[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Zg07hrb_RMUu2tq7@FVFF77S0Q05N>
Date: Wed, 3 Apr 2024 12:20:38 +0100
From: Mark Rutland <mark.rutland@....com>
To: linux-kernel@...r.kernel.org
Cc: agordeev@...ux.ibm.com, anil.s.keshavamurthy@...el.com,
aou@...s.berkeley.edu, bp@...en8.de, catalin.marinas@....com,
dave.hansen@...ux.intel.com, davem@...emloft.net, gor@...ux.ibm.com,
hca@...ux.ibm.com, jarkko@...nel.org, jcalvinowens@...il.com,
linux-arm-kernel@...ts.infradead.org, mhiramat@...nel.org,
mingo@...hat.com, mpe@...erman.id.au, naveen.n.rao@...ux.ibm.com,
palmer@...belt.com, paul.walmsley@...ive.com, tglx@...utronix.de,
will@...nel.org
Subject: Re: [PATCH 4/4] kprobes: Remove core dependency on modules
On Tue, Mar 26, 2024 at 04:36:24PM +0000, Mark Rutland wrote:
> From: Jarkko Sakkinen <jarkko@...nel.org>
>
> Tracing with kprobes while running a monolithic kernel is currently
> impossible because KPROBES depends on MODULES. While this dependency is
> necessary when KPROBES_USE_MODULE_ALLOC=y, all the other module-specific
> code only exist to handle the case when MODULES=y, and can be hidden
> behind ifdeffery.
>
> Add the necessary ifdeffery, and remove the dependency on MODULES=N when
> KPROBES_USE_MODULE_ALLOC=n.
>
> Currently this allows kprobes to be used when CONFIG_MODULES=n on arm64
> and riscv, and other architectures can enable support by implementing
> their own kprobes_alloc_insn_page() and kprobes_free_insn_page() which
> do not depend on MODULES.
>
> Signed-off-by: Jarkko Sakkinen <jarkko@...nel.org>
> Link: https://lore.kernel.org/all/20240326012102.27438-1-jarkko@kernel.org/
> [Mark: Remove execmem changes, depend on !KPROBES_USE_MODULE_ALLOC]
> Signed-off-by: Mark Rutland <mark.rutland@....com>
> Cc: Anil S Keshavamurthy <anil.s.keshavamurthy@...el.com>
> Cc: David S. Miller <davem@...emloft.net>
> Cc: Jarkko Sakkinen <jarkko@...nel.org>
> Cc: Masami Hiramatsu <mhiramat@...nel.org>
> Cc: Naveen N. Rao <naveen.n.rao@...ux.ibm.com>
> ---
> arch/Kconfig | 2 +-
> kernel/kprobes.c | 12 +++++++++++-
> kernel/trace/trace_kprobe.c | 15 +++++++++++++--
> 3 files changed, 25 insertions(+), 4 deletions(-)
>
> diff --git a/arch/Kconfig b/arch/Kconfig
> index 85bb59f7b8c07..cf43de9ffb5b9 100644
> --- a/arch/Kconfig
> +++ b/arch/Kconfig
> @@ -52,7 +52,7 @@ config GENERIC_ENTRY
>
> config KPROBES
> bool "Kprobes"
> - depends on MODULES
> + depends on MODULES || !KPROBES_USE_MODULE_ALLOC
Whoops; that should be:
depends on MODULES || HAVE_KPROBES_ALLOC
.. with similar fixups in the commit message to describe HAVE_KPROBES_ALLOC
rather than KPROBES_USE_MODULE_ALLOC (which does not exist in any version of
the series that got sent to the list).
I'll send a v2 with that fixed (and the other changes from Jarkko's v7 base
patch) once I've locally tested that for architectures with and without
HAVE_KPROBES_ALLOC.
Mark.
Powered by blists - more mailing lists