All of lore.kernel.org
 help / color / mirror / Atom feed
From: Richard Henderson <richard.henderson@linaro.org>
To: qemu-devel@nongnu.org
Cc: Luis Pires <luis.pires@eldorado.org.br>
Subject: [PULL 34/34] docs/devel: Explain in more detail the TB chaining mechanisms
Date: Fri, 11 Jun 2021 16:41:44 -0700	[thread overview]
Message-ID: <20210611234144.653682-35-richard.henderson@linaro.org> (raw)
In-Reply-To: <20210611234144.653682-1-richard.henderson@linaro.org>

From: Luis Pires <luis.pires@eldorado.org.br>

Signed-off-by: Luis Pires <luis.pires@eldorado.org.br>
Message-Id: <20210601125143.191165-1-luis.pires@eldorado.org.br>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
 docs/devel/tcg.rst | 101 ++++++++++++++++++++++++++++++++++++++++-----
 1 file changed, 90 insertions(+), 11 deletions(-)

diff --git a/docs/devel/tcg.rst b/docs/devel/tcg.rst
index 4ebde44b9d..a65fb7b1c4 100644
--- a/docs/devel/tcg.rst
+++ b/docs/devel/tcg.rst
@@ -11,13 +11,14 @@ performances.
 QEMU's dynamic translation backend is called TCG, for "Tiny Code
 Generator". For more information, please take a look at ``tcg/README``.
 
-Some notable features of QEMU's dynamic translator are:
+The following sections outline some notable features and implementation
+details of QEMU's dynamic translator.
 
 CPU state optimisations
 -----------------------
 
-The target CPUs have many internal states which change the way it
-evaluates instructions. In order to achieve a good speed, the
+The target CPUs have many internal states which change the way they
+evaluate instructions. In order to achieve a good speed, the
 translation phase considers that some state information of the virtual
 CPU cannot change in it. The state is recorded in the Translation
 Block (TB). If the state changes (e.g. privilege level), a new TB will
@@ -31,17 +32,95 @@ Direct block chaining
 ---------------------
 
 After each translated basic block is executed, QEMU uses the simulated
-Program Counter (PC) and other cpu state information (such as the CS
+Program Counter (PC) and other CPU state information (such as the CS
 segment base value) to find the next basic block.
 
-In order to accelerate the most common cases where the new simulated PC
-is known, QEMU can patch a basic block so that it jumps directly to the
-next one.
+In its simplest, less optimized form, this is done by exiting from the
+current TB, going through the TB epilogue, and then back to the
+main loop. That’s where QEMU looks for the next TB to execute,
+translating it from the guest architecture if it isn’t already available
+in memory. Then QEMU proceeds to execute this next TB, starting at the
+prologue and then moving on to the translated instructions.
 
-The most portable code uses an indirect jump. An indirect jump makes
-it easier to make the jump target modification atomic. On some host
-architectures (such as x86 or PowerPC), the ``JUMP`` opcode is
-directly patched so that the block chaining has no overhead.
+Exiting from the TB this way will cause the ``cpu_exec_interrupt()``
+callback to be re-evaluated before executing additional instructions.
+It is mandatory to exit this way after any CPU state changes that may
+unmask interrupts.
+
+In order to accelerate the cases where the TB for the new
+simulated PC is already available, QEMU has mechanisms that allow
+multiple TBs to be chained directly, without having to go back to the
+main loop as described above. These mechanisms are:
+
+``lookup_and_goto_ptr``
+^^^^^^^^^^^^^^^^^^^^^^^
+
+Calling ``tcg_gen_lookup_and_goto_ptr()`` will emit a call to
+``helper_lookup_tb_ptr``. This helper will look for an existing TB that
+matches the current CPU state. If the destination TB is available its
+code address is returned, otherwise the address of the JIT epilogue is
+returned. The call to the helper is always followed by the tcg ``goto_ptr``
+opcode, which branches to the returned address. In this way, we either
+branch to the next TB or return to the main loop.
+
+``goto_tb + exit_tb``
+^^^^^^^^^^^^^^^^^^^^^
+
+The translation code usually implements branching by performing the
+following steps:
+
+1. Call ``tcg_gen_goto_tb()`` passing a jump slot index (either 0 or 1)
+   as a parameter.
+
+2. Emit TCG instructions to update the CPU state with any information
+   that has been assumed constant and is required by the main loop to
+   correctly locate and execute the next TB. For most guests, this is
+   just the PC of the branch destination, but others may store additional
+   data. The information updated in this step must be inferable from both
+   ``cpu_get_tb_cpu_state()`` and ``cpu_restore_state()``.
+
+3. Call ``tcg_gen_exit_tb()`` passing the address of the current TB and
+   the jump slot index again.
+
+Step 1, ``tcg_gen_goto_tb()``, will emit a ``goto_tb`` TCG
+instruction that later on gets translated to a jump to an address
+associated with the specified jump slot. Initially, this is the address
+of step 2's instructions, which update the CPU state information. Step 3,
+``tcg_gen_exit_tb()``, exits from the current TB returning a tagged
+pointer composed of the last executed TB’s address and the jump slot
+index.
+
+The first time this whole sequence is executed, step 1 simply jumps
+to step 2. Then the CPU state information gets updated and we exit from
+the current TB. As a result, the behavior is very similar to the less
+optimized form described earlier in this section.
+
+Next, the main loop looks for the next TB to execute using the
+current CPU state information (creating the TB if it wasn’t already
+available) and, before starting to execute the new TB’s instructions,
+patches the previously executed TB by associating one of its jump
+slots (the one specified in the call to ``tcg_gen_exit_tb()``) with the
+address of the new TB.
+
+The next time this previous TB is executed and we get to that same
+``goto_tb`` step, it will already be patched (assuming the destination TB
+is still in memory) and will jump directly to the first instruction of
+the destination TB, without going back to the main loop.
+
+For the ``goto_tb + exit_tb`` mechanism to be used, the following
+conditions need to be satisfied:
+
+* The change in CPU state must be constant, e.g., a direct branch and
+  not an indirect branch.
+
+* The direct branch cannot cross a page boundary. Memory mappings
+  may change, causing the code at the destination address to change.
+
+Note that, on step 3 (``tcg_gen_exit_tb()``), in addition to the
+jump slot index, the address of the TB just executed is also returned.
+This address corresponds to the TB that will be patched; it may be
+different than the one that was directly executed from the main loop
+if the latter had already been chained to other TBs.
 
 Self-modifying code and translated code invalidation
 ----------------------------------------------------
-- 
2.25.1



  parent reply	other threads:[~2021-06-12  0:16 UTC|newest]

Thread overview: 42+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-06-11 23:41 [PULL 00/34] tcg patch queue Richard Henderson
2021-06-11 23:41 ` [PULL 01/34] meson: Split out tcg/meson.build Richard Henderson
2021-06-11 23:41 ` [PULL 02/34] meson: Split out fpu/meson.build Richard Henderson
2021-06-11 23:41 ` [PULL 03/34] tcg: Re-order tcg_region_init vs tcg_prologue_init Richard Henderson
2021-06-14 14:36   ` Christian Borntraeger
2021-06-11 23:41 ` [PULL 04/34] tcg: Remove error return from tcg_region_initial_alloc__locked Richard Henderson
2021-06-11 23:41 ` [PULL 05/34] tcg: Split out tcg_region_initial_alloc Richard Henderson
2021-06-11 23:41 ` [PULL 06/34] tcg: Split out tcg_region_prologue_set Richard Henderson
2021-06-11 23:41 ` [PULL 07/34] tcg: Split out region.c Richard Henderson
2021-06-11 23:41 ` [PULL 08/34] accel/tcg: Inline cpu_gen_init Richard Henderson
2021-06-11 23:41 ` [PULL 09/34] accel/tcg: Move alloc_code_gen_buffer to tcg/region.c Richard Henderson
2021-06-11 23:41 ` [PULL 10/34] accel/tcg: Rename tcg_init to tcg_init_machine Richard Henderson
2021-06-11 23:41 ` [PULL 11/34] tcg: Create tcg_init Richard Henderson
2021-06-11 23:41 ` [PULL 12/34] accel/tcg: Merge tcg_exec_init into tcg_init_machine Richard Henderson
2021-06-11 23:41 ` [PULL 13/34] accel/tcg: Use MiB in tcg_init_machine Richard Henderson
2021-06-11 23:41 ` [PULL 14/34] accel/tcg: Pass down max_cpus to tcg_init Richard Henderson
2021-06-11 23:41 ` [PULL 15/34] tcg: Introduce tcg_max_ctxs Richard Henderson
2021-06-11 23:41 ` [PULL 16/34] tcg: Move MAX_CODE_GEN_BUFFER_SIZE to tcg-target.h Richard Henderson
2021-06-11 23:41 ` [PULL 17/34] tcg: Replace region.end with region.total_size Richard Henderson
2021-06-11 23:41 ` [PULL 18/34] tcg: Rename region.start to region.after_prologue Richard Henderson
2021-06-11 23:41 ` [PULL 19/34] tcg: Tidy tcg_n_regions Richard Henderson
2021-06-11 23:41 ` [PULL 20/34] tcg: Tidy split_cross_256mb Richard Henderson
2021-06-11 23:41 ` [PULL 21/34] tcg: Move in_code_gen_buffer and tests to region.c Richard Henderson
2021-06-11 23:41 ` [PULL 22/34] tcg: Allocate code_gen_buffer into struct tcg_region_state Richard Henderson
2021-06-11 23:41 ` [PULL 23/34] tcg: Return the map protection from alloc_code_gen_buffer Richard Henderson
2021-06-11 23:41 ` [PULL 24/34] tcg: Sink qemu_madvise call to common code Richard Henderson
2021-06-11 23:41 ` [PULL 25/34] util/osdep: Add qemu_mprotect_rw Richard Henderson
2021-06-11 23:41 ` [PULL 26/34] tcg: Round the tb_size default from qemu_get_host_physmem Richard Henderson
2021-06-11 23:41 ` [PULL 27/34] tcg: Merge buffer protection and guard page protection Richard Henderson
2021-06-11 23:41 ` [PULL 28/34] tcg: When allocating for !splitwx, begin with PROT_NONE Richard Henderson
2021-06-11 23:41 ` [PULL 29/34] tcg: Move tcg_init_ctx and tcg_ctx from accel/tcg/ Richard Henderson
2021-06-11 23:41 ` [PULL 30/34] tcg: Introduce tcg_remove_ops_after Richard Henderson
2021-06-11 23:41 ` [PULL 31/34] tcg: Fix documentation for tcg_constant_* vs tcg_temp_free_* Richard Henderson
2021-06-11 23:41 ` [PULL 32/34] tcg/arm: Fix tcg_out_op function signature Richard Henderson
2021-06-11 23:41 ` [PULL 33/34] softfloat: Fix tp init in float32_exp2 Richard Henderson
2021-06-11 23:41 ` Richard Henderson [this message]
2021-06-13 15:13 ` [PULL 00/34] tcg patch queue Peter Maydell
2021-06-13 17:10   ` Peter Maydell
2021-06-14  1:37     ` Richard Henderson
2021-06-14  9:28       ` Peter Maydell
2021-06-14  9:35       ` Alex Bennée
2021-06-14 15:00         ` Philippe Mathieu-Daudé

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=20210611234144.653682-35-richard.henderson@linaro.org \
    --to=richard.henderson@linaro.org \
    --cc=luis.pires@eldorado.org.br \
    --cc=qemu-devel@nongnu.org \
    --subject='Re: [PULL 34/34] docs/devel: Explain in more detail the TB chaining mechanisms' \
    /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

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.