[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20201109111043.GD2594@hirez.programming.kicks-ass.net>
Date: Mon, 9 Nov 2020 12:10:43 +0100
From: Peter Zijlstra <peterz@...radead.org>
To: Steven Rostedt <rostedt@...dmis.org>
Cc: linux-kernel@...r.kernel.org, Ingo Molnar <mingo@...nel.org>,
Andrew Morton <akpm@...ux-foundation.org>,
Miroslav Benes <mbenes@...e.cz>,
Thomas Gleixner <tglx@...utronix.de>,
Masami Hiramatsu <mhiramat@...nel.org>
Subject: Re: [PATCH 2/3 v4] ftrace/x86: Allow for arguments to be passed in
to ftrace_regs by default
On Fri, Nov 06, 2020 at 04:42:36PM -0500, Steven Rostedt wrote:
> +static __always_inline struct pt_regs *
> +arch_ftrace_get_regs(struct ftrace_regs *fregs)
> +{
> + /* Only when FL_SAVE_REGS is set, cs will be non zero */
> + if (!fregs->regs.cs)
> + return NULL;
> + return &fregs->regs;
> +}
> +#endif
> +
> #endif /* CONFIG_DYNAMIC_FTRACE */
> #endif /* __ASSEMBLY__ */
> #endif /* CONFIG_FUNCTION_TRACER */
> diff --git a/arch/x86/kernel/ftrace_64.S b/arch/x86/kernel/ftrace_64.S
> index ac3d5f22fe64..60e3b64f5ea6 100644
> --- a/arch/x86/kernel/ftrace_64.S
> +++ b/arch/x86/kernel/ftrace_64.S
> @@ -140,12 +140,19 @@ SYM_FUNC_START(ftrace_caller)
> /* save_mcount_regs fills in first two parameters */
> save_mcount_regs
>
> + /* Stack - skipping return address of ftrace_caller */
> + leaq MCOUNT_REG_SIZE+8(%rsp), %rcx
> + movq %rcx, RSP(%rsp)
> +
> SYM_INNER_LABEL(ftrace_caller_op_ptr, SYM_L_GLOBAL)
> /* Load the ftrace_ops into the 3rd parameter */
> movq function_trace_op(%rip), %rdx
>
> - /* regs go into 4th parameter (but make it NULL) */
> - movq $0, %rcx
> + /* regs go into 4th parameter */
> + leaq (%rsp), %rcx
> +
> + /* Only ops with REGS flag set should have CS register set */
> + movq $0, CS(%rsp)
>
> SYM_INNER_LABEL(ftrace_call, SYM_L_GLOBAL)
> call ftrace_stub
You now seem to be relying on save_mcount_regs() resulting in a cleared
CS, however, AFAICT CS is uninitialized stack garbage.
Powered by blists - more mailing lists