From: Miroslav Benes <mbenes@suse.cz>
To: Torsten Duwe <duwe@lst.de>
Cc: Will Deacon <will.deacon@arm.com>,
Catalin Marinas <catalin.marinas@arm.com>,
Julien Thierry <julien.thierry@arm.com>,
Steven Rostedt <rostedt@goodmis.org>,
Josh Poimboeuf <jpoimboe@redhat.com>,
Ingo Molnar <mingo@redhat.com>,
Ard Biesheuvel <ard.biesheuvel@linaro.org>,
Arnd Bergmann <arnd@arndb.de>,
AKASHI Takahiro <takahiro.akashi@linaro.org>,
linux-arm-kernel@lists.infradead.org,
linux-kernel@vger.kernel.org, live-patching@vger.kernel.org,
jeyu@kernel.org
Subject: Re: [PATCH v3 3/4] arm64: implement live patching
Date: Wed, 17 Oct 2018 15:39:31 +0200 (CEST) [thread overview]
Message-ID: <alpine.LSU.2.21.1810171051560.21847@pobox.suse.cz> (raw)
In-Reply-To: <20181001141652.5478C68BE1@newverein.lst.de>
On Mon, 1 Oct 2018, Torsten Duwe wrote:
> Based on ftrace with regs, do the usual thing. Also allocate a
> task flag for whatever consistency handling will be used.
> Watch out for interactions with the graph tracer.
Similar to what Mark wrote about 2/4, I'd appreciate a better commit log.
Could you explain traditional "what/why/how", please?
> Signed-off-by: Torsten Duwe <duwe@suse.de>
>
> --- a/arch/arm64/Kconfig
> +++ b/arch/arm64/Kconfig
> @@ -119,6 +119,7 @@ config ARM64
> select HAVE_GENERIC_DMA_COHERENT
> select HAVE_HW_BREAKPOINT if PERF_EVENTS
> select HAVE_IRQ_TIME_ACCOUNTING
> + select HAVE_LIVEPATCH
> select HAVE_MEMBLOCK
> select HAVE_MEMBLOCK_NODE_MAP if NUMA
> select HAVE_NMI
> @@ -1349,4 +1350,6 @@ if CRYPTO
> source "arch/arm64/crypto/Kconfig"
> endif
>
> +source "kernel/livepatch/Kconfig"
> +
> source "lib/Kconfig"
> --- a/arch/arm64/include/asm/thread_info.h
> +++ b/arch/arm64/include/asm/thread_info.h
> @@ -76,6 +76,7 @@ void arch_release_task_struct(struct tas
> #define TIF_FOREIGN_FPSTATE 3 /* CPU's FP state is not current's */
> #define TIF_UPROBE 4 /* uprobe breakpoint or singlestep */
> #define TIF_FSCHECK 5 /* Check FS is USER_DS on return */
> +#define TIF_PATCH_PENDING 6
> #define TIF_NOHZ 7
> #define TIF_SYSCALL_TRACE 8
> #define TIF_SYSCALL_AUDIT 9
> @@ -94,6 +95,7 @@ void arch_release_task_struct(struct tas
> #define _TIF_NEED_RESCHED (1 << TIF_NEED_RESCHED)
> #define _TIF_NOTIFY_RESUME (1 << TIF_NOTIFY_RESUME)
> #define _TIF_FOREIGN_FPSTATE (1 << TIF_FOREIGN_FPSTATE)
> +#define _TIF_PATCH_PENDING (1 << TIF_PATCH_PENDING)
> #define _TIF_NOHZ (1 << TIF_NOHZ)
> #define _TIF_SYSCALL_TRACE (1 << TIF_SYSCALL_TRACE)
> #define _TIF_SYSCALL_AUDIT (1 << TIF_SYSCALL_AUDIT)
> @@ -106,7 +108,8 @@ void arch_release_task_struct(struct tas
>
> #define _TIF_WORK_MASK (_TIF_NEED_RESCHED | _TIF_SIGPENDING | \
> _TIF_NOTIFY_RESUME | _TIF_FOREIGN_FPSTATE | \
> - _TIF_UPROBE | _TIF_FSCHECK)
> + _TIF_UPROBE | _TIF_FSCHECK | \
> + _TIF_PATCH_PENDING)
Could you add a note to the changelog what this means? My ability to read
arm64 entry.S is very limited, but I can see that _TIF_WORK_MASK is
process in a syscall exit and irq return paths. That's good. It is also
called (via "b ret_to_user") in a couple of different places (el0_*
labels). I guess those are returns from exception handling. A comment
about it in the changelog would be appreciated.
> #define _TIF_SYSCALL_WORK (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_AUDIT | \
> _TIF_SYSCALL_TRACEPOINT | _TIF_SECCOMP | \
> --- /dev/null
> +++ b/arch/arm64/include/asm/livepatch.h
> @@ -0,0 +1,36 @@
> +/* SPDX-License-Identifier: GPL-2.0
> + *
> + * livepatch.h - arm64-specific Kernel Live Patching Core
> + *
> + * Copyright (C) 2016,2018 SUSE
> + *
> + * This program is free software; you can redistribute it and/or
> + * modify it under the terms of the GNU General Public License
> + * as published by the Free Software Foundation; either version 2
> + * of the License, or (at your option) any later version.
> + *
> + * This program is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> + * GNU General Public License for more details.
> + *
> + * You should have received a copy of the GNU General Public License
> + * along with this program; if not, see <http://www.gnu.org/licenses/>.
> + */
> +#ifndef _ASM_ARM64_LIVEPATCH_H
> +#define _ASM_ARM64_LIVEPATCH_H
> +
> +#include <linux/module.h>
> +#include <linux/ftrace.h>
I think that only
#include <asm/ptrace.h>
is in fact needed, because of "struct pt_regs".
Ad relocations. I checked that everything in struct mod_arch_specific
stays after the module is load. Both core and init get SHF_ALLOC set
(mod->arch.core.plt->sh_flags in module_frob_arch_sections(). It is
important because apply_relocate_add() may use those sections
through module_emit_plt_entry() call.
ftrace_trampoline section gets SHF_ALLOC as well. Btw, don't you want to
do something similar for new ftrace_regs_trampoline in
module_frob_arch_sections()? Perhaps even edit
arch/arm64/kernel/module.lds? I'm completely clueless here, so it may be
ok. And it applies to 2/4 patch.
The last thing is count_plts() function called from
module_frob_arch_sections(). It needed to be changed in 2016 as well. See
Jessica's patch
(20160713001113.GA30925@packer-debian-8-amd64.digitalocean.com). The logic
in the function has changed since then. If I am not mistaken,
count_plts() is fine as it is right now. It does not consider SHN_UNDEF
anymore, it looks at the destination section (where a symbol should
resolved to) only.
Jessica, could you doublecheck, please?
Torsten, please doublecheck all this as well and add a comment about all
this to the changelog, so we don't have to analyze it from the beginning
again in the future.
Thanks,
Miroslav
next prev parent reply other threads:[~2018-10-17 13:39 UTC|newest]
Thread overview: 48+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-10-01 14:09 [PATCH v3 0/4] arm64 live patching Torsten Duwe
2018-10-01 14:16 ` [PATCH v3 1/4] DYNAMIC_FTRACE configurable with and without REGS Torsten Duwe
2018-10-01 14:52 ` Ard Biesheuvel
2018-10-01 15:03 ` Torsten Duwe
2018-10-01 15:06 ` Ard Biesheuvel
2018-10-01 15:10 ` Torsten Duwe
2018-10-01 15:14 ` Steven Rostedt
2018-10-01 14:16 ` [PATCH v3 2/4] arm64: implement ftrace with regs Torsten Duwe
2018-10-01 15:57 ` Ard Biesheuvel
2018-10-02 10:02 ` Torsten Duwe
2018-10-02 10:39 ` Ard Biesheuvel
2018-10-02 11:27 ` Mark Rutland
2018-10-02 12:18 ` Torsten Duwe
2018-10-02 12:57 ` Mark Rutland
2018-10-01 14:16 ` [PATCH v3 3/4] arm64: implement live patching Torsten Duwe
2018-10-17 13:39 ` Miroslav Benes [this message]
2018-10-18 12:58 ` Jessica Yu
2018-10-19 11:59 ` Miroslav Benes
2018-10-19 12:18 ` Jessica Yu
2018-10-19 15:14 ` Miroslav Benes
2018-10-19 13:46 ` Torsten Duwe
2018-10-19 13:52 ` Ard Biesheuvel
2018-10-19 15:21 ` Miroslav Benes
2018-10-20 14:10 ` Ard Biesheuvel
2018-10-22 12:53 ` Miroslav Benes
2018-10-22 14:54 ` Torsten Duwe
2018-10-23 17:55 ` [PATCH] arm64/module: use mod->klp_info section header information Jessica Yu
2018-10-23 19:32 ` kbuild test robot
2018-10-24 11:57 ` Miroslav Benes
2018-10-25 8:08 ` Petr Mladek
2018-10-25 9:00 ` Miroslav Benes
2018-10-25 11:42 ` Jessica Yu
2018-10-26 17:25 ` [PATCH v2] arm64/module: use mod->klp_info section header information for livepatch modules Jessica Yu
2018-10-29 13:24 ` Miroslav Benes
2018-10-29 13:32 ` Jessica Yu
2018-10-29 15:28 ` Will Deacon
2018-10-30 13:19 ` Jessica Yu
2018-11-01 15:18 ` Miroslav Benes
2018-11-01 16:07 ` Will Deacon
2018-11-05 12:30 ` Ard Biesheuvel
2018-11-05 17:57 ` [PATCH] arm64/module: use plt section indices for relocations Jessica Yu
2018-11-05 18:04 ` Ard Biesheuvel
2018-11-05 18:53 ` [PATCH v2] " Jessica Yu
2018-11-05 18:56 ` Ard Biesheuvel
2018-11-05 19:26 ` Will Deacon
2018-11-05 19:49 ` Jessica Yu
2018-11-06 9:44 ` Miroslav Benes
2018-10-01 14:16 ` [PATCH v3 4/4] arm64: reliable stacktraces Torsten Duwe
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=alpine.LSU.2.21.1810171051560.21847@pobox.suse.cz \
--to=mbenes@suse.cz \
--cc=ard.biesheuvel@linaro.org \
--cc=arnd@arndb.de \
--cc=catalin.marinas@arm.com \
--cc=duwe@lst.de \
--cc=jeyu@kernel.org \
--cc=jpoimboe@redhat.com \
--cc=julien.thierry@arm.com \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=live-patching@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=rostedt@goodmis.org \
--cc=takahiro.akashi@linaro.org \
--cc=will.deacon@arm.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 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).