From: Mark Rutland <mark.rutland@arm.com>
To: Youling Tang <tangyouling@loongson.cn>
Cc: Huacai Chen <chenhuacai@kernel.org>,
Steven Rostedt <rostedt@goodmis.org>,
Masami Hiramatsu <mhiramat@kernel.org>,
WANG Xuerui <kernel@xen0n.name>,
linux-kernel@vger.kernel.org, linux-trace-kernel@vger.kernel.org,
loongarch@lists.linux.dev, loongson-kernel@lists.loongnix.cn,
Qing Zhang <zhangqing@loongson.cn>
Subject: Re: [PATCH 3/5] LoongArch: ftrace: Add direct call support
Date: Wed, 26 Apr 2023 14:43:17 +0100 [thread overview]
Message-ID: <ZEkqdfNk4YKVvxxM@FVFF77S0Q05N> (raw)
In-Reply-To: <1682473754-25077-4-git-send-email-tangyouling@loongson.cn>
Hi,
On Wed, Apr 26, 2023 at 09:49:12AM +0800, Youling Tang wrote:
> select the DYNAMIC_FTRACE_WITH_DIRECT_CALLS to provide the
> register_ftrace_direct[_multi] interfaces allowing users to register
> the customed trampoline (direct_caller) as the mcount for one or
> more target functions. And modify_ftrace_direct[_multi] are also
> provided for modifying direct_caller.
>
> There are a few cases to distinguish:
> - If a direct call ops is the only one tracing a function:
> - If the direct called trampoline is within the reach of a 'bl'
> instruction
> -> the ftrace patchsite jumps to the trampoline
> - Else
> -> the ftrace patchsite jumps to the ftrace_regs_caller trampoline
> which reads the ops pointer in the patchsite and jumps to the
> direct call address stored in the ops
I think you forgot to update this wording; there's no ops pointer in the
patchsite as you don't implement DYNAMIC_FTRACE_WITH_CALL_OPS.
I think you can delete the "Else" case here, and replace the above with:
- If a direct call ops is the only one tracing a function AND the direct called
trampoline is within the reach of a 'bl' instruction
-> the ftrace patchsite jumps to the trampoline
> - Else
> -> the ftrace patchsite jumps to the ftrace_regs_caller trampoline and
> its ops literal points to ftrace_list_ops so it iterates over all
> registered ftrace ops, including the direct call ops and calls its
> call_direct_funcs handler which stores the direct called trampoline's
> address in the ftrace_regs and the ftrace_regs_caller trampoline will
Likewise here, there's no "ops literal" associated with the patchsite.
Otherwise, this looks sane to me.
Mark.
> return to that address instead of returning to the traced function
>
> Signed-off-by: Qing Zhang <zhangqing@loongson.cn>
> Signed-off-by: Youling Tang <tangyouling@loongson.cn>
> ---
> arch/loongarch/Kconfig | 1 +
> arch/loongarch/include/asm/ftrace.h | 12 ++++++++++++
> arch/loongarch/kernel/ftrace_dyn.c | 8 ++++++++
> arch/loongarch/kernel/mcount_dyn.S | 7 ++++++-
> 4 files changed, 27 insertions(+), 1 deletion(-)
>
> diff --git a/arch/loongarch/Kconfig b/arch/loongarch/Kconfig
> index 3ddde336e6a5..d6068a88d53f 100644
> --- a/arch/loongarch/Kconfig
> +++ b/arch/loongarch/Kconfig
> @@ -92,6 +92,7 @@ config LOONGARCH
> select HAVE_DMA_CONTIGUOUS
> select HAVE_DYNAMIC_FTRACE
> select HAVE_DYNAMIC_FTRACE_WITH_ARGS
> + select HAVE_DYNAMIC_FTRACE_WITH_DIRECT_CALLS
> select HAVE_DYNAMIC_FTRACE_WITH_REGS
> select HAVE_EBPF_JIT
> select HAVE_EFFICIENT_UNALIGNED_ACCESS if !ARCH_STRICT_ALIGN
> diff --git a/arch/loongarch/include/asm/ftrace.h b/arch/loongarch/include/asm/ftrace.h
> index 3418d32d4fc7..f789e680f633 100644
> --- a/arch/loongarch/include/asm/ftrace.h
> +++ b/arch/loongarch/include/asm/ftrace.h
> @@ -57,6 +57,18 @@ static __always_inline struct pt_regs *arch_ftrace_get_regs(struct ftrace_regs *
> #define ftrace_graph_func ftrace_graph_func
> void ftrace_graph_func(unsigned long ip, unsigned long parent_ip,
> struct ftrace_ops *op, struct ftrace_regs *fregs);
> +
> +#ifdef CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS
> +static inline void
> +__arch_ftrace_set_direct_caller(struct pt_regs *regs, unsigned long addr)
> +{
> + regs->regs[13] = addr; /* t1 */
> +}
> +
> +#define arch_ftrace_set_direct_caller(fregs, addr) \
> + __arch_ftrace_set_direct_caller(&(fregs)->regs, addr)
> +#endif /* CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS */
> +
> #endif
>
> #endif /* __ASSEMBLY__ */
> diff --git a/arch/loongarch/kernel/ftrace_dyn.c b/arch/loongarch/kernel/ftrace_dyn.c
> index 3cc4f8159f48..4062a7e63137 100644
> --- a/arch/loongarch/kernel/ftrace_dyn.c
> +++ b/arch/loongarch/kernel/ftrace_dyn.c
> @@ -66,6 +66,14 @@ static bool ftrace_find_callable_addr(struct dyn_ftrace *rec, struct module *mod
> unsigned long pc = rec->ip + LOONGARCH_INSN_SIZE;
> struct plt_entry *plt;
>
> + /*
> + * If a custom trampoline is unreachable, rely on the ftrace_regs_caller
> + * trampoline which knows how to indirectly reach that trampoline
> + * through ops->direct_call.
> + */
> + if (*addr != FTRACE_ADDR && *addr != FTRACE_REGS_ADDR && !reachable_by_bl(*addr, pc))
> + *addr = FTRACE_REGS_ADDR;
> +
> /*
> * When the target is within range of the 'bl' instruction, use 'addr'
> * as-is and branch to that directly.
> diff --git a/arch/loongarch/kernel/mcount_dyn.S b/arch/loongarch/kernel/mcount_dyn.S
> index bbabf06244c2..1008fbc3cabc 100644
> --- a/arch/loongarch/kernel/mcount_dyn.S
> +++ b/arch/loongarch/kernel/mcount_dyn.S
> @@ -42,7 +42,6 @@
> .if \allregs
> PTR_S tp, sp, PT_R2
> PTR_S t0, sp, PT_R12
> - PTR_S t1, sp, PT_R13
> PTR_S t2, sp, PT_R14
> PTR_S t3, sp, PT_R15
> PTR_S t4, sp, PT_R16
> @@ -64,6 +63,8 @@
> PTR_S zero, sp, PT_R0
> .endif
> PTR_S ra, sp, PT_ERA /* Save trace function ra at PT_ERA */
> + move t1, zero
> + PTR_S t1, sp, PT_R13
> PTR_ADDI t8, sp, PT_SIZE
> PTR_S t8, sp, PT_R3
> .endm
> @@ -104,8 +105,12 @@ ftrace_common_return:
> PTR_L a7, sp, PT_R11
> PTR_L fp, sp, PT_R22
> PTR_L t0, sp, PT_ERA
> + PTR_L t1, sp, PT_R13
> PTR_ADDI sp, sp, PT_SIZE
> + bnez t1,.Ldirect
> jr t0
> +.Ldirect:
> + jr t1
> SYM_CODE_END(ftrace_common)
>
> SYM_CODE_START(ftrace_caller)
> --
> 2.37.1
>
next prev parent reply other threads:[~2023-04-26 13:43 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-04-26 1:49 [PATCH 0/5] LoongArch: ftrace: Add direct call support code simplification Youling Tang
2023-04-26 1:49 ` [PATCH 1/5] LoongArch: Fix build error if CONFIG_DYNAMIC_FTRACE_WITH_REGS is not set Youling Tang
2023-04-26 1:49 ` [PATCH 2/5] LoongArch: ftrace: Implement ftrace_find_callable_addr() to simplify code Youling Tang
2023-04-26 1:49 ` [PATCH 3/5] LoongArch: ftrace: Add direct call support Youling Tang
2023-04-26 13:43 ` Mark Rutland [this message]
2023-04-26 14:30 ` Youling Tang
2023-04-26 1:49 ` [PATCH 4/5] LoongArch: ftrace: Add direct call trampoline samples support Youling Tang
2023-04-26 1:49 ` [PATCH 5/5] LoongArch: Abstract DYNAMIC_FTRACE_WITH_ARGS accesses Youling Tang
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=ZEkqdfNk4YKVvxxM@FVFF77S0Q05N \
--to=mark.rutland@arm.com \
--cc=chenhuacai@kernel.org \
--cc=kernel@xen0n.name \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-trace-kernel@vger.kernel.org \
--cc=loongarch@lists.linux.dev \
--cc=loongson-kernel@lists.loongnix.cn \
--cc=mhiramat@kernel.org \
--cc=rostedt@goodmis.org \
--cc=tangyouling@loongson.cn \
--cc=zhangqing@loongson.cn \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®