[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Qemu-devel] [PULL 09/58] tcg: cpu-exec: remove tb_lock from the hot-pat
From: |
Paolo Bonzini |
Subject: |
[Qemu-devel] [PULL 09/58] tcg: cpu-exec: remove tb_lock from the hot-path |
Date: |
Tue, 13 Sep 2016 19:15:40 +0200 |
From: Alex Bennée <address@hidden>
Lock contention in the hot path of moving between existing patched
TranslationBlocks is the main drag in multithreaded performance. This
patch pushes the tb_lock() usage down to the two places that really need
it:
- code generation (tb_gen_code)
- jump patching (tb_add_jump)
The rest of the code doesn't really need to hold a lock as it is either
using per-CPU structures, atomically updated or designed to be used in
concurrent read situations (qht_lookup).
To keep things simple I removed the #ifdef CONFIG_USER_ONLY stuff as the
locks become NOPs anyway until the MTTCG work is completed.
Signed-off-by: Alex Bennée <address@hidden>
Reviewed-by: Richard Henderson <address@hidden>
Reviewed-by: Sergey Fedorov <address@hidden>
Signed-off-by: Sergey Fedorov <address@hidden>
Message-Id: <address@hidden>
Signed-off-by: Paolo Bonzini <address@hidden>
---
cpu-exec.c | 48 +++++++++++++++++++++---------------------------
1 file changed, 21 insertions(+), 27 deletions(-)
diff --git a/cpu-exec.c b/cpu-exec.c
index cdaab1d..7ca2b71 100644
--- a/cpu-exec.c
+++ b/cpu-exec.c
@@ -287,35 +287,29 @@ static TranslationBlock *tb_find_slow(CPUState *cpu,
TranslationBlock *tb;
tb = tb_find_physical(cpu, pc, cs_base, flags);
- if (tb) {
- goto found;
- }
+ if (!tb) {
-#ifdef CONFIG_USER_ONLY
- /* mmap_lock is needed by tb_gen_code, and mmap_lock must be
- * taken outside tb_lock. Since we're momentarily dropping
- * tb_lock, there's a chance that our desired tb has been
- * translated.
- */
- tb_unlock();
- mmap_lock();
- tb_lock();
- tb = tb_find_physical(cpu, pc, cs_base, flags);
- if (tb) {
- mmap_unlock();
- goto found;
- }
-#endif
+ /* mmap_lock is needed by tb_gen_code, and mmap_lock must be
+ * taken outside tb_lock. As system emulation is currently
+ * single threaded the locks are NOPs.
+ */
+ mmap_lock();
+ tb_lock();
- /* if no translated code available, then translate it now */
- tb = tb_gen_code(cpu, pc, cs_base, flags, 0);
+ /* There's a chance that our desired tb has been translated while
+ * taking the locks so we check again inside the lock.
+ */
+ tb = tb_find_physical(cpu, pc, cs_base, flags);
+ if (!tb) {
+ /* if no translated code available, then translate it now */
+ tb = tb_gen_code(cpu, pc, cs_base, flags, 0);
+ }
-#ifdef CONFIG_USER_ONLY
- mmap_unlock();
-#endif
+ tb_unlock();
+ mmap_unlock();
+ }
-found:
- /* we add the TB in the virtual pc hash table */
+ /* We add the TB in the virtual pc hash table for the fast lookup */
atomic_set(&cpu->tb_jmp_cache[tb_jmp_cache_hash_func(pc)], tb);
return tb;
}
@@ -333,7 +327,6 @@ static inline TranslationBlock *tb_find_fast(CPUState *cpu,
always be the same before a given translated block
is executed. */
cpu_get_tb_cpu_state(env, &pc, &cs_base, &flags);
- tb_lock();
tb = atomic_rcu_read(&cpu->tb_jmp_cache[tb_jmp_cache_hash_func(pc)]);
if (unlikely(!tb || tb->pc != pc || tb->cs_base != cs_base ||
tb->flags != flags)) {
@@ -350,14 +343,15 @@ static inline TranslationBlock *tb_find_fast(CPUState
*cpu,
#endif
/* See if we can patch the calling TB. */
if (last_tb && !qemu_loglevel_mask(CPU_LOG_TB_NOCHAIN)) {
+ tb_lock();
/* Check if translation buffer has been flushed */
if (cpu->tb_flushed) {
cpu->tb_flushed = false;
} else if (!tb->invalid) {
tb_add_jump(last_tb, tb_exit, tb);
}
+ tb_unlock();
}
- tb_unlock();
return tb;
}
--
1.8.3.1
- [Qemu-devel] [PULL 00/58] First round of misc patches for QEMU 2.8, Paolo Bonzini, 2016/09/13
- [Qemu-devel] [PULL 01/58] qtail: clean up direct access to tqe_prev field, Paolo Bonzini, 2016/09/13
- [Qemu-devel] [PULL 04/58] tcg: Pass last_tb by value to tb_find_fast(), Paolo Bonzini, 2016/09/13
- [Qemu-devel] [PULL 03/58] util: fix some coding style issue, Paolo Bonzini, 2016/09/13
- [Qemu-devel] [PULL 05/58] tcg: Prepare safe tb_jmp_cache lookup out of tb_lock, Paolo Bonzini, 2016/09/13
- [Qemu-devel] [PULL 08/58] tcg: set up tb->page_addr before insertion, Paolo Bonzini, 2016/09/13
- [Qemu-devel] [PULL 02/58] util/qemu-sockets: revert Yoda Conditions to normal, Paolo Bonzini, 2016/09/13
- [Qemu-devel] [PULL 07/58] tcg: Prepare TB invalidation for lockless TB lookup, Paolo Bonzini, 2016/09/13
- [Qemu-devel] [PULL 11/58] tcg: Merge tb_find_slow() and tb_find_fast(), Paolo Bonzini, 2016/09/13
- [Qemu-devel] [PULL 09/58] tcg: cpu-exec: remove tb_lock from the hot-path,
Paolo Bonzini <=
- [Qemu-devel] [PULL 12/58] tcg: rename tb_find_physical(), Paolo Bonzini, 2016/09/13
- [Qemu-devel] [PULL 06/58] tcg: Prepare safe access to tb_flushed out of tb_lock, Paolo Bonzini, 2016/09/13
- [Qemu-devel] [PULL 10/58] tcg: Avoid bouncing tb_lock between tb_gen_code() and tb_add_jump(), Paolo Bonzini, 2016/09/13
- [Qemu-devel] [PULL 13/58] rules.mak: Don't extract libs from .mo-libs in link command, Paolo Bonzini, 2016/09/13
- [Qemu-devel] [PULL 14/58] timer: update comments, Paolo Bonzini, 2016/09/13
- [Qemu-devel] [PULL 15/58] cpus: rename local variable to meaningful one, Paolo Bonzini, 2016/09/13
- [Qemu-devel] [PULL 21/58] lsi: never set DMA FIFO Empty (DFE) bit in DSTAT register, Paolo Bonzini, 2016/09/13
- [Qemu-devel] [PULL 23/58] scsi-disk: change disk serial length from 20 to 36, Paolo Bonzini, 2016/09/13
- [Qemu-devel] [PULL 24/58] vmw_pvscsi: check page count while initialising descriptor rings, Paolo Bonzini, 2016/09/13
- [Qemu-devel] [PULL 16/58] cpus: update comments, Paolo Bonzini, 2016/09/13