[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Qemu-devel] [PATCH v4 11/35] cputlb: Move most of iotlb code out of lin
From: |
Richard Henderson |
Subject: |
[Qemu-devel] [PATCH v4 11/35] cputlb: Move most of iotlb code out of line |
Date: |
Fri, 16 Sep 2016 10:46:33 -0700 |
Saves 2k code size off of a cold path.
Signed-off-by: Richard Henderson <address@hidden>
---
cputlb.c | 37 +++++++++++++++++++++++++++++++++++++
softmmu_template.h | 52 ++++++++++------------------------------------------
2 files changed, 47 insertions(+), 42 deletions(-)
diff --git a/cputlb.c b/cputlb.c
index 0c9b77b..1bee47d 100644
--- a/cputlb.c
+++ b/cputlb.c
@@ -498,6 +498,43 @@ tb_page_addr_t get_page_addr_code(CPUArchState *env1,
target_ulong addr)
return qemu_ram_addr_from_host_nofail(p);
}
+static uint64_t io_readx(CPUArchState *env, CPUIOTLBEntry *iotlbentry,
+ target_ulong addr, uintptr_t retaddr, int size)
+{
+ CPUState *cpu = ENV_GET_CPU(env);
+ hwaddr physaddr = iotlbentry->addr;
+ MemoryRegion *mr = iotlb_to_region(cpu, physaddr, iotlbentry->attrs);
+ uint64_t val;
+
+ physaddr = (physaddr & TARGET_PAGE_MASK) + addr;
+ cpu->mem_io_pc = retaddr;
+ if (mr != &io_mem_rom && mr != &io_mem_notdirty && !cpu->can_do_io) {
+ cpu_io_recompile(cpu, retaddr);
+ }
+
+ cpu->mem_io_vaddr = addr;
+ memory_region_dispatch_read(mr, physaddr, &val, size, iotlbentry->attrs);
+ return val;
+}
+
+static void io_writex(CPUArchState *env, CPUIOTLBEntry *iotlbentry,
+ uint64_t val, target_ulong addr,
+ uintptr_t retaddr, int size)
+{
+ CPUState *cpu = ENV_GET_CPU(env);
+ hwaddr physaddr = iotlbentry->addr;
+ MemoryRegion *mr = iotlb_to_region(cpu, physaddr, iotlbentry->attrs);
+
+ physaddr = (physaddr & TARGET_PAGE_MASK) + addr;
+ if (mr != &io_mem_rom && mr != &io_mem_notdirty && !cpu->can_do_io) {
+ cpu_io_recompile(cpu, retaddr);
+ }
+
+ cpu->mem_io_vaddr = addr;
+ cpu->mem_io_pc = retaddr;
+ memory_region_dispatch_write(mr, physaddr, val, size, iotlbentry->attrs);
+}
+
/* Return true if ADDR is present in the victim tlb, and has been copied
back to the main tlb. */
static bool victim_tlb_hit(CPUArchState *env, size_t mmu_idx, size_t index,
diff --git a/softmmu_template.h b/softmmu_template.h
index b9532a4..035ffc8 100644
--- a/softmmu_template.h
+++ b/softmmu_template.h
@@ -112,25 +112,12 @@
#ifndef SOFTMMU_CODE_ACCESS
static inline DATA_TYPE glue(io_read, SUFFIX)(CPUArchState *env,
- CPUIOTLBEntry *iotlbentry,
+ size_t mmu_idx, size_t index,
target_ulong addr,
uintptr_t retaddr)
{
- uint64_t val;
- CPUState *cpu = ENV_GET_CPU(env);
- hwaddr physaddr = iotlbentry->addr;
- MemoryRegion *mr = iotlb_to_region(cpu, physaddr, iotlbentry->attrs);
-
- physaddr = (physaddr & TARGET_PAGE_MASK) + addr;
- cpu->mem_io_pc = retaddr;
- if (mr != &io_mem_rom && mr != &io_mem_notdirty && !cpu->can_do_io) {
- cpu_io_recompile(cpu, retaddr);
- }
-
- cpu->mem_io_vaddr = addr;
- memory_region_dispatch_read(mr, physaddr, &val, DATA_SIZE,
- iotlbentry->attrs);
- return val;
+ CPUIOTLBEntry *iotlbentry = &env->iotlb[mmu_idx][index];
+ return io_readx(env, iotlbentry, addr, retaddr, DATA_SIZE);
}
#endif
@@ -161,15 +148,13 @@ WORD_TYPE helper_le_ld_name(CPUArchState *env,
target_ulong addr,
/* Handle an IO access. */
if (unlikely(tlb_addr & ~TARGET_PAGE_MASK)) {
- CPUIOTLBEntry *iotlbentry;
if ((addr & (DATA_SIZE - 1)) != 0) {
goto do_unaligned_access;
}
- iotlbentry = &env->iotlb[mmu_idx][index];
/* ??? Note that the io helpers always read data in the target
byte ordering. We should push the LE/BE request down into io. */
- res = glue(io_read, SUFFIX)(env, iotlbentry, addr, retaddr);
+ res = glue(io_read, SUFFIX)(env, mmu_idx, index, addr, retaddr);
res = TGT_LE(res);
return res;
}
@@ -230,15 +215,13 @@ WORD_TYPE helper_be_ld_name(CPUArchState *env,
target_ulong addr,
/* Handle an IO access. */
if (unlikely(tlb_addr & ~TARGET_PAGE_MASK)) {
- CPUIOTLBEntry *iotlbentry;
if ((addr & (DATA_SIZE - 1)) != 0) {
goto do_unaligned_access;
}
- iotlbentry = &env->iotlb[mmu_idx][index];
/* ??? Note that the io helpers always read data in the target
byte ordering. We should push the LE/BE request down into io. */
- res = glue(io_read, SUFFIX)(env, iotlbentry, addr, retaddr);
+ res = glue(io_read, SUFFIX)(env, mmu_idx, index, addr, retaddr);
res = TGT_BE(res);
return res;
}
@@ -289,24 +272,13 @@ WORD_TYPE helper_be_lds_name(CPUArchState *env,
target_ulong addr,
#endif
static inline void glue(io_write, SUFFIX)(CPUArchState *env,
- CPUIOTLBEntry *iotlbentry,
+ size_t mmu_idx, size_t index,
DATA_TYPE val,
target_ulong addr,
uintptr_t retaddr)
{
- CPUState *cpu = ENV_GET_CPU(env);
- hwaddr physaddr = iotlbentry->addr;
- MemoryRegion *mr = iotlb_to_region(cpu, physaddr, iotlbentry->attrs);
-
- physaddr = (physaddr & TARGET_PAGE_MASK) + addr;
- if (mr != &io_mem_rom && mr != &io_mem_notdirty && !cpu->can_do_io) {
- cpu_io_recompile(cpu, retaddr);
- }
-
- cpu->mem_io_vaddr = addr;
- cpu->mem_io_pc = retaddr;
- memory_region_dispatch_write(mr, physaddr, val, DATA_SIZE,
- iotlbentry->attrs);
+ CPUIOTLBEntry *iotlbentry = &env->iotlb[mmu_idx][index];
+ return io_writex(env, iotlbentry, val, addr, retaddr, DATA_SIZE);
}
void helper_le_st_name(CPUArchState *env, target_ulong addr, DATA_TYPE val,
@@ -334,16 +306,14 @@ void helper_le_st_name(CPUArchState *env, target_ulong
addr, DATA_TYPE val,
/* Handle an IO access. */
if (unlikely(tlb_addr & ~TARGET_PAGE_MASK)) {
- CPUIOTLBEntry *iotlbentry;
if ((addr & (DATA_SIZE - 1)) != 0) {
goto do_unaligned_access;
}
- iotlbentry = &env->iotlb[mmu_idx][index];
/* ??? Note that the io helpers always read data in the target
byte ordering. We should push the LE/BE request down into io. */
val = TGT_LE(val);
- glue(io_write, SUFFIX)(env, iotlbentry, val, addr, retaddr);
+ glue(io_write, SUFFIX)(env, mmu_idx, index, val, addr, retaddr);
return;
}
@@ -412,16 +382,14 @@ void helper_be_st_name(CPUArchState *env, target_ulong
addr, DATA_TYPE val,
/* Handle an IO access. */
if (unlikely(tlb_addr & ~TARGET_PAGE_MASK)) {
- CPUIOTLBEntry *iotlbentry;
if ((addr & (DATA_SIZE - 1)) != 0) {
goto do_unaligned_access;
}
- iotlbentry = &env->iotlb[mmu_idx][index];
/* ??? Note that the io helpers always read data in the target
byte ordering. We should push the LE/BE request down into io. */
val = TGT_BE(val);
- glue(io_write, SUFFIX)(env, iotlbentry, val, addr, retaddr);
+ glue(io_write, SUFFIX)(env, mmu_idx, index, val, addr, retaddr);
return;
}
--
2.5.5
- [Qemu-devel] [PATCH v4 02/35] atomics: add atomic_op_fetch variants, (continued)
- [Qemu-devel] [PATCH v4 02/35] atomics: add atomic_op_fetch variants, Richard Henderson, 2016/09/16
- [Qemu-devel] [PATCH v4 04/35] int128: Use __int128 if available, Richard Henderson, 2016/09/16
- [Qemu-devel] [PATCH v4 03/35] exec: Avoid direct references to Int128 parts, Richard Henderson, 2016/09/16
- [Qemu-devel] [PATCH v4 05/35] int128: Add int128_make128, Richard Henderson, 2016/09/16
- [Qemu-devel] [PATCH v4 07/35] HACK: Always enable parallel_cpus, Richard Henderson, 2016/09/16
- [Qemu-devel] [PATCH v4 08/35] cputlb: Replace SHIFT with DATA_SIZE, Richard Henderson, 2016/09/16
- [Qemu-devel] [PATCH v4 10/35] cputlb: Remove includes from softmmu_template.h, Richard Henderson, 2016/09/16
- [Qemu-devel] [PATCH v4 09/35] cputlb: Move probe_write out of softmmu_template.h, Richard Henderson, 2016/09/16
- [Qemu-devel] [PATCH v4 12/35] cputlb: Tidy some macros, Richard Henderson, 2016/09/16
- [Qemu-devel] [PATCH v4 16/35] tcg: Emit barriers with parallel_cpus, Richard Henderson, 2016/09/16
- [Qemu-devel] [PATCH v4 11/35] cputlb: Move most of iotlb code out of line,
Richard Henderson <=
- [Qemu-devel] [PATCH v4 17/35] target-i386: emulate LOCK'ed cmpxchg using cmpxchg helpers, Richard Henderson, 2016/09/16
- [Qemu-devel] [PATCH v4 19/35] target-i386: emulate LOCK'ed INC using atomic helper, Richard Henderson, 2016/09/16
- [Qemu-devel] [PATCH v4 15/35] tcg: Add CONFIG_ATOMIC64, Richard Henderson, 2016/09/16
- [Qemu-devel] [PATCH v4 14/35] tcg: Add atomic128 helpers, Richard Henderson, 2016/09/16
- [Qemu-devel] [PATCH v4 24/35] target-i386: emulate XCHG using atomic helper, Richard Henderson, 2016/09/16
- [Qemu-devel] [PATCH v4 23/35] target-i386: emulate LOCK'ed BTX ops using atomic helpers, Richard Henderson, 2016/09/16
- [Qemu-devel] [PATCH v4 25/35] target-i386: remove helper_lock(), Richard Henderson, 2016/09/16
- [Qemu-devel] [PATCH v4 26/35] tests: add atomic_add-bench, Richard Henderson, 2016/09/16