From: "Philippe Mathieu-Daudé" <philmd@redhat.com>
To: "Alex Bennée" <alex.bennee@linaro.org>, qemu-devel@nongnu.org
Cc: Paolo Bonzini <pbonzini@redhat.com>, Richard Henderson <rth@twiddle.net>
Subject: Re: [PATCH v1 10/10] translate-all: include guest address in out_asm output
Date: Thu, 21 May 2020 16:04:51 +0200 [thread overview]
Message-ID: <f4862b01-8a96-7c8c-c62e-ccb9fa5cffbb@redhat.com> (raw)
In-Reply-To: <20200513175134.19619-11-alex.bennee@linaro.org>
Hi Alex,
On 5/13/20 7:51 PM, Alex Bennée wrote:
> We already have information about where each guest instructions
> representation starts stored in the tcg_ctx->gen_insn_data so we can
> rectify the PC for faults. We can re-use this information to annotate
> the out_asm output with guest instruction address which makes it a bit
> easier to work out where you are especially with longer blocks. A
> minor wrinkle is that some instructions get optimised away so we have
> to scan forward until we find some actual generated code.
>
> Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
>
> ---
> v1
> - better logic for doing chunk at a time
> - use new "note" facility to tag address
> - rewrite the commit log
> v2
> - don't terminate gen_insn_end_off, trust your termination
> conditions ;-)
> ---
> accel/tcg/translate-all.c | 39 +++++++++++++++++++++++++++++++++------
> 1 file changed, 33 insertions(+), 6 deletions(-)
>
> diff --git a/accel/tcg/translate-all.c b/accel/tcg/translate-all.c
> index cdf58bb420e..42ce1dfcff7 100644
> --- a/accel/tcg/translate-all.c
> +++ b/accel/tcg/translate-all.c
> @@ -1794,14 +1794,43 @@ TranslationBlock *tb_gen_code(CPUState *cpu,
> if (qemu_loglevel_mask(CPU_LOG_TB_OUT_ASM) &&
> qemu_log_in_addr_range(tb->pc)) {
> FILE *logfile = qemu_log_lock();
> + int code_size, data_size = 0;
> + g_autoptr(GString) note = g_string_new("[tb header & initial instruction]");
> + size_t chunk_start = 0;
> + int insn = 0;
> qemu_log("OUT: [size=%d]\n", gen_code_size);
> if (tcg_ctx->data_gen_ptr) {
> - size_t code_size = tcg_ctx->data_gen_ptr - tb->tc.ptr;
> - size_t data_size = gen_code_size - code_size;
> - size_t i;
> + code_size = tcg_ctx->data_gen_ptr - tb->tc.ptr;
> + data_size = gen_code_size - code_size;
> + } else {
> + code_size = gen_code_size;
> + }
>
> - log_disas(tb->tc.ptr, code_size, NULL);
> + /* Dump header and the first instruction */
> + chunk_start = tcg_ctx->gen_insn_end_off[insn];
> + log_disas(tb->tc.ptr, chunk_start, note->str);
>
> + /*
> + * Dump each instruction chunk, wrapping up empty chunks into
> + * the next instruction. The whole array is offset so the
> + * first entry is the beginning of the 2nd instruction.
> + */
> + while (insn <= tb->icount && chunk_start < code_size) {
> + size_t chunk_end = tcg_ctx->gen_insn_end_off[insn];
> + if (chunk_end > chunk_start) {
> + g_string_printf(note, "[guest addr: " TARGET_FMT_lx "]",
> + tcg_ctx->gen_insn_data[insn][0]);
> + log_disas(tb->tc.ptr + chunk_start, chunk_end - chunk_start,
> + note->str);
> + chunk_start = chunk_end;
> + }
> + insn++;
> + }
> +
> + /* Finally dump any data we may have after the block */
> + if (data_size) {
It seems we can simplify checking tcg_ctx->data_gen_ptr here, and
declaring data_size in this reduced scope. Doing so as a preliminary
patch makes the rest of this patch easier to review. What do you think?
> + int i;
> + qemu_log(" data: [size=%d]\n", data_size);
> for (i = 0; i < data_size; i += sizeof(tcg_target_ulong)) {
> if (sizeof(tcg_target_ulong) == 8) {
> qemu_log("0x%08" PRIxPTR ": .quad 0x%016" PRIx64 "\n",
> @@ -1813,8 +1842,6 @@ TranslationBlock *tb_gen_code(CPUState *cpu,
> *(uint32_t *)(tcg_ctx->data_gen_ptr + i));
> }
> }
> - } else {
> - log_disas(tb->tc.ptr, gen_code_size, NULL);
> }
> qemu_log("\n");
> qemu_log_flush();
>
prev parent reply other threads:[~2020-05-21 14:05 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-05-13 17:51 [PATCH v1 00/10] testing and tcg tweaks Alex Bennée
2020-05-13 17:51 ` [PATCH v1 01/10] tests/guest-debug: catch hanging guests Alex Bennée
2020-05-21 14:06 ` Philippe Mathieu-Daudé
2020-05-13 17:51 ` [PATCH v1 02/10] travis.yml: Improve the --disable-tcg test on s390x Alex Bennée
2020-05-13 17:51 ` [PATCH v1 03/10] tests/docker: Kludge <linux/swab.h> breakage by pinning linux-libc-dev Alex Bennée
2020-05-13 17:51 ` [PATCH v1 04/10] linux-user: completely re-write init_guest_space Alex Bennée
2020-05-21 4:43 ` Thomas Huth
2020-05-21 8:21 ` Alex Bennée
2020-05-21 15:39 ` Richard Henderson
2020-05-22 10:24 ` Alex Bennée
2020-05-22 10:36 ` Peter Maydell
2020-05-13 17:51 ` [PATCH v1 05/10] exec/cpu-all: Use bool for have_guest_base Alex Bennée
2020-05-13 18:18 ` Laurent Vivier
2020-05-13 17:51 ` [PATCH v1 06/10] accel/tcg: Relax va restrictions on 64-bit guests Alex Bennée
2020-05-13 17:51 ` [PATCH v1 07/10] accel/tcg: don't disable exec_tb trace events Alex Bennée
2020-05-13 17:51 ` [PATCH v1 08/10] disas: include an optional note for the start of disassembly Alex Bennée
2020-05-13 17:51 ` [PATCH v1 09/10] disas: add optional note support to cap_disas Alex Bennée
2020-05-13 17:51 ` [PATCH v1 10/10] translate-all: include guest address in out_asm output Alex Bennée
2020-05-21 14:04 ` Philippe Mathieu-Daudé [this message]
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=f4862b01-8a96-7c8c-c62e-ccb9fa5cffbb@redhat.com \
--to=philmd@redhat.com \
--cc=alex.bennee@linaro.org \
--cc=pbonzini@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=rth@twiddle.net \
/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).