From: Borislav Petkov <bp@alien8.de>
To: Andy Lutomirski <luto@amacapital.net>,
Masami Hiramatsu <mhiramat@kernel.org>
Cc: X86 ML <x86@kernel.org>, LKML <linux-kernel@vger.kernel.org>
Subject: [RFC PATCH v0 00/19] x86/insn: Add an insn_decode() API
Date: Tue, 24 Nov 2020 11:19:33 +0100 [thread overview]
Message-ID: <20201124101952.7909-1-bp@alien8.de> (raw)
From: Borislav Petkov <bp@suse.de>
Hi,
here's what I had in mind, finally split into proper patches. The final
goal is for all users of the decoder to simply call insn_decode() and
look at its retval. Simple.
As to amluto's question about detecting partial insns, see the diff
below.
Running that gives:
insn buffer:
0x48 0xcf 0x48 0x83 0x90 0x90 0x90 0x90 0x90 0x90 0x90 0x90 0x90 0x90 0x90
supplied buf size: 15, ret 0
supplied buf size: 2, ret 0
supplied buf size: 3, ret 0
supplied buf size: 4, ret 0
and the return value is always success.
Which means that that buf_len that gets supplied to the decoder
functions doesn't really work and I need to look into it.
That is, provided this is how we want to control what the instruction
decoder decodes - by supplying the length of the buffer it should look
at.
We could also say that probably there should be a way to say "decode
only the first insn in the buffer and ignore the rest". That is all up
to the use cases so I'm looking for suggestions here.
In any case, at least the case where I give it
0x48 0xcf 0x48 0x83
and say that buf size is 4, should return an error because the second
insn is incomplete. So I need to go look at that now.
Thx.
---
diff --git a/arch/x86/tools/insn_sanity.c b/arch/x86/tools/insn_sanity.c
index 51309df285b4..41e1ae0cd833 100644
--- a/arch/x86/tools/insn_sanity.c
+++ b/arch/x86/tools/insn_sanity.c
@@ -220,6 +220,45 @@ static void parse_args(int argc, char **argv)
}
}
+static void run_insn_limits_test(void)
+{
+ unsigned char b[MAX_INSN_SIZE];
+ struct insn insn;
+ int ret, i, size;
+
+ memset(b, INSN_NOP, MAX_INSN_SIZE);
+
+ /* IRETQ */
+ b[0] = 0x48;
+ b[1] = 0xcf;
+
+ /* partial add $0x8, %rsp */
+ b[2] = 0x48;
+ b[3] = 0x83;
+
+ printf("insn buffer:\n");
+
+ for (i = 0; i < MAX_INSN_SIZE; i++)
+ printf("0x%hhx ", b[i]);
+ printf("\n");
+
+ size = MAX_INSN_SIZE;
+ ret = insn_decode(&insn, b, size, INSN_MODE_64);
+ printf("supplied buf size: %d, ret %d\n", size, ret);
+
+ size = 2;
+ ret = insn_decode(&insn, b, size, INSN_MODE_64);
+ printf("supplied buf size: %d, ret %d\n", size, ret);
+
+ size = 3;
+ ret = insn_decode(&insn, b, size, INSN_MODE_64);
+ printf("supplied buf size: %d, ret %d\n", size, ret);
+
+ size = 4;
+ ret = insn_decode(&insn, b, size, INSN_MODE_64);
+ printf("supplied buf size: %d, ret %d\n", size, ret);
+}
+
int main(int argc, char **argv)
{
int insns = 0, ret;
@@ -265,5 +304,7 @@ int main(int argc, char **argv)
errors,
seed);
+ run_insn_limits_test();
+
return errors ? 1 : 0;
}
Borislav Petkov (19):
x86/insn: Rename insn_decode() to insn_decode_regs()
x86/insn: Add @buf_len param to insn_init() kernel-doc comment
x86/insn: Add an insn_decode() API
x86/insn-eval: Handle return values from the decoder
x86/boot/compressed/sev-es: Convert to insn_decode()
perf/x86/intel/ds: Check insn_get_length() retval
perf/x86/intel/ds: Check return values of insn decoder functions
x86/alternative: Use insn_decode()
x86/mce: Convert to insn_decode()
x86/kprobes: Convert to insn_decode()
x86/sev-es: Convert to insn_decode()
x86/traps: Convert to insn_decode()
x86/uprobes: Convert to insn_decode()
x86/tools/insn_decoder_test: Convert to insn_decode()
tools/objtool: Convert to insn_decode()
x86/tools/insn_sanity: Convert to insn_decode()
tools/perf: Convert to insn_decode()
x86/insn: Remove kernel_insn_init()
x86/insn: Make insn_complete() static
arch/x86/boot/compressed/sev-es.c | 11 +-
arch/x86/events/intel/ds.c | 4 +-
arch/x86/events/intel/lbr.c | 10 +-
arch/x86/include/asm/insn-eval.h | 4 +-
arch/x86/include/asm/insn.h | 42 ++--
arch/x86/kernel/alternative.c | 6 +-
arch/x86/kernel/cpu/mce/severity.c | 12 +-
arch/x86/kernel/kprobes/core.c | 17 +-
arch/x86/kernel/kprobes/opt.c | 9 +-
arch/x86/kernel/sev-es.c | 15 +-
arch/x86/kernel/traps.c | 7 +-
arch/x86/kernel/umip.c | 2 +-
arch/x86/kernel/uprobes.c | 8 +-
arch/x86/lib/insn-eval.c | 20 +-
arch/x86/lib/insn.c | 190 ++++++++++++++----
arch/x86/tools/insn_decoder_test.c | 10 +-
arch/x86/tools/insn_sanity.c | 8 +-
tools/arch/x86/include/asm/insn.h | 42 ++--
tools/arch/x86/lib/insn.c | 190 ++++++++++++++----
tools/include/linux/kconfig.h | 73 +++++++
tools/objtool/arch/x86/decode.c | 9 +-
tools/perf/arch/x86/tests/insn-x86.c | 9 +-
tools/perf/arch/x86/util/archinsn.c | 9 +-
.../intel-pt-decoder/intel-pt-insn-decoder.c | 17 +-
24 files changed, 507 insertions(+), 217 deletions(-)
create mode 100644 tools/include/linux/kconfig.h
--
2.21.0
next reply other threads:[~2020-11-24 10:20 UTC|newest]
Thread overview: 34+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-11-24 10:19 Borislav Petkov [this message]
2020-11-24 10:19 ` [RFC PATCH v0 01/19] x86/insn: Rename insn_decode() to insn_decode_regs() Borislav Petkov
2020-11-24 10:19 ` [RFC PATCH v0 02/19] x86/insn: Add @buf_len param to insn_init() kernel-doc comment Borislav Petkov
2020-11-24 10:19 ` [RFC PATCH v0 03/19] x86/insn: Add an insn_decode() API Borislav Petkov
2020-11-25 16:53 ` Masami Hiramatsu
2020-11-25 19:25 ` Borislav Petkov
2020-11-26 1:37 ` Masami Hiramatsu
2020-11-26 17:50 ` Borislav Petkov
2020-11-27 5:54 ` Masami Hiramatsu
2020-11-24 10:19 ` [RFC PATCH v0 04/19] x86/insn-eval: Handle return values from the decoder Borislav Petkov
2020-11-24 10:19 ` [RFC PATCH v0 05/19] x86/boot/compressed/sev-es: Convert to insn_decode() Borislav Petkov
2020-11-24 10:19 ` [RFC PATCH v0 06/19] perf/x86/intel/ds: Check insn_get_length() retval Borislav Petkov
2020-11-24 10:19 ` [RFC PATCH v0 07/19] perf/x86/intel/ds: Check return values of insn decoder functions Borislav Petkov
2020-11-24 10:19 ` [RFC PATCH v0 08/19] x86/alternative: Use insn_decode() Borislav Petkov
2020-11-24 10:19 ` [RFC PATCH v0 09/19] x86/mce: Convert to insn_decode() Borislav Petkov
2020-11-24 10:19 ` [RFC PATCH v0 10/19] x86/kprobes: " Borislav Petkov
2020-11-25 17:19 ` Masami Hiramatsu
2020-11-24 10:19 ` [RFC PATCH v0 11/19] x86/sev-es: " Borislav Petkov
2020-11-24 10:19 ` [RFC PATCH v0 12/19] x86/traps: " Borislav Petkov
2020-11-24 10:19 ` [RFC PATCH v0 13/19] x86/uprobes: " Borislav Petkov
2020-11-24 10:19 ` [RFC PATCH v0 14/19] x86/tools/insn_decoder_test: " Borislav Petkov
2020-11-24 10:19 ` [RFC PATCH v0 15/19] tools/objtool: " Borislav Petkov
2020-11-24 10:19 ` [RFC PATCH v0 16/19] x86/tools/insn_sanity: " Borislav Petkov
2020-11-24 10:19 ` [RFC PATCH v0 17/19] tools/perf: " Borislav Petkov
2020-11-24 10:19 ` [RFC PATCH v0 18/19] x86/insn: Remove kernel_insn_init() Borislav Petkov
2020-11-24 10:19 ` [RFC PATCH v0 19/19] x86/insn: Make insn_complete() static Borislav Petkov
2020-11-24 17:46 ` [RFC PATCH v0 00/19] x86/insn: Add an insn_decode() API Borislav Petkov
2020-11-25 8:03 ` Masami Hiramatsu
2020-11-27 17:45 ` Andy Lutomirski
2020-11-27 22:35 ` Borislav Petkov
2020-11-29 8:50 ` Masami Hiramatsu
2020-11-30 13:44 ` Borislav Petkov
2020-11-30 17:21 ` Masami Hiramatsu
2020-12-02 18:04 ` Borislav Petkov
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=20201124101952.7909-1-bp@alien8.de \
--to=bp@alien8.de \
--cc=linux-kernel@vger.kernel.org \
--cc=luto@amacapital.net \
--cc=mhiramat@kernel.org \
--cc=x86@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).