linux-kernel - Re: [patch part-II V2 10/13] x86/entry/common: Split prepare_exit_to_usermode() and syscall_return

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: <66bc932f-3ef7-1bcd-c5bc-c546036e2f3f@oracle.com>
Date:   Tue, 10 Mar 2020 14:37:30 +0100
From:   Alexandre Chartre <alexandre.chartre@...cle.com>
To:     Thomas Gleixner <tglx@...utronix.de>,
        LKML <linux-kernel@...r.kernel.org>
Cc:     x86@...nel.org, Steven Rostedt <rostedt@...dmis.org>,
        Brian Gerst <brgerst@...il.com>,
        Juergen Gross <jgross@...e.com>,
        Frederic Weisbecker <frederic@...nel.org>
Subject: Re: [patch part-II V2 10/13] x86/entry/common: Split
 prepare_exit_to_usermode() and syscall_return_slowpath()


On 3/8/20 11:24 PM, Thomas Gleixner wrote:
> Split the functions into traceable and probale parts and a part protected

typo: probale -> probeable ?

> from instrumentation.
> 
> This is required because after calling user_exit_irqsoff() kprobes and
> tracepoints/function entry/exit which can be utilized by e.g. BPF are not
> longer safe vs. RCU.
> 
> Preparatory step to move irq flags tracing to C code.
> 
> Signed-off-by: Thomas Gleixner <tglx@...utronix.de>
> ---
> V2: New patch to reduce the churn
> ---
>   arch/x86/entry/common.c |   25 ++++++++++++++++++-------
>   1 file changed, 18 insertions(+), 7 deletions(-)
> 
> --- a/arch/x86/entry/common.c
> +++ b/arch/x86/entry/common.c
> @@ -206,7 +206,7 @@ static void exit_to_usermode_loop(struct
>   }
>   
>   /* Called with IRQs disabled. */
> -__visible inline void prepare_exit_to_usermode(struct pt_regs *regs)
> +static noinline void __prepare_exit_to_usermode(struct pt_regs *regs)

Why noinline?

Reviewed-by: Alexandre Chartre <alexandre.chartre@...cle.com>

alex.


>   {
>   	struct thread_info *ti = current_thread_info();
>   	u32 cached_flags;
> @@ -245,11 +245,16 @@ static void exit_to_usermode_loop(struct
>   	 */
>   	ti->status &= ~(TS_COMPAT|TS_I386_REGS_POKED);
>   #endif
> +}
>   
> -	user_enter_irqoff();
> +__visible inline notrace void prepare_exit_to_usermode(struct pt_regs *regs)
> +{
> +	__prepare_exit_to_usermode(regs);
>   
> +	user_enter_irqoff();
>   	mds_user_clear_cpu_buffers();
>   }
> +NOKPROBE_SYMBOL(prepare_exit_to_usermode);
>   
>   #define SYSCALL_EXIT_WORK_FLAGS				\
>   	(_TIF_SYSCALL_TRACE | _TIF_SYSCALL_AUDIT |	\
> @@ -277,11 +282,7 @@ static void syscall_slow_exit_work(struc
>   		tracehook_report_syscall_exit(regs, step);
>   }
>   
> -/*
> - * Called with IRQs on and fully valid regs.  Returns with IRQs off in a
> - * state such that we can immediately switch to user mode.
> - */
> -__visible inline void syscall_return_slowpath(struct pt_regs *regs)
> +static void __syscall_return_slowpath(struct pt_regs *regs)
>   {
>   	struct thread_info *ti = current_thread_info();
>   	u32 cached_flags = READ_ONCE(ti->flags);
> @@ -302,8 +303,18 @@ static void syscall_slow_exit_work(struc
>   		syscall_slow_exit_work(regs, cached_flags);
>   
>   	local_irq_disable();
> +}
> +
> +/*
> + * Called with IRQs on and fully valid regs.  Returns with IRQs off in a
> + * state such that we can immediately switch to user mode.
> + */
> +__visible inline notrace void syscall_return_slowpath(struct pt_regs *regs)
> +{
> +	__syscall_return_slowpath(regs);
>   	prepare_exit_to_usermode(regs);
>   }
> +NOKPROBE_SYMBOL(syscall_return_slowpath);
>   
>   #ifdef CONFIG_X86_64
>   static __always_inline
>