qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Aaron Lindsay via <qemu-devel@nongnu.org>
To: "Alex Bennée" <alex.bennee@linaro.org>
Cc: qemu-devel@nongnu.org, cota@braap.org, kuhn.chenqun@huawei.com,
	robhenry@microsoft.com, mahmoudabdalghany@outlook.com,
	Richard Henderson <richard.henderson@linaro.org>,
	Paolo Bonzini <pbonzini@redhat.com>
Subject: Re: [PATCH  v2 20/21] accel/tcg: allow plugin instrumentation to be disable via cflags
Date: Thu, 11 Feb 2021 19:53:56 -0500	[thread overview]
Message-ID: <YCXRpCiekU+TgoAX@strawberry.localdomain> (raw)
In-Reply-To: <20210210221053.18050-21-alex.bennee@linaro.org>

On Feb 10 22:10, Alex Bennée wrote:
> When icount is enabled and we recompile an MMIO access we end up
> double counting the instruction execution. To avoid this we introduce
> the CF_NOINSTR cflag which disables instrumentation for the next TB.
> As this is part of the hashed compile flags we will only execute the
> generated TB while coming out of a cpu_io_recompile.

Unfortunately this patch works a little too well!

With this change, the memory access callbacks registered via
`qemu_plugin_register_vcpu_mem_cb()` are never called for the
re-translated instruction making the IO access, since we've disabled all
instrumentation.

Is it possible to selectively disable only instruction callbacks using
this mechanism, while still allowing others that would not yet have been
called for the re-translated instruction?

-Aaron

> While we are at it delete the old TODO. We might as well keep the
> translation handy as it's likely you will repeatedly hit it on each
> MMIO access.
> 
> Reported-by: Aaron Lindsay <aaron@os.amperecomputing.com>
> Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
> Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
> Message-Id: <20210209182749.31323-12-alex.bennee@linaro.org>
> 
> ---
> v2
>   - squashed CH_HASHMASK to ~CF_INVALID
> ---
>  include/exec/exec-all.h   |  6 +++---
>  accel/tcg/translate-all.c | 17 ++++++++---------
>  accel/tcg/translator.c    |  2 +-
>  3 files changed, 12 insertions(+), 13 deletions(-)
> 
> diff --git a/include/exec/exec-all.h b/include/exec/exec-all.h
> index e08179de34..299282cc59 100644
> --- a/include/exec/exec-all.h
> +++ b/include/exec/exec-all.h
> @@ -454,14 +454,14 @@ struct TranslationBlock {
>      uint32_t cflags;    /* compile flags */
>  #define CF_COUNT_MASK  0x00007fff
>  #define CF_LAST_IO     0x00008000 /* Last insn may be an IO access.  */
> +#define CF_NOINSTR     0x00010000 /* Disable instrumentation of TB */
>  #define CF_USE_ICOUNT  0x00020000
>  #define CF_INVALID     0x00040000 /* TB is stale. Set with @jmp_lock held */
>  #define CF_PARALLEL    0x00080000 /* Generate code for a parallel context */
>  #define CF_CLUSTER_MASK 0xff000000 /* Top 8 bits are cluster ID */
>  #define CF_CLUSTER_SHIFT 24
> -/* cflags' mask for hashing/comparison */
> -#define CF_HASH_MASK   \
> -    (CF_COUNT_MASK | CF_LAST_IO | CF_USE_ICOUNT | CF_PARALLEL | CF_CLUSTER_MASK)
> +/* cflags' mask for hashing/comparison, basically ignore CF_INVALID */
> +#define CF_HASH_MASK   (~CF_INVALID)
>  
>      /* Per-vCPU dynamic tracing state used to generate this TB */
>      uint32_t trace_vcpu_dstate;
> diff --git a/accel/tcg/translate-all.c b/accel/tcg/translate-all.c
> index 0666f9ef14..32a3d8fe24 100644
> --- a/accel/tcg/translate-all.c
> +++ b/accel/tcg/translate-all.c
> @@ -2399,7 +2399,8 @@ void tb_check_watchpoint(CPUState *cpu, uintptr_t retaddr)
>  }
>  
>  #ifndef CONFIG_USER_ONLY
> -/* in deterministic execution mode, instructions doing device I/Os
> +/*
> + * In deterministic execution mode, instructions doing device I/Os
>   * must be at the end of the TB.
>   *
>   * Called by softmmu_template.h, with iothread mutex not held.
> @@ -2430,19 +2431,17 @@ void cpu_io_recompile(CPUState *cpu, uintptr_t retaddr)
>          n = 2;
>      }
>  
> -    /* Generate a new TB executing the I/O insn.  */
> -    cpu->cflags_next_tb = curr_cflags() | CF_LAST_IO | n;
> +    /*
> +     * Exit the loop and potentially generate a new TB executing the
> +     * just the I/O insns. We also disable instrumentation so we don't
> +     * double count the instruction.
> +     */
> +    cpu->cflags_next_tb = curr_cflags() | CF_NOINSTR | CF_LAST_IO | n;
>  
>      qemu_log_mask_and_addr(CPU_LOG_EXEC, tb->pc,
>                             "cpu_io_recompile: rewound execution of TB to "
>                             TARGET_FMT_lx "\n", tb->pc);
>  
> -    /* TODO: If env->pc != tb->pc (i.e. the faulting instruction was not
> -     * the first in the TB) then we end up generating a whole new TB and
> -     *  repeating the fault, which is horribly inefficient.
> -     *  Better would be to execute just this insn uncached, or generate a
> -     *  second new TB.
> -     */
>      cpu_loop_exit_noexc(cpu);
>  }
>  
> diff --git a/accel/tcg/translator.c b/accel/tcg/translator.c
> index a49a794065..14d1ea795d 100644
> --- a/accel/tcg/translator.c
> +++ b/accel/tcg/translator.c
> @@ -58,7 +58,7 @@ void translator_loop(const TranslatorOps *ops, DisasContextBase *db,
>      ops->tb_start(db, cpu);
>      tcg_debug_assert(db->is_jmp == DISAS_NEXT);  /* no early exit */
>  
> -    plugin_enabled = plugin_gen_tb_start(cpu, tb);
> +    plugin_enabled = !(tb_cflags(db->tb) & CF_NOINSTR) && plugin_gen_tb_start(cpu, tb);
>  
>      while (true) {
>          db->num_insns++;
> -- 
> 2.20.1
> 


  reply	other threads:[~2021-02-12  0:56 UTC|newest]

Thread overview: 53+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-02-10 22:10 [PATCH v2 00/21] plugins/next pre-PR (hwprofile, regression fixes, icount count fix) Alex Bennée
2021-02-10 22:10 ` [PATCH v2 01/21] hw/virtio/pci: include vdev name in registered PCI sections Alex Bennée
2021-02-10 22:10 ` [PATCH v2 02/21] plugins: add API to return a name for a IO device Alex Bennée
2021-02-10 22:10 ` [PATCH v2 03/21] plugins: new hwprofile plugin Alex Bennée
2021-02-10 22:10 ` [PATCH v2 04/21] contrib: Don't use '#' flag of printf format Alex Bennée
2021-02-10 22:10 ` [PATCH v2 05/21] contrib: Fix some code style problems, ERROR: "foo * bar" should be "foo *bar" Alex Bennée
2021-02-10 22:10 ` [PATCH v2 06/21] contrib: Add spaces around operator Alex Bennée
2021-02-10 22:10 ` [PATCH v2 07/21] contrib: space required after that ',' Alex Bennée
2021-02-10 22:10 ` [PATCH v2 08/21] contrib: Open brace '{' following struct go on the same line Alex Bennée
2021-02-10 22:10 ` [PATCH v2 09/21] accel/tcg/plugin-gen: fix the call signature for inline callbacks Alex Bennée
2021-02-10 22:10 ` [PATCH v2 10/21] exec: Move TranslationBlock typedef to qemu/typedefs.h Alex Bennée
2021-02-11 10:14   ` Philippe Mathieu-Daudé
2021-02-11 10:24     ` Alex Bennée
2021-02-10 22:10 ` [PATCH v2 11/21] accel/tcg: Create io_recompile_replay_branch hook Alex Bennée
2021-02-11 10:12   ` Philippe Mathieu-Daudé
2021-02-10 22:10 ` [PATCH v2 12/21] target/mips: Create mips_io_recompile_replay_branch Alex Bennée
2021-02-11 10:10   ` Philippe Mathieu-Daudé
2021-02-10 22:10 ` [PATCH v2 13/21] target/sh4: Create superh_io_recompile_replay_branch Alex Bennée
2021-02-11 10:13   ` Philippe Mathieu-Daudé
2021-02-10 22:10 ` [PATCH v2 14/21] tests/plugin: expand insn test to detect duplicate instructions Alex Bennée
2021-02-10 22:10 ` [PATCH v2 15/21] tests/acceptance: add a new set of tests to exercise plugins Alex Bennée
2021-02-11 10:31   ` Philippe Mathieu-Daudé
2021-02-11 18:59     ` Wainer dos Santos Moschetta
2021-02-11 19:51   ` Wainer dos Santos Moschetta
2021-02-10 22:10 ` [PATCH v2 16/21] accel/tcg: actually cache our partial icount TB Alex Bennée
2021-02-11 10:21   ` Philippe Mathieu-Daudé
2021-02-11 18:48     ` Richard Henderson
2021-02-12 15:40       ` Philippe Mathieu-Daudé
2021-02-12 17:06         ` Alex Bennée
2021-02-11 18:48   ` Richard Henderson
2021-02-10 22:10 ` [PATCH v2 17/21] accel/tcg: cache single instruction TB on pending replay exception Alex Bennée
2021-02-11 19:12   ` Richard Henderson
2021-02-11 20:00     ` Alex Bennée
2021-02-10 22:10 ` [PATCH v2 18/21] accel/tcg: re-factor non-RAM execution code Alex Bennée
2021-02-11 19:19   ` Richard Henderson
2021-02-10 22:10 ` [PATCH v2 19/21] accel/tcg: remove CF_NOCACHE and special cases Alex Bennée
2021-02-10 22:10 ` [PATCH v2 20/21] accel/tcg: allow plugin instrumentation to be disable via cflags Alex Bennée
2021-02-12  0:53   ` Aaron Lindsay via [this message]
2021-02-12 11:22     ` Alex Bennée
2021-02-12 14:31       ` Aaron Lindsay via
2021-02-12 14:59         ` Alex Bennée
2021-02-12 14:43     ` Alex Bennée
2021-02-12 15:41       ` Aaron Lindsay via
2021-02-12 16:04         ` Alex Bennée
2021-02-12 16:50           ` Aaron Lindsay via
2021-02-12 17:19             ` Alex Bennée
2021-02-16 10:34             ` Alex Bennée
2021-02-17 16:32               ` Aaron Lindsay via
2021-02-12 16:00       ` Alex Bennée
2021-02-12 17:04         ` Aaron Lindsay via
2021-02-10 22:10 ` [PATCH v2 21/21] tests/acceptance: add a new tests to detect counting errors Alex Bennée
2021-02-11 10:24   ` Philippe Mathieu-Daudé
2021-02-11 19:56   ` Wainer dos Santos Moschetta

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=YCXRpCiekU+TgoAX@strawberry.localdomain \
    --to=qemu-devel@nongnu.org \
    --cc=aaron@os.amperecomputing.com \
    --cc=alex.bennee@linaro.org \
    --cc=cota@braap.org \
    --cc=kuhn.chenqun@huawei.com \
    --cc=mahmoudabdalghany@outlook.com \
    --cc=pbonzini@redhat.com \
    --cc=richard.henderson@linaro.org \
    --cc=robhenry@microsoft.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).