All of lore.kernel.org
 help / color / mirror / Atom feed
From: Mark Rutland <mark.rutland@arm.com>
To: Xu Kuohai <xukuohai@huawei.com>
Cc: bpf@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org, netdev@vger.kernel.org,
	linux-kselftest@vger.kernel.org,
	Catalin Marinas <catalin.marinas@arm.com>,
	Will Deacon <will@kernel.org>,
	Steven Rostedt <rostedt@goodmis.org>,
	Ingo Molnar <mingo@redhat.com>,
	Daniel Borkmann <daniel@iogearbox.net>,
	Alexei Starovoitov <ast@kernel.org>,
	Zi Shen Lim <zlim.lnx@gmail.com>,
	Andrii Nakryiko <andrii@kernel.org>,
	Martin KaFai Lau <kafai@fb.com>, Song Liu <songliubraving@fb.com>,
	Yonghong Song <yhs@fb.com>,
	John Fastabend <john.fastabend@gmail.com>,
	KP Singh <kpsingh@kernel.org>,
	"David S . Miller" <davem@davemloft.net>,
	Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>,
	David Ahern <dsahern@kernel.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	Borislav Petkov <bp@alien8.de>,
	Dave Hansen <dave.hansen@linux.intel.com>,
	x86@kernel.org, hpa@zytor.com, Shuah Khan <shuah@kernel.org>,
	Jakub Kicinski <kuba@kernel.org>,
	Jesper Dangaard Brouer <hawk@kernel.org>,
	Pasha Tatashin <pasha.tatashin@soleen.com>,
	Ard Biesheuvel <ardb@kernel.org>,
	Daniel Kiss <daniel.kiss@arm.com>,
	Steven Price <steven.price@arm.com>,
	Sudeep Holla <sudeep.holla@arm.com>,
	Marc Zyngier <maz@kernel.org>,
	Peter Collingbourne <pcc@google.com>,
	Mark Brown <broonie@kernel.org>, Delyan Kratunov <delyank@fb.com>,
	Kumar Kartikeya Dwivedi <memxor@gmail.com>
Subject: Re: [PATCH bpf-next v5 2/6] ftrace: Fix deadloop caused by direct call in ftrace selftest
Date: Wed, 25 May 2022 14:43:55 +0100	[thread overview]
Message-ID: <Yo4ymwu92gM75/Z5@FVFF77S0Q05N> (raw)
In-Reply-To: <20220518131638.3401509-3-xukuohai@huawei.com>

On Wed, May 18, 2022 at 09:16:34AM -0400, Xu Kuohai wrote:
> After direct call is enabled for arm64, ftrace selftest enters a
> dead loop:

IIUC this means that patch 1 alone is broken, and presumably this patch should
have been part of it?

> <trace_selftest_dynamic_test_func>:
> 00  bti     c
> 01  mov     x9, x30                            <trace_direct_tramp>:
> 02  bl      <trace_direct_tramp>    ---------->     ret
>                                                      |
>                                          lr/x30 is 03, return to 03
>                                                      |
> 03  mov     w0, #0x0   <-----------------------------|
>      |                                               |
>      |                   dead loop!                  |
>      |                                               |
> 04  ret   ---- lr/x30 is still 03, go back to 03 ----|
> 
> The reason is that when the direct caller trace_direct_tramp() returns
> to the patched function trace_selftest_dynamic_test_func(), lr is still
> the address after the instrumented instruction in the patched function,
> so when the patched function exits, it returns to itself!
> 
> To fix this issue, we need to restore lr before trace_direct_tramp()
> exits, so rewrite a dedicated trace_direct_tramp() for arm64.

As mentioned on patch 1 I'd prefer we solved this through indirection, which
would avoid the need for this and would make things more robust generally by
keeping the unusual calling convention private to the patch-site and regular
trampoline.

Thanks,
Mark.

> Reported-by: Li Huafei <lihuafei1@huawei.com>
> Signed-off-by: Xu Kuohai <xukuohai@huawei.com>
> ---
>  arch/arm64/include/asm/ftrace.h  | 10 ++++++++++
>  arch/arm64/kernel/entry-ftrace.S | 10 ++++++++++
>  kernel/trace/trace_selftest.c    |  2 ++
>  3 files changed, 22 insertions(+)
> 
> diff --git a/arch/arm64/include/asm/ftrace.h b/arch/arm64/include/asm/ftrace.h
> index 14a35a5df0a1..6f6b184e72fb 100644
> --- a/arch/arm64/include/asm/ftrace.h
> +++ b/arch/arm64/include/asm/ftrace.h
> @@ -126,6 +126,16 @@ static inline bool arch_syscall_match_sym_name(const char *sym,
>  	 */
>  	return !strcmp(sym + 8, name);
>  }
> +
> +#ifdef CONFIG_FTRACE_SELFTEST
> +#ifdef CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS
> +
> +#define trace_direct_tramp trace_direct_tramp
> +extern void trace_direct_tramp(void);
> +
> +#endif /* CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS */
> +#endif /* CONFIG_FTRACE_SELFTEST */
> +
>  #endif /* ifndef __ASSEMBLY__ */
>  
>  #endif /* __ASM_FTRACE_H */
> diff --git a/arch/arm64/kernel/entry-ftrace.S b/arch/arm64/kernel/entry-ftrace.S
> index dfe62c55e3a2..a47e87d4d3dd 100644
> --- a/arch/arm64/kernel/entry-ftrace.S
> +++ b/arch/arm64/kernel/entry-ftrace.S
> @@ -357,3 +357,13 @@ SYM_CODE_START(return_to_handler)
>  	ret
>  SYM_CODE_END(return_to_handler)
>  #endif /* CONFIG_FUNCTION_GRAPH_TRACER */
> +
> +#ifdef CONFIG_FTRACE_SELFTEST
> +#ifdef CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS
> +SYM_FUNC_START(trace_direct_tramp)
> +	mov x10, x30
> +	mov x30, x9
> +	ret x10
> +SYM_FUNC_END(trace_direct_tramp)
> +#endif /* CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS */
> +#endif /* CONFIG_FTRACE_SELFTEST */
> diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c
> index abcadbe933bb..e7ccd0d10c39 100644
> --- a/kernel/trace/trace_selftest.c
> +++ b/kernel/trace/trace_selftest.c
> @@ -785,8 +785,10 @@ static struct fgraph_ops fgraph_ops __initdata  = {
>  };
>  
>  #ifdef CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS
> +#ifndef trace_direct_tramp
>  noinline __noclone static void trace_direct_tramp(void) { }
>  #endif
> +#endif
>  
>  /*
>   * Pretty much the same than for the function tracer from which the selftest
> -- 
> 2.30.2
> 

WARNING: multiple messages have this Message-ID (diff)
From: Mark Rutland <mark.rutland@arm.com>
To: Xu Kuohai <xukuohai@huawei.com>
Cc: bpf@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org, netdev@vger.kernel.org,
	linux-kselftest@vger.kernel.org,
	Catalin Marinas <catalin.marinas@arm.com>,
	Will Deacon <will@kernel.org>,
	Steven Rostedt <rostedt@goodmis.org>,
	Ingo Molnar <mingo@redhat.com>,
	Daniel Borkmann <daniel@iogearbox.net>,
	Alexei Starovoitov <ast@kernel.org>,
	Zi Shen Lim <zlim.lnx@gmail.com>,
	Andrii Nakryiko <andrii@kernel.org>,
	Martin KaFai Lau <kafai@fb.com>, Song Liu <songliubraving@fb.com>,
	Yonghong Song <yhs@fb.com>,
	John Fastabend <john.fastabend@gmail.com>,
	KP Singh <kpsingh@kernel.org>,
	"David S . Miller" <davem@davemloft.net>,
	Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>,
	David Ahern <dsahern@kernel.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	Borislav Petkov <bp@alien8.de>,
	Dave Hansen <dave.hansen@linux.intel.com>,
	x86@kernel.org, hpa@zytor.com, Shuah Khan <shuah@kernel.org>,
	Jakub Kicinski <kuba@kernel.org>,
	Jesper Dangaard Brouer <hawk@kernel.org>,
	Pasha Tatashin <pasha.tatashin@soleen.com>,
	Ard Biesheuvel <ardb@kernel.org>,
	Daniel Kiss <daniel.kiss@arm.com>,
	Steven Price <steven.price@arm.com>,
	Sudeep Holla <sudeep.holla@arm.com>,
	Marc Zyngier <maz@kernel.org>,
	Peter Collingbourne <pcc@google.com>,
	Mark Brown <broonie@kernel.org>, Delyan Kratunov <delyank@fb.com>,
	Kumar Kartikeya Dwivedi <memxor@gmail.com>
Subject: Re: [PATCH bpf-next v5 2/6] ftrace: Fix deadloop caused by direct call in ftrace selftest
Date: Wed, 25 May 2022 14:43:55 +0100	[thread overview]
Message-ID: <Yo4ymwu92gM75/Z5@FVFF77S0Q05N> (raw)
In-Reply-To: <20220518131638.3401509-3-xukuohai@huawei.com>

On Wed, May 18, 2022 at 09:16:34AM -0400, Xu Kuohai wrote:
> After direct call is enabled for arm64, ftrace selftest enters a
> dead loop:

IIUC this means that patch 1 alone is broken, and presumably this patch should
have been part of it?

> <trace_selftest_dynamic_test_func>:
> 00  bti     c
> 01  mov     x9, x30                            <trace_direct_tramp>:
> 02  bl      <trace_direct_tramp>    ---------->     ret
>                                                      |
>                                          lr/x30 is 03, return to 03
>                                                      |
> 03  mov     w0, #0x0   <-----------------------------|
>      |                                               |
>      |                   dead loop!                  |
>      |                                               |
> 04  ret   ---- lr/x30 is still 03, go back to 03 ----|
> 
> The reason is that when the direct caller trace_direct_tramp() returns
> to the patched function trace_selftest_dynamic_test_func(), lr is still
> the address after the instrumented instruction in the patched function,
> so when the patched function exits, it returns to itself!
> 
> To fix this issue, we need to restore lr before trace_direct_tramp()
> exits, so rewrite a dedicated trace_direct_tramp() for arm64.

As mentioned on patch 1 I'd prefer we solved this through indirection, which
would avoid the need for this and would make things more robust generally by
keeping the unusual calling convention private to the patch-site and regular
trampoline.

Thanks,
Mark.

> Reported-by: Li Huafei <lihuafei1@huawei.com>
> Signed-off-by: Xu Kuohai <xukuohai@huawei.com>
> ---
>  arch/arm64/include/asm/ftrace.h  | 10 ++++++++++
>  arch/arm64/kernel/entry-ftrace.S | 10 ++++++++++
>  kernel/trace/trace_selftest.c    |  2 ++
>  3 files changed, 22 insertions(+)
> 
> diff --git a/arch/arm64/include/asm/ftrace.h b/arch/arm64/include/asm/ftrace.h
> index 14a35a5df0a1..6f6b184e72fb 100644
> --- a/arch/arm64/include/asm/ftrace.h
> +++ b/arch/arm64/include/asm/ftrace.h
> @@ -126,6 +126,16 @@ static inline bool arch_syscall_match_sym_name(const char *sym,
>  	 */
>  	return !strcmp(sym + 8, name);
>  }
> +
> +#ifdef CONFIG_FTRACE_SELFTEST
> +#ifdef CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS
> +
> +#define trace_direct_tramp trace_direct_tramp
> +extern void trace_direct_tramp(void);
> +
> +#endif /* CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS */
> +#endif /* CONFIG_FTRACE_SELFTEST */
> +
>  #endif /* ifndef __ASSEMBLY__ */
>  
>  #endif /* __ASM_FTRACE_H */
> diff --git a/arch/arm64/kernel/entry-ftrace.S b/arch/arm64/kernel/entry-ftrace.S
> index dfe62c55e3a2..a47e87d4d3dd 100644
> --- a/arch/arm64/kernel/entry-ftrace.S
> +++ b/arch/arm64/kernel/entry-ftrace.S
> @@ -357,3 +357,13 @@ SYM_CODE_START(return_to_handler)
>  	ret
>  SYM_CODE_END(return_to_handler)
>  #endif /* CONFIG_FUNCTION_GRAPH_TRACER */
> +
> +#ifdef CONFIG_FTRACE_SELFTEST
> +#ifdef CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS
> +SYM_FUNC_START(trace_direct_tramp)
> +	mov x10, x30
> +	mov x30, x9
> +	ret x10
> +SYM_FUNC_END(trace_direct_tramp)
> +#endif /* CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS */
> +#endif /* CONFIG_FTRACE_SELFTEST */
> diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c
> index abcadbe933bb..e7ccd0d10c39 100644
> --- a/kernel/trace/trace_selftest.c
> +++ b/kernel/trace/trace_selftest.c
> @@ -785,8 +785,10 @@ static struct fgraph_ops fgraph_ops __initdata  = {
>  };
>  
>  #ifdef CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS
> +#ifndef trace_direct_tramp
>  noinline __noclone static void trace_direct_tramp(void) { }
>  #endif
> +#endif
>  
>  /*
>   * Pretty much the same than for the function tracer from which the selftest
> -- 
> 2.30.2
> 

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

  reply	other threads:[~2022-05-25 13:44 UTC|newest]

Thread overview: 60+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-05-18 13:16 [PATCH bpf-next v5 0/6] bpf trampoline for arm64 Xu Kuohai
2022-05-18 13:16 ` Xu Kuohai
2022-05-18 13:16 ` [PATCH bpf-next v5 1/6] arm64: ftrace: Add ftrace direct call support Xu Kuohai
2022-05-18 13:16   ` Xu Kuohai
2022-05-23  1:39   ` KP Singh
2022-05-23  1:39     ` KP Singh
2022-05-25 13:38   ` Mark Rutland
2022-05-25 13:38     ` Mark Rutland
2022-05-26  9:45     ` Xu Kuohai
2022-05-26  9:45       ` Xu Kuohai
2022-05-26 10:06       ` Mark Rutland
2022-05-26 10:06         ` Mark Rutland
2022-05-26 14:48         ` Xu Kuohai
2022-05-26 14:48           ` Xu Kuohai
2022-06-06 16:35           ` Mark Rutland
2022-06-06 16:35             ` Mark Rutland
2022-06-09  4:27             ` Xu Kuohai
2022-06-09  4:27               ` Xu Kuohai
2022-08-09 17:03               ` Florent Revest
2022-08-09 17:03                 ` Florent Revest
2022-08-10  8:10                 ` Xu Kuohai
2022-08-10  8:10                   ` Xu Kuohai
2022-05-18 13:16 ` [PATCH bpf-next v5 2/6] ftrace: Fix deadloop caused by direct call in ftrace selftest Xu Kuohai
2022-05-18 13:16   ` Xu Kuohai
2022-05-25 13:43   ` Mark Rutland [this message]
2022-05-25 13:43     ` Mark Rutland
2022-05-26  9:45     ` Xu Kuohai
2022-05-26  9:45       ` Xu Kuohai
2022-05-18 13:16 ` [PATCH bpf-next v5 3/6] bpf: Remove is_valid_bpf_tramp_flags() Xu Kuohai
2022-05-18 13:16   ` Xu Kuohai
2022-05-25 13:45   ` Mark Rutland
2022-05-25 13:45     ` Mark Rutland
2022-05-26  9:45     ` Xu Kuohai
2022-05-26  9:45       ` Xu Kuohai
2022-05-26 10:12       ` Mark Rutland
2022-05-26 10:12         ` Mark Rutland
2022-05-26 14:46         ` Xu Kuohai
2022-05-26 14:46           ` Xu Kuohai
2022-05-18 13:16 ` [PATCH bpf-next v5 4/6] bpf, arm64: Impelment bpf_arch_text_poke() for arm64 Xu Kuohai
2022-05-18 13:16   ` Xu Kuohai
2022-05-23  1:41   ` KP Singh
2022-05-23  1:41     ` KP Singh
2022-05-25 14:10   ` Mark Rutland
2022-05-25 14:10     ` Mark Rutland
2022-05-26  9:45     ` Xu Kuohai
2022-05-26  9:45       ` Xu Kuohai
2022-05-26 10:34       ` Mark Rutland
2022-05-26 10:34         ` Mark Rutland
2022-05-26 14:47         ` Xu Kuohai
2022-05-26 14:47           ` Xu Kuohai
2022-05-18 13:16 ` [PATCH bpf-next v5 5/6] bpf, arm64: bpf trampoline " Xu Kuohai
2022-05-18 13:16   ` Xu Kuohai
2022-05-20 21:18   ` Alexei Starovoitov
2022-05-20 21:18     ` Alexei Starovoitov
2022-05-23 16:09     ` Mark Rutland
2022-05-23 16:09       ` Mark Rutland
2022-05-23  1:36   ` KP Singh
2022-05-23  1:36     ` KP Singh
2022-05-18 13:16 ` [PATCH bpf-next v5 6/6] selftests/bpf: Fix trivial typo in fentry_fexit.c Xu Kuohai
2022-05-18 13:16   ` Xu Kuohai

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=Yo4ymwu92gM75/Z5@FVFF77S0Q05N \
    --to=mark.rutland@arm.com \
    --cc=andrii@kernel.org \
    --cc=ardb@kernel.org \
    --cc=ast@kernel.org \
    --cc=bp@alien8.de \
    --cc=bpf@vger.kernel.org \
    --cc=broonie@kernel.org \
    --cc=catalin.marinas@arm.com \
    --cc=daniel.kiss@arm.com \
    --cc=daniel@iogearbox.net \
    --cc=dave.hansen@linux.intel.com \
    --cc=davem@davemloft.net \
    --cc=delyank@fb.com \
    --cc=dsahern@kernel.org \
    --cc=hawk@kernel.org \
    --cc=hpa@zytor.com \
    --cc=john.fastabend@gmail.com \
    --cc=kafai@fb.com \
    --cc=kpsingh@kernel.org \
    --cc=kuba@kernel.org \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-kselftest@vger.kernel.org \
    --cc=maz@kernel.org \
    --cc=memxor@gmail.com \
    --cc=mingo@redhat.com \
    --cc=netdev@vger.kernel.org \
    --cc=pasha.tatashin@soleen.com \
    --cc=pcc@google.com \
    --cc=rostedt@goodmis.org \
    --cc=shuah@kernel.org \
    --cc=songliubraving@fb.com \
    --cc=steven.price@arm.com \
    --cc=sudeep.holla@arm.com \
    --cc=tglx@linutronix.de \
    --cc=will@kernel.org \
    --cc=x86@kernel.org \
    --cc=xukuohai@huawei.com \
    --cc=yhs@fb.com \
    --cc=yoshfuji@linux-ipv6.org \
    --cc=zlim.lnx@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.