qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: BALATON Zoltan <balaton@eik.bme.hu>
To: Richard Henderson <richard.henderson@linaro.org>
Cc: lvivier@redhat.com, alistair23@gmail.com, qemu-devel@nongnu.org
Subject: Re: [PATCH] tcg: Restart code generation when we run out of temps
Date: Sun, 24 Jan 2021 14:48:45 +0100 (CET)	[thread overview]
Message-ID: <9f216d18-7d62-1e5-795e-3d2a2ae2ac4f@eik.bme.hu> (raw)
In-Reply-To: <20210123230105.2076270-1-richard.henderson@linaro.org>

On Sat, 23 Jan 2021, Richard Henderson wrote:
> Some large translation blocks can generate so many unique
> constants that we run out of temps to hold them.  In this
> case, longjmp back to the start of code generation and
> restart with a smaller translation block.
>
> Buglink: https://bugs.launchpad.net/bugs/1912065
> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Tested-by: BALATON Zoltan <balaton@eik.bme.hu>

Regards,
BALATON Zoltan

> ---
>
> This replaces both the patch to increase the number of temps,
> and the buggy patch set that dynamically allocated the temps.
>
>
> r~
>
> ---
> include/tcg/tcg.h         |  3 +++
> accel/tcg/translate-all.c | 15 ++++++++++++++-
> tcg/tcg.c                 | 11 ++++++++---
> 3 files changed, 25 insertions(+), 4 deletions(-)
>
> diff --git a/include/tcg/tcg.h b/include/tcg/tcg.h
> index c5a9d65d5f..0f0695e90d 100644
> --- a/include/tcg/tcg.h
> +++ b/include/tcg/tcg.h
> @@ -680,6 +680,9 @@ struct TCGContext {
>
>     uint16_t gen_insn_end_off[TCG_MAX_INSNS];
>     target_ulong gen_insn_data[TCG_MAX_INSNS][TARGET_INSN_START_WORDS];
> +
> +    /* Exit to translator on overflow. */
> +    sigjmp_buf jmp_trans;
> };
>
> static inline bool temp_readonly(TCGTemp *ts)
> diff --git a/accel/tcg/translate-all.c b/accel/tcg/translate-all.c
> index d09c187e0f..81d4c83f22 100644
> --- a/accel/tcg/translate-all.c
> +++ b/accel/tcg/translate-all.c
> @@ -1926,11 +1926,17 @@ TranslationBlock *tb_gen_code(CPUState *cpu,
>     ti = profile_getclock();
> #endif
>
> +    gen_code_size = sigsetjmp(tcg_ctx->jmp_trans, 0);
> +    if (unlikely(gen_code_size != 0)) {
> +        goto error_return;
> +    }
> +
>     tcg_func_start(tcg_ctx);
>
>     tcg_ctx->cpu = env_cpu(env);
>     gen_intermediate_code(cpu, tb, max_insns);
>     tcg_ctx->cpu = NULL;
> +    max_insns = tb->icount;
>
>     trace_translate_block(tb, tb->pc, tb->tc.ptr);
>
> @@ -1955,6 +1961,7 @@ TranslationBlock *tb_gen_code(CPUState *cpu,
>
>     gen_code_size = tcg_gen_code(tcg_ctx, tb);
>     if (unlikely(gen_code_size < 0)) {
> + error_return:
>         switch (gen_code_size) {
>         case -1:
>             /*
> @@ -1966,6 +1973,9 @@ TranslationBlock *tb_gen_code(CPUState *cpu,
>              * flush the TBs, allocate a new TB, re-initialize it per
>              * above, and re-do the actual code generation.
>              */
> +            qemu_log_mask(CPU_LOG_TB_OP | CPU_LOG_TB_OP_OPT,
> +                          "Restarting code generation for "
> +                          "code_gen_buffer overflow\n");
>             goto buffer_overflow;
>
>         case -2:
> @@ -1978,9 +1988,12 @@ TranslationBlock *tb_gen_code(CPUState *cpu,
>              * Try again with half as many insns as we attempted this time.
>              * If a single insn overflows, there's a bug somewhere...
>              */
> -            max_insns = tb->icount;
>             assert(max_insns > 1);
>             max_insns /= 2;
> +            qemu_log_mask(CPU_LOG_TB_OP | CPU_LOG_TB_OP_OPT,
> +                          "Restarting code generation with "
> +                          "smaller translation block (max %d insns)\n",
> +                          max_insns);
>             goto tb_overflow;
>
>         default:
> diff --git a/tcg/tcg.c b/tcg/tcg.c
> index 67b08f708d..9e1b0d73c7 100644
> --- a/tcg/tcg.c
> +++ b/tcg/tcg.c
> @@ -1205,18 +1205,23 @@ void tcg_func_start(TCGContext *s)
>     QSIMPLEQ_INIT(&s->labels);
> }
>
> -static inline TCGTemp *tcg_temp_alloc(TCGContext *s)
> +static TCGTemp *tcg_temp_alloc(TCGContext *s)
> {
>     int n = s->nb_temps++;
> -    tcg_debug_assert(n < TCG_MAX_TEMPS);
> +
> +    if (n >= TCG_MAX_TEMPS) {
> +        /* Signal overflow, starting over with fewer guest insns. */
> +        siglongjmp(s->jmp_trans, -2);
> +    }
>     return memset(&s->temps[n], 0, sizeof(TCGTemp));
> }
>
> -static inline TCGTemp *tcg_global_alloc(TCGContext *s)
> +static TCGTemp *tcg_global_alloc(TCGContext *s)
> {
>     TCGTemp *ts;
>
>     tcg_debug_assert(s->nb_globals == s->nb_temps);
> +    tcg_debug_assert(s->nb_globals < TCG_MAX_TEMPS);
>     s->nb_globals++;
>     ts = tcg_temp_alloc(s);
>     ts->kind = TEMP_GLOBAL;
>


      parent reply	other threads:[~2021-01-24 13:50 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-01-23 23:01 [PATCH] tcg: Restart code generation when we run out of temps Richard Henderson
2021-01-24  0:27 ` Philippe Mathieu-Daudé
2021-01-24 13:48 ` BALATON Zoltan [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=9f216d18-7d62-1e5-795e-3d2a2ae2ac4f@eik.bme.hu \
    --to=balaton@eik.bme.hu \
    --cc=alistair23@gmail.com \
    --cc=lvivier@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=richard.henderson@linaro.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).