From: Christophe Leroy <christophe.leroy@csgroup.eu>
To: Naveen N Rao <naveen@kernel.org>,
"linuxppc-dev@lists.ozlabs.org" <linuxppc-dev@lists.ozlabs.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Subject: Re: [PATCH 16/17] powerpc/ftrace: Add support for -fpatchable-function-entry
Date: Fri, 23 Jun 2023 05:37:59 +0000 [thread overview]
Message-ID: <66372e0f-a5e0-d883-e9b3-a837a17cd593@csgroup.eu> (raw)
In-Reply-To: <68586d22981a2c3bb45f27a2b621173d10a7d092.1687166935.git.naveen@kernel.org>
Le 19/06/2023 à 11:47, Naveen N Rao a écrit :
> GCC v13.1 updated support for -fpatchable-function-entry on ppc64le to
> emit nops after the local entry point, rather than before it. This
> allows us to use this in the kernel for ftrace purposes. A new script is
> added under arch/powerpc/tools/ to help detect if nops are emitted after
> the function local entry point, or before the global entry point.
>
> With -fpatchable-function-entry, we no longer have the profiling
> instructions generated at function entry, so we only need to validate
> the presence of two nops at the ftrace location in ftrace_init_nop(). We
> patch the preceding instruction with 'mflr r0' to match the
> -mprofile-kernel ABI for subsequent ftrace use.
>
> This changes the profiling instructions used on ppc32. The default -pg
> option emits an additional 'stw' instruction after 'mflr r0' and before
> the branch to _mcount 'bl _mcount'. This is very similar to the original
> -mprofile-kernel implementation on ppc64le, where an additional 'std'
> instruction was used to save LR to its save location in the caller's
> stackframe. Subsequently, this additional store was removed in later
> compiler versions for performance reasons. The same reasons apply for
> ppc32 so we only patch in a 'mflr r0'.
>
> Signed-off-by: Naveen N Rao <naveen@kernel.org>
Reviewed-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Nit below
> ---
> arch/powerpc/Kconfig | 14 +++++++---
> arch/powerpc/Makefile | 5 ++++
> arch/powerpc/include/asm/ftrace.h | 6 +++--
> arch/powerpc/include/asm/vermagic.h | 4 ++-
> arch/powerpc/kernel/module_64.c | 2 +-
> arch/powerpc/kernel/trace/ftrace.c | 14 ++++++++--
> arch/powerpc/kernel/trace/ftrace_entry.S | 2 ++
> .../gcc-check-fpatchable-function-entry.sh | 26 +++++++++++++++++++
> 8 files changed, 64 insertions(+), 9 deletions(-)
> create mode 100755 arch/powerpc/tools/gcc-check-fpatchable-function-entry.sh
>
> diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig
> index bff5820b7cda14..9352d8e68152e1 100644
> --- a/arch/powerpc/Kconfig
> +++ b/arch/powerpc/Kconfig
> @@ -187,6 +187,7 @@ config PPC
> select DYNAMIC_FTRACE if FUNCTION_TRACER
> select EDAC_ATOMIC_SCRUB
> select EDAC_SUPPORT
> + select FTRACE_MCOUNT_USE_PATCHABLE_FUNCTION_ENTRY if ARCH_USING_PATCHABLE_FUNCTION_ENTRY
> select GENERIC_ATOMIC64 if PPC32
> select GENERIC_CLOCKEVENTS_BROADCAST if SMP
> select GENERIC_CMOS_UPDATE
> @@ -227,8 +228,8 @@ config PPC
> select HAVE_DEBUG_KMEMLEAK
> select HAVE_DEBUG_STACKOVERFLOW
> select HAVE_DYNAMIC_FTRACE
> - select HAVE_DYNAMIC_FTRACE_WITH_ARGS if MPROFILE_KERNEL || PPC32
> - select HAVE_DYNAMIC_FTRACE_WITH_REGS if MPROFILE_KERNEL || PPC32
> + select HAVE_DYNAMIC_FTRACE_WITH_ARGS if ARCH_USING_PATCHABLE_FUNCTION_ENTRY || MPROFILE_KERNEL || PPC32
> + select HAVE_DYNAMIC_FTRACE_WITH_REGS if ARCH_USING_PATCHABLE_FUNCTION_ENTRY || MPROFILE_KERNEL || PPC32
ARCH_USING_PATCHABLE_FUNCTION_ENTRY defaults to y if PPC32, so you can
remove PPC32 from the condition here.
> select HAVE_EBPF_JIT
> select HAVE_EFFICIENT_UNALIGNED_ACCESS
> select HAVE_FAST_GUP
> @@ -256,7 +257,7 @@ config PPC
> select HAVE_MOD_ARCH_SPECIFIC
> select HAVE_NMI if PERF_EVENTS || (PPC64 && PPC_BOOK3S)
> select HAVE_OPTPROBES
> - select HAVE_OBJTOOL if PPC32 || MPROFILE_KERNEL
> + select HAVE_OBJTOOL if ARCH_USING_PATCHABLE_FUNCTION_ENTRY || MPROFILE_KERNEL || PPC32
Same
> select HAVE_OBJTOOL_MCOUNT if HAVE_OBJTOOL
> select HAVE_PERF_EVENTS
> select HAVE_PERF_EVENTS_NMI if PPC64
> @@ -550,6 +551,13 @@ config MPROFILE_KERNEL
> depends on PPC64 && CPU_LITTLE_ENDIAN && FUNCTION_TRACER
> def_bool $(success,$(srctree)/arch/powerpc/tools/gcc-check-mprofile-kernel.sh $(CC) -I$(srctree)/include -D__KERNEL__)
>
> +config ARCH_USING_PATCHABLE_FUNCTION_ENTRY
> + depends on FUNCTION_TRACER && (PPC32 || PPC64_ELF_ABI_V2)
> + depends on $(cc-option,-fpatchable-function-entry=2)
> + def_bool y if PPC32
> + def_bool $(success,$(srctree)/arch/powerpc/tools/gcc-check-fpatchable-function-entry.sh $(CC) -mlittle-endian) if PPC64 && CPU_LITTLE_ENDIAN
> + def_bool $(success,$(srctree)/arch/powerpc/tools/gcc-check-fpatchable-function-entry.sh $(CC) -mbig-endian) if PPC64 && CPU_BIG_ENDIAN
> +
> config HOTPLUG_CPU
> bool "Support for enabling/disabling CPUs"
> depends on SMP && (PPC_PSERIES || \
> diff --git a/arch/powerpc/Makefile b/arch/powerpc/Makefile
> index dca73f673d7046..de39478b1c9e9f 100644
> --- a/arch/powerpc/Makefile
> +++ b/arch/powerpc/Makefile
> @@ -148,11 +148,16 @@ CFLAGS-$(CONFIG_PPC32) += $(call cc-option, $(MULTIPLEWORD))
> CFLAGS-$(CONFIG_PPC32) += $(call cc-option,-mno-readonly-in-sdata)
>
> ifdef CONFIG_FUNCTION_TRACER
> +ifdef CONFIG_ARCH_USING_PATCHABLE_FUNCTION_ENTRY
> +KBUILD_CPPFLAGS += -DCC_USING_PATCHABLE_FUNCTION_ENTRY
> +CC_FLAGS_FTRACE := -fpatchable-function-entry=2
> +else
> CC_FLAGS_FTRACE := -pg
> ifdef CONFIG_MPROFILE_KERNEL
> CC_FLAGS_FTRACE += -mprofile-kernel
> endif
> endif
> +endif
>
> CFLAGS-$(CONFIG_TARGET_CPU_BOOL) += -mcpu=$(CONFIG_TARGET_CPU)
> AFLAGS-$(CONFIG_TARGET_CPU_BOOL) += -mcpu=$(CONFIG_TARGET_CPU)
> diff --git a/arch/powerpc/include/asm/ftrace.h b/arch/powerpc/include/asm/ftrace.h
> index ef9f0b97670d1c..9e5a39b6a3114b 100644
> --- a/arch/powerpc/include/asm/ftrace.h
> +++ b/arch/powerpc/include/asm/ftrace.h
> @@ -11,7 +11,7 @@
> #define HAVE_FUNCTION_GRAPH_RET_ADDR_PTR
>
> /* Ignore unused weak functions which will have larger offsets */
> -#ifdef CONFIG_MPROFILE_KERNEL
> +#if defined(CONFIG_MPROFILE_KERNEL) || defined(CONFIG_ARCH_USING_PATCHABLE_FUNCTION_ENTRY)
> #define FTRACE_MCOUNT_MAX_OFFSET 16
> #elif defined(CONFIG_PPC32)
> #define FTRACE_MCOUNT_MAX_OFFSET 8
> @@ -22,7 +22,9 @@ extern void _mcount(void);
>
> static inline unsigned long ftrace_call_adjust(unsigned long addr)
> {
> - /* relocation of mcount call site is the same as the address */
> + if (IS_ENABLED(CONFIG_ARCH_USING_PATCHABLE_FUNCTION_ENTRY))
> + addr += MCOUNT_INSN_SIZE;
> +
> return addr;
> }
>
> diff --git a/arch/powerpc/include/asm/vermagic.h b/arch/powerpc/include/asm/vermagic.h
> index b054a8576e5deb..6f250fe506bd1c 100644
> --- a/arch/powerpc/include/asm/vermagic.h
> +++ b/arch/powerpc/include/asm/vermagic.h
> @@ -2,7 +2,9 @@
> #ifndef _ASM_VERMAGIC_H
> #define _ASM_VERMAGIC_H
>
> -#ifdef CONFIG_MPROFILE_KERNEL
> +#ifdef CONFIG_ARCH_USING_PATCHABLE_FUNCTION_ENTRY
> +#define MODULE_ARCH_VERMAGIC_FTRACE "patchable-function-entry "
> +#elif defined(CONFIG_MPROFILE_KERNEL)
> #define MODULE_ARCH_VERMAGIC_FTRACE "mprofile-kernel "
> #else
> #define MODULE_ARCH_VERMAGIC_FTRACE ""
> diff --git a/arch/powerpc/kernel/module_64.c b/arch/powerpc/kernel/module_64.c
> index 92570289ce08f5..7112adc597a80b 100644
> --- a/arch/powerpc/kernel/module_64.c
> +++ b/arch/powerpc/kernel/module_64.c
> @@ -465,7 +465,7 @@ int module_frob_arch_sections(Elf64_Ehdr *hdr,
> return 0;
> }
>
> -#ifdef CONFIG_MPROFILE_KERNEL
> +#if defined(CONFIG_MPROFILE_KERNEL) || defined(CONFIG_ARCH_USING_PATCHABLE_FUNCTION_ENTRY)
>
> static u32 stub_insns[] = {
> #ifdef CONFIG_PPC_KERNEL_PCREL
> diff --git a/arch/powerpc/kernel/trace/ftrace.c b/arch/powerpc/kernel/trace/ftrace.c
> index cf9dce77527920..82010629cf887c 100644
> --- a/arch/powerpc/kernel/trace/ftrace.c
> +++ b/arch/powerpc/kernel/trace/ftrace.c
> @@ -220,7 +220,12 @@ int ftrace_init_nop(struct module *mod, struct dyn_ftrace *rec)
> int ret = 0;
>
> /* Verify instructions surrounding the ftrace location */
> - if (IS_ENABLED(CONFIG_PPC32)) {
> + if (IS_ENABLED(CONFIG_ARCH_USING_PATCHABLE_FUNCTION_ENTRY)) {
> + /* Expect nops */
> + ret = ftrace_validate_inst(ip - 4, ppc_inst(PPC_RAW_NOP()));
> + if (!ret)
> + ret = ftrace_validate_inst(ip, ppc_inst(PPC_RAW_NOP()));
> + } else if (IS_ENABLED(CONFIG_PPC32)) {
> /* Expected sequence: 'mflr r0', 'stw r0,4(r1)', 'bl _mcount' */
> ret = ftrace_validate_inst(ip - 8, ppc_inst(PPC_RAW_MFLR(_R0)));
> if (!ret)
> @@ -250,7 +255,12 @@ int ftrace_init_nop(struct module *mod, struct dyn_ftrace *rec)
> /* Nop-out the ftrace location */
> new = ppc_inst(PPC_RAW_NOP());
> addr = MCOUNT_ADDR;
> - if (is_offset_in_branch_range(addr - ip)) {
> + if (IS_ENABLED(CONFIG_ARCH_USING_PATCHABLE_FUNCTION_ENTRY)) {
> + /* we instead patch-in the 'mflr r0' */
> + old = ppc_inst(PPC_RAW_NOP());
> + new = ppc_inst(PPC_RAW_MFLR(_R0));
> + ret = ftrace_modify_code(ip - 4, old, new);
> + } else if (is_offset_in_branch_range(addr - ip)) {
> /* Within range */
> old = ftrace_create_branch_inst(ip, addr, 1);
> ret = ftrace_modify_code(ip, old, new);
> diff --git a/arch/powerpc/kernel/trace/ftrace_entry.S b/arch/powerpc/kernel/trace/ftrace_entry.S
> index e8339706e735b1..bab3ab1368a33f 100644
> --- a/arch/powerpc/kernel/trace/ftrace_entry.S
> +++ b/arch/powerpc/kernel/trace/ftrace_entry.S
> @@ -250,6 +250,7 @@ livepatch_handler:
> blr
> #endif /* CONFIG_LIVEPATCH */
>
> +#ifndef CONFIG_ARCH_USING_PATCHABLE_FUNCTION_ENTRY
> _GLOBAL(mcount)
> _GLOBAL(_mcount)
> EXPORT_SYMBOL(_mcount)
> @@ -257,6 +258,7 @@ EXPORT_SYMBOL(_mcount)
> mtctr r12
> mtlr r0
> bctr
> +#endif
>
> #ifdef CONFIG_FUNCTION_GRAPH_TRACER
> _GLOBAL(return_to_handler)
> diff --git a/arch/powerpc/tools/gcc-check-fpatchable-function-entry.sh b/arch/powerpc/tools/gcc-check-fpatchable-function-entry.sh
> new file mode 100755
> index 00000000000000..06706903503b6c
> --- /dev/null
> +++ b/arch/powerpc/tools/gcc-check-fpatchable-function-entry.sh
> @@ -0,0 +1,26 @@
> +#!/bin/bash
> +# SPDX-License-Identifier: GPL-2.0
> +
> +set -e
> +set -o pipefail
> +
> +# To debug, uncomment the following line
> +# set -x
> +
> +# Output from -fpatchable-function-entry can only vary on ppc64 elfv2, so this
> +# should not be invoked for other targets. Therefore we can pass in -m64 and
> +# -mabi explicitly, to take care of toolchains defaulting to other targets.
> +
> +# Test whether the compile option -fpatchable-function-entry exists and
> +# generates appropriate code
> +echo "int func() { return 0; }" | \
> + $* -m64 -mabi=elfv2 -S -x c -O2 -fpatchable-function-entry=2 - -o - 2> /dev/null | \
> + grep -q "__patchable_function_entries"
> +
> +# Test whether nops are generated after the local entry point
> +echo "int x; int func() { return x; }" | \
> + $* -m64 -mabi=elfv2 -S -x c -O2 -fpatchable-function-entry=2 - -o - 2> /dev/null | \
> + awk 'BEGIN { RS = ";" } /\.localentry.*nop.*\n[[:space:]]*nop/ { print $0 }' | \
> + grep -q "func:"
> +
> +exit 0
next prev parent reply other threads:[~2023-06-23 5:39 UTC|newest]
Thread overview: 39+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-06-19 9:47 [PATCH 00/17] powerpc/ftrace: refactor and add support for -fpatchable-function-entry Naveen N Rao
2023-06-19 9:47 ` [PATCH 01/17] powerpc/ftrace: Fix dropping weak symbols with older toolchains Naveen N Rao
2023-06-23 5:10 ` Christophe Leroy
2023-06-19 9:47 ` [PATCH 02/17] powerpc/module: Remove unused .ftrace.tramp section Naveen N Rao
2023-06-23 5:12 ` Christophe Leroy
2023-06-19 9:47 ` [PATCH 03/17] powerpc64/ftrace: Move ELFv1 and -pg support code into a separate file Naveen N Rao
2023-06-23 5:13 ` Christophe Leroy
2023-06-19 9:47 ` [PATCH 04/17] powerpc/ftrace: Simplify function_graph support in ftrace.c Naveen N Rao
2023-06-23 5:14 ` Christophe Leroy
2023-06-19 9:47 ` [PATCH 05/17] powerpc/ftrace: Use FTRACE_REGS_ADDR to identify the correct ftrace trampoline Naveen N Rao
2023-06-23 5:15 ` Christophe Leroy
2023-06-19 9:47 ` [PATCH 06/17] powerpc/ftrace: Extend ftrace support for large kernels to ppc32 Naveen N Rao
2023-06-23 5:21 ` Christophe Leroy
2023-06-19 9:47 ` [PATCH 07/17] powerpc/ftrace: Consolidate ftrace support into fewer files Naveen N Rao
2023-06-23 5:25 ` Christophe Leroy
2023-06-28 7:32 ` Naveen N Rao
2023-06-19 9:47 ` [PATCH 08/17] powerpc/ftrace: Refactor ftrace_modify_code() Naveen N Rao
2023-06-23 5:27 ` Christophe Leroy
2023-06-19 9:47 ` [PATCH 09/17] powerpc/ftrace: Stop re-purposing linker generated long branches for ftrace Naveen N Rao
2023-06-23 5:28 ` Christophe Leroy
2023-06-19 9:47 ` [PATCH 10/17] powerpc/ftrace: Add separate ftrace_init_nop() with additional validation Naveen N Rao
2023-06-23 5:29 ` Christophe Leroy
2023-06-19 9:47 ` [PATCH 11/17] powerpc/ftrace: Simplify ftrace_make_nop() Naveen N Rao
2023-06-23 5:30 ` Christophe Leroy
2023-06-19 9:47 ` [PATCH 12/17] powerpc/ftrace: Simplify ftrace_make_call() Naveen N Rao
2023-06-23 5:30 ` Christophe Leroy
2023-06-19 9:47 ` [PATCH 13/17] powerpc/ftrace: Simplify ftrace_modify_call() Naveen N Rao
2023-06-23 5:31 ` Christophe Leroy
2023-06-19 9:47 ` [PATCH 14/17] powerpc/ftrace: Replace use of ftrace_call_replace() with ftrace_create_branch_inst() Naveen N Rao
2023-06-23 5:32 ` Christophe Leroy
2023-06-19 9:47 ` [PATCH 15/17] powerpc/ftrace: Implement ftrace_replace_code() Naveen N Rao
2023-06-23 5:32 ` Christophe Leroy
2023-06-19 9:47 ` [PATCH 16/17] powerpc/ftrace: Add support for -fpatchable-function-entry Naveen N Rao
2023-06-23 5:37 ` Christophe Leroy [this message]
2023-06-28 7:40 ` Naveen N Rao
2023-06-19 9:47 ` [PATCH 17/17] powerpc/ftrace: Create a dummy stackframe to fix stack unwind Naveen N Rao
2023-06-23 5:40 ` Christophe Leroy
2023-06-28 7:43 ` Naveen N Rao
2023-08-23 11:55 ` [PATCH 00/17] powerpc/ftrace: refactor and add support for -fpatchable-function-entry Michael Ellerman
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=66372e0f-a5e0-d883-e9b3-a837a17cd593@csgroup.eu \
--to=christophe.leroy@csgroup.eu \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=naveen@kernel.org \
--cc=rostedt@goodmis.org \
/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;
as well as URLs for NNTP newsgroup(s).