caller's original
caller.