From: "Jarkko Sakkinen" <jarkko@kernel.org>
To: "Mark Rutland" <mark.rutland@arm.com>, <linux-kernel@vger.kernel.org>
Cc: <anil.s.keshavamurthy@intel.com>, <aou@eecs.berkeley.edu>,
<catalin.marinas@arm.com>, <davem@davemloft.net>,
<linux-arm-kernel@lists.infradead.org>, <mhiramat@kernel.org>,
<naveen.n.rao@linux.ibm.com>, <palmer@dabbelt.com>,
<paul.walmsley@sifive.com>, <will@kernel.org>
Subject: Re: [PATCH v2 3/4] kprobes/treewide: Explicitly override alloc/free functions
Date: Wed, 03 Apr 2024 19:40:33 +0300 [thread overview]
Message-ID: <D0AMXFFH9WX8.3G0IX3GW60V7W@kernel.org> (raw)
In-Reply-To: <20240403150154.667649-4-mark.rutland@arm.com>
On Wed Apr 3, 2024 at 6:01 PM EEST, Mark Rutland wrote:
> Currently architectures can override kprobes_alloc_insn_page(), but
> kprobes_free_insn_page() is always implemented using module_memfree(),
> which might not be what an architecture needs, especially as we'd like
> to make it possible to use kprobes without requiring MODULES.
>
> It would be nicer if architectures either:
>
> (a) Used only the generic kprobes_alloc_insn_page() and
> kprobes_free_insn_page(), implicitly depending on MODULES.
>
> (b) Provided their own implementation of both kprobes_alloc_insn_page()
> and kprobes_free_insn_page(), handling the relevant dependencies
> themselves.
>
> This patch applies that split treewide:
>
> (a) Architectures using the generic kprobes_free_insn_page() and
> kprobes_free_insn_page() are left as-is. The __weak annotation is
> removed from the generic implementations so that accidental
> overrides/misuse can be detected easily.
>
> (b) Architectures which provide their own kprobes_free_insn_page() are
> given a matching implementation of kprobes_free_insn_page(), and
> select HAVE_KPROBES_ALLOC.
>
> This new Kconfig symbol will allow subsequent patches to relax the
> dependency on MODULES to (MODULES || HAVE_KPROBES_ALLOC) once other
> module dependencies in the core kprobes code are cleaned up.
>
> Architectures which use module_alloc() are given an implementation
> using module_memfree() along with an explicit dependency on MODULES.
>
> Architectures using __vmalloc_node_range() are given an
> implementation using vfree(). This loses the warning for
> in_interrupt(), but vfree() can handle this via vfree_atomic(), so
> the warning isn't necessary.
>
> On riscv, the allocator depends on !XIP_KERNEL, which is already a
> dependency for HAVE_KPROBES in arch/riscv/Kconfig.
>
> As of this patch arm64 and riscv have kprobe allocation functions which
> do not explicitly depend on MODULES. The core kprobes code still depends
> on MODULES.
>
> Signed-off-by: Mark Rutland <mark.rutland@arm.com>
> Cc: Albert Ou <aou@eecs.berkeley.edu>
> Cc: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
> Cc: Catalin Marinas <catalin.marinas@arm.com>
> Cc: David S. Miller <davem@davemloft.net>
> Cc: Jarkko Sakkinen <jarkko@kernel.org>
> Cc: Masami Hiramatsu <mhiramat@kernel.org>
> Cc: Naveen N. Rao <naveen.n.rao@linux.ibm.com>
> Cc: Palmer Dabbelt <palmer@dabbelt.com>
> Cc: Paul Walmsley <paul.walmsley@sifive.com>
> Cc: Will Deacon <will@kernel.org>
> ---
> arch/Kconfig | 3 +++
> arch/arm64/Kconfig | 1 +
> arch/arm64/kernel/probes/kprobes.c | 5 +++++
> arch/powerpc/Kconfig | 3 ++-
> arch/powerpc/kernel/kprobes.c | 5 +++++
> arch/riscv/Kconfig | 1 +
> arch/riscv/kernel/probes/kprobes.c | 5 +++++
> arch/s390/Kconfig | 3 ++-
> arch/s390/kernel/kprobes.c | 5 +++++
> arch/x86/Kconfig | 3 ++-
> arch/x86/kernel/kprobes/core.c | 5 +++++
> include/linux/kprobes.h | 1 +
> kernel/kprobes.c | 6 ++++--
> 13 files changed, 41 insertions(+), 5 deletions(-)
>
> diff --git a/arch/Kconfig b/arch/Kconfig
> index 9f066785bb71d..85bb59f7b8c07 100644
> --- a/arch/Kconfig
> +++ b/arch/Kconfig
> @@ -206,6 +206,9 @@ config HAVE_IOREMAP_PROT
> config HAVE_KPROBES
> bool
>
> +config HAVE_KPROBES_ALLOC
> + bool
> +
> config HAVE_KRETPROBES
> bool
>
> diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
> index 7b11c98b3e84b..bda7913d6c9b8 100644
> --- a/arch/arm64/Kconfig
> +++ b/arch/arm64/Kconfig
> @@ -233,6 +233,7 @@ config ARM64
> select HAVE_STACKPROTECTOR
> select HAVE_SYSCALL_TRACEPOINTS
> select HAVE_KPROBES
> + select HAVE_KPROBES_ALLOC
> select HAVE_KRETPROBES
> select HAVE_GENERIC_VDSO
> select HOTPLUG_CORE_SYNC_DEAD if HOTPLUG_CPU
> diff --git a/arch/arm64/kernel/probes/kprobes.c b/arch/arm64/kernel/probes/kprobes.c
> index 4b6ab7b1fa211..69d19a390cd48 100644
> --- a/arch/arm64/kernel/probes/kprobes.c
> +++ b/arch/arm64/kernel/probes/kprobes.c
> @@ -136,6 +136,11 @@ void *kprobes_alloc_insn_page(void)
> NUMA_NO_NODE, __builtin_return_address(0));
> }
>
> +void kprobes_free_insn_page(void *page)
> +{
> + vfree(page);
> +}
> +
> /* arm kprobe: install breakpoint in text */
> void __kprobes arch_arm_kprobe(struct kprobe *p)
> {
> diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig
> index 1c4be33736860..13e0fc51dcdcf 100644
> --- a/arch/powerpc/Kconfig
> +++ b/arch/powerpc/Kconfig
> @@ -254,7 +254,8 @@ config PPC
> select HAVE_KERNEL_LZMA if DEFAULT_UIMAGE
> select HAVE_KERNEL_LZO if DEFAULT_UIMAGE
> select HAVE_KERNEL_XZ if PPC_BOOK3S || 44x
> - select HAVE_KPROBES
> + select HAVE_KPROBES if MODULES
> + select HAVE_KPROBES_ALLOC
> select HAVE_KPROBES_ON_FTRACE
> select HAVE_KRETPROBES
> select HAVE_LD_DEAD_CODE_DATA_ELIMINATION if HAVE_OBJTOOL_MCOUNT && (!ARCH_USING_PATCHABLE_FUNCTION_ENTRY || (!CC_IS_GCC || GCC_VERSION >= 110100))
> diff --git a/arch/powerpc/kernel/kprobes.c b/arch/powerpc/kernel/kprobes.c
> index 0b297718d5de6..d0332aaebab09 100644
> --- a/arch/powerpc/kernel/kprobes.c
> +++ b/arch/powerpc/kernel/kprobes.c
> @@ -146,6 +146,11 @@ void *kprobes_alloc_insn_page(void)
> return NULL;
> }
>
> +void kprobes_free_insn_page(void *page)
> +{
> + module_memfree(page);
> +}
> +
> int arch_prepare_kprobe(struct kprobe *p)
> {
> int ret = 0;
> diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
> index be09c8836d56b..4e22549a522a5 100644
> --- a/arch/riscv/Kconfig
> +++ b/arch/riscv/Kconfig
> @@ -139,6 +139,7 @@ config RISCV
> select HAVE_GENERIC_VDSO if MMU && 64BIT
> select HAVE_IRQ_TIME_ACCOUNTING
> select HAVE_KPROBES if !XIP_KERNEL
> + select HAVE_KPROBES_ALLOC
> select HAVE_KPROBES_ON_FTRACE if !XIP_KERNEL
> select HAVE_KRETPROBES if !XIP_KERNEL
> # https://github.com/ClangBuiltLinux/linux/issues/1881
> diff --git a/arch/riscv/kernel/probes/kprobes.c b/arch/riscv/kernel/probes/kprobes.c
> index 75201ce721057..37fdfa952d999 100644
> --- a/arch/riscv/kernel/probes/kprobes.c
> +++ b/arch/riscv/kernel/probes/kprobes.c
> @@ -112,6 +112,11 @@ void *kprobes_alloc_insn_page(void)
> VM_FLUSH_RESET_PERMS, NUMA_NO_NODE,
> __builtin_return_address(0));
> }
> +
> +void kprobes_free_insn_page(void *page)
> +{
> + vfree(page);
> +}
> #endif
>
> /* install breakpoint in text */
> diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig
> index 8f01ada6845e3..635eddc3fce80 100644
> --- a/arch/s390/Kconfig
> +++ b/arch/s390/Kconfig
> @@ -193,7 +193,8 @@ config S390
> select HAVE_KERNEL_UNCOMPRESSED
> select HAVE_KERNEL_XZ
> select HAVE_KERNEL_ZSTD
> - select HAVE_KPROBES
> + select HAVE_KPROBES if MODULES
> + select HAVE_KPROBES_ALLOC
> select HAVE_KPROBES_ON_FTRACE
> select HAVE_KRETPROBES
> select HAVE_LIVEPATCH
> diff --git a/arch/s390/kernel/kprobes.c b/arch/s390/kernel/kprobes.c
> index 91ca4d501d4ef..a5b142b8eb0f7 100644
> --- a/arch/s390/kernel/kprobes.c
> +++ b/arch/s390/kernel/kprobes.c
> @@ -45,6 +45,11 @@ void *kprobes_alloc_insn_page(void)
> return page;
> }
>
> +void kprobes_free_insn_page(void *page)
> +{
> + module_memfree(page);
> +}
> +
> static void *alloc_s390_insn_page(void)
> {
> if (xchg(&insn_page_in_use, 1) == 1)
> diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
> index 4fff6ed46e902..0810cd0bdeca9 100644
> --- a/arch/x86/Kconfig
> +++ b/arch/x86/Kconfig
> @@ -240,7 +240,8 @@ config X86
> select HAVE_KERNEL_LZO
> select HAVE_KERNEL_XZ
> select HAVE_KERNEL_ZSTD
> - select HAVE_KPROBES
> + select HAVE_KPROBES if MODULES
> + select HAVE_KPROBES_ALLOC
> select HAVE_KPROBES_ON_FTRACE
> select HAVE_FUNCTION_ERROR_INJECTION
> select HAVE_KRETPROBES
> diff --git a/arch/x86/kernel/kprobes/core.c b/arch/x86/kernel/kprobes/core.c
> index 7f01bbbfa9e2a..5f093b94d9b40 100644
> --- a/arch/x86/kernel/kprobes/core.c
> +++ b/arch/x86/kernel/kprobes/core.c
> @@ -508,6 +508,11 @@ void *kprobes_alloc_insn_page(void)
> return page;
> }
>
> +void kprobes_free_insn_page(void *page)
> +{
> + module_memfree(page);
> +}
> +
> /* Kprobe x86 instruction emulation - only regs->ip or IF flag modifiers */
>
> static void kprobe_emulate_ifmodifiers(struct kprobe *p, struct pt_regs *regs)
> diff --git a/include/linux/kprobes.h b/include/linux/kprobes.h
> index ad4b561100f9e..651c807727bea 100644
> --- a/include/linux/kprobes.h
> +++ b/include/linux/kprobes.h
> @@ -431,6 +431,7 @@ int enable_kprobe(struct kprobe *kp);
> void dump_kprobe(struct kprobe *kp);
>
> void *kprobes_alloc_insn_page(void);
> +void kprobes_free_insn_page(void *page);
>
> void *kprobes_alloc_optinsn_page(void);
> void kprobes_free_optinsn_page(void *page);
> diff --git a/kernel/kprobes.c b/kernel/kprobes.c
> index 35adf56430c9b..fa2ee4e59eca2 100644
> --- a/kernel/kprobes.c
> +++ b/kernel/kprobes.c
> @@ -110,7 +110,8 @@ enum kprobe_slot_state {
> SLOT_USED = 2,
> };
>
> -void __weak *kprobes_alloc_insn_page(void)
> +#ifndef CONFIG_HAVE_KPROBES_ALLOC
> +void *kprobes_alloc_insn_page(void)
> {
> /*
> * Use module_alloc() so this page is within +/- 2GB of where the
> @@ -121,10 +122,11 @@ void __weak *kprobes_alloc_insn_page(void)
> return module_alloc(PAGE_SIZE);
> }
>
> -static void kprobes_free_insn_page(void *page)
> +void kprobes_free_insn_page(void *page)
> {
> module_memfree(page);
> }
> +#endif
>
> struct kprobe_insn_cache kprobe_insn_slots = {
> .mutex = __MUTEX_INITIALIZER(kprobe_insn_slots.mutex),
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
BR, Jarkko
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
next prev parent reply other threads:[~2024-04-03 16:40 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-04-03 15:01 [PATCH v2 0/4] kprobes: permit use without modules Mark Rutland
2024-04-03 15:01 ` [PATCH v2 1/4] arm64: patching: always use fixmap Mark Rutland
2024-04-03 16:20 ` Jarkko Sakkinen
2024-04-03 16:51 ` Mark Rutland
2024-04-04 14:48 ` Jarkko Sakkinen
2024-04-03 17:52 ` Catalin Marinas
2024-04-03 23:44 ` Masami Hiramatsu
2024-04-03 15:01 ` [PATCH v2 2/4] kprobes/treewide: Add kprobes_ prefix to insn alloc/free functions Mark Rutland
2024-04-03 16:39 ` Jarkko Sakkinen
2024-04-03 15:01 ` [PATCH v2 3/4] kprobes/treewide: Explicitly override " Mark Rutland
2024-04-03 16:40 ` Jarkko Sakkinen [this message]
2024-04-03 15:01 ` [PATCH v2 4/4] kprobes: Remove core dependency on modules Mark Rutland
2024-04-03 16:41 ` Jarkko Sakkinen
2024-04-04 8:15 ` Jarkko Sakkinen
2024-04-04 15:18 ` Jarkko Sakkinen
2024-04-04 16:10 ` Masami Hiramatsu
2024-04-04 16:47 ` Mark Rutland
2024-04-05 3:13 ` Masami Hiramatsu
2024-04-13 20:16 ` Jarkko Sakkinen
2024-04-11 22:54 ` Jarkko Sakkinen
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=D0AMXFFH9WX8.3G0IX3GW60V7W@kernel.org \
--to=jarkko@kernel.org \
--cc=anil.s.keshavamurthy@intel.com \
--cc=aou@eecs.berkeley.edu \
--cc=catalin.marinas@arm.com \
--cc=davem@davemloft.net \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mark.rutland@arm.com \
--cc=mhiramat@kernel.org \
--cc=naveen.n.rao@linux.ibm.com \
--cc=palmer@dabbelt.com \
--cc=paul.walmsley@sifive.com \
--cc=will@kernel.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).