More mem.c fixes - simplified some unnecessary complexities introduced this morning and rewrote do_mmutranslate() again, fixes #1351.
This commit is contained in:
168
src/mem/mem.c
168
src/mem/mem.c
@@ -993,7 +993,6 @@ uint16_t
|
||||
readmemwl_no_mmut(uint32_t addr, uint64_t *addr64)
|
||||
{
|
||||
mem_mapping_t *map;
|
||||
int i;
|
||||
|
||||
mem_logical_addr = addr;
|
||||
|
||||
@@ -1002,10 +1001,8 @@ readmemwl_no_mmut(uint32_t addr, uint64_t *addr64)
|
||||
cycles -= timing_misaligned;
|
||||
if ((addr & 0xfff) > 0xffe) {
|
||||
if (cr0 >> 31) {
|
||||
for (i = 0; i < 2; i++) {
|
||||
if (addr64[i] > 0xffffffffULL)
|
||||
return 0xffff;
|
||||
}
|
||||
if ((addr64[0] > 0xffffffffULL) || (addr64[1] > 0xffffffffULL))
|
||||
return 0xffff;
|
||||
}
|
||||
|
||||
return readmembl_no_mmut(addr, addr64[0]) |
|
||||
@@ -1041,7 +1038,6 @@ void
|
||||
writememwl_no_mmut(uint32_t addr, uint64_t *addr64, uint16_t val)
|
||||
{
|
||||
mem_mapping_t *map;
|
||||
int i;
|
||||
|
||||
mem_logical_addr = addr;
|
||||
|
||||
@@ -1050,10 +1046,8 @@ writememwl_no_mmut(uint32_t addr, uint64_t *addr64, uint16_t val)
|
||||
cycles -= timing_misaligned;
|
||||
if ((addr & 0xfff) > 0xffe) {
|
||||
if (cr0 >> 31) {
|
||||
for (i = 0; i < 2; i++) {
|
||||
if (addr64[i] > 0xffffffffULL)
|
||||
return;
|
||||
}
|
||||
if ((addr64[0] > 0xffffffffULL) || (addr64[1] > 0xffffffffULL))
|
||||
return;
|
||||
}
|
||||
|
||||
writemembl_no_mmut(addr, addr64[0], val);
|
||||
@@ -1099,7 +1093,7 @@ readmemll(uint32_t addr)
|
||||
{
|
||||
uint64_t addr64[4];
|
||||
mem_mapping_t *map;
|
||||
int i, wrap_i = 1;
|
||||
int i;
|
||||
|
||||
for (i = 0; i < 4; i++)
|
||||
addr64[i] = (uint64_t) (addr + i);
|
||||
@@ -1111,24 +1105,14 @@ readmemll(uint32_t addr)
|
||||
cycles -= timing_misaligned;
|
||||
if ((addr & 0xfff) > 0xffc) {
|
||||
if (cr0 >> 31) {
|
||||
wrap_i = 4 - (addr & 0x3);
|
||||
for (i = 0; i < 4; i++) {
|
||||
if ((i == 0) || !((addr + i) & 0xfff))
|
||||
addr64[i] = mmutranslate_read(addr + i);
|
||||
else
|
||||
addr64[i] = (addr64[i - 1] & ~0xfffLL) | ((uint64_t) ((addr + i) & 0xfff));
|
||||
|
||||
addr64[0] = mmutranslate_read(addr);
|
||||
|
||||
if (addr64[0] > 0xffffffffULL)
|
||||
return 0xffffffff;
|
||||
|
||||
for (i = 1; i < wrap_i; i++)
|
||||
addr64[i] = (addr64[0] & ~0xfffLL) | ((uint64_t) ((addr + i) & 0xfff));
|
||||
|
||||
addr64[wrap_i] = mmutranslate_read(addr + wrap_i);
|
||||
|
||||
if (addr64[wrap_i] > 0xffffffffULL)
|
||||
return 0xffffffff;
|
||||
|
||||
if (wrap_i != 3) {
|
||||
for (i = (wrap_i) + 1; i <= 3; i++)
|
||||
addr64[i] = (addr64[wrap_i] & ~0xfffLL) | ((uint64_t) ((addr + i) & 0xfff));
|
||||
if (addr64[i] > 0xffffffffULL)
|
||||
return 0xffff;
|
||||
}
|
||||
}
|
||||
|
||||
@@ -1173,7 +1157,7 @@ writememll(uint32_t addr, uint32_t val)
|
||||
{
|
||||
uint64_t addr64[4];
|
||||
mem_mapping_t *map;
|
||||
int i, wrap_i = 1;
|
||||
int i;
|
||||
|
||||
for (i = 0; i < 4; i++)
|
||||
addr64[i] = (uint64_t) (addr + i);
|
||||
@@ -1185,30 +1169,18 @@ writememll(uint32_t addr, uint32_t val)
|
||||
cycles -= timing_misaligned;
|
||||
if ((addr & 0xfff) > 0xffc) {
|
||||
if (cr0 >> 31) {
|
||||
wrap_i = 4 - (addr & 0x3);
|
||||
for (i = 0; i < 4; i++) {
|
||||
/* Do not translate a page that has a valid lookup, as that is by definition valid
|
||||
and the whole purpose of the lookup is to avoid repeat identical translations. */
|
||||
if (!page_lookup[(addr + i) >> 12] || !page_lookup[(addr + i) >> 12]->write_b) {
|
||||
if ((i == 0) || !((addr + i) & 0xfff))
|
||||
addr64[i] = mmutranslate_write(addr + i);
|
||||
else
|
||||
addr64[i] = (addr64[i - 1] & ~0xfffLL) | ((uint64_t) ((addr + i) & 0xfff));
|
||||
|
||||
if (!page_lookup[(addr + 0) >> 12] || !page_lookup[(addr + 0) >> 12]->write_b)
|
||||
addr64[0] = mmutranslate_write(addr);
|
||||
else
|
||||
addr64[0] = 0xffffffff;
|
||||
|
||||
if (addr64[0] > 0xffffffffULL)
|
||||
return;
|
||||
|
||||
for (i = 1; i < wrap_i; i++)
|
||||
addr64[i] = (addr64[0] & ~0xfffLL) | ((uint64_t) ((addr + i) & 0xfff));
|
||||
|
||||
if (!page_lookup[(addr + wrap_i) >> 12] || !page_lookup[(addr + wrap_i) >> 12]->write_b)
|
||||
addr64[wrap_i] = mmutranslate_write(addr + wrap_i);
|
||||
else
|
||||
addr64[wrap_i] = 0xffffffff;
|
||||
|
||||
if (addr64[wrap_i] > 0xffffffffULL)
|
||||
return;
|
||||
|
||||
if (wrap_i != 3) {
|
||||
for (i = (wrap_i) + 1; i <= 3; i++)
|
||||
addr64[i] = (addr64[wrap_i] & ~0xfffLL) | ((uint64_t) ((addr + i) & 0xfff));
|
||||
if (addr64[i] > 0xffffffffULL)
|
||||
return;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
@@ -1263,7 +1235,6 @@ uint32_t
|
||||
readmemll_no_mmut(uint32_t addr, uint64_t *addr64)
|
||||
{
|
||||
mem_mapping_t *map;
|
||||
int i;
|
||||
|
||||
mem_logical_addr = addr;
|
||||
|
||||
@@ -1272,10 +1243,8 @@ readmemll_no_mmut(uint32_t addr, uint64_t *addr64)
|
||||
cycles -= timing_misaligned;
|
||||
if ((addr & 0xfff) > 0xffc) {
|
||||
if (cr0 >> 31) {
|
||||
for (i = 0; i < 4; i += 2) {
|
||||
if (addr64[i] > 0xffffffffULL)
|
||||
return 0xffffffff;
|
||||
}
|
||||
if ((addr64[0] > 0xffffffffULL) || (addr64[3] > 0xffffffffULL))
|
||||
return 0xffffffff;
|
||||
}
|
||||
|
||||
return readmemwl_no_mmut(addr, addr64) |
|
||||
@@ -1316,7 +1285,6 @@ void
|
||||
writememll_no_mmut(uint32_t addr, uint64_t *addr64, uint32_t val)
|
||||
{
|
||||
mem_mapping_t *map;
|
||||
int i;
|
||||
|
||||
mem_logical_addr = addr;
|
||||
|
||||
@@ -1325,10 +1293,8 @@ writememll_no_mmut(uint32_t addr, uint64_t *addr64, uint32_t val)
|
||||
cycles -= timing_misaligned;
|
||||
if ((addr & 0xfff) > 0xffc) {
|
||||
if (cr0 >> 31) {
|
||||
for (i = 0; i < 4; i += 2) {
|
||||
if (addr64[i] > 0xffffffffULL)
|
||||
return;
|
||||
}
|
||||
if ((addr64[0] > 0xffffffffULL) || (addr64[3] > 0xffffffffULL))
|
||||
return;
|
||||
}
|
||||
|
||||
writememwl_no_mmut(addr, &(addr64[0]), val);
|
||||
@@ -1390,26 +1356,17 @@ readmemql(uint32_t addr)
|
||||
cycles -= timing_misaligned;
|
||||
if ((addr & 0xfff) > 0xff8) {
|
||||
if (cr0 >> 31) {
|
||||
wrap_i = 8 - (addr & 0x7);
|
||||
|
||||
addr64[0] = mmutranslate_read(addr);
|
||||
|
||||
if (addr64[0] > 0xffffffffULL)
|
||||
return 0xffffffffffffffffULL;
|
||||
|
||||
for (i = 1; i < wrap_i; i++)
|
||||
addr64[i] = (addr64[0] & ~0xfffLL) | ((uint64_t) ((addr + i) & 0xfff));
|
||||
|
||||
addr64[wrap_i] = mmutranslate_read(addr + wrap_i);
|
||||
|
||||
if (addr64[wrap_i] > 0xffffffffULL)
|
||||
return 0xffffffffffffffffULL;
|
||||
|
||||
if (wrap_i != 7) {
|
||||
for (i = (wrap_i) + 1; i <= 7; i++)
|
||||
for (i = 0; i < 8; i++) {
|
||||
if ((i == 0) || !((addr + i) & 0xfff))
|
||||
addr64[i] = mmutranslate_read(addr + i);
|
||||
else
|
||||
addr64[i] = (addr64[wrap_i] & ~0xfffLL) | ((uint64_t) ((addr + i) & 0xfff));
|
||||
|
||||
if (addr64[i] > 0xffffffffULL)
|
||||
return 0xffff;
|
||||
}
|
||||
}
|
||||
|
||||
/* No need to waste precious CPU host cycles on mmutranslate's that were already done, just pass
|
||||
their result as a parameter to be used if needed. */
|
||||
return readmemll_no_mmut(addr, addr64) |
|
||||
@@ -1440,7 +1397,7 @@ writememql(uint32_t addr, uint64_t val)
|
||||
{
|
||||
uint64_t addr64[8];
|
||||
mem_mapping_t *map;
|
||||
int i, wrap_i;
|
||||
int i;
|
||||
|
||||
for (i = 0; i < 8; i++)
|
||||
addr64[i] = (uint64_t) (addr + i);
|
||||
@@ -1451,30 +1408,18 @@ writememql(uint32_t addr, uint64_t val)
|
||||
cycles -= timing_misaligned;
|
||||
if ((addr & 0xfff) > 0xff8) {
|
||||
if (cr0 >> 31) {
|
||||
wrap_i = 8 - (addr & 0x7);
|
||||
for (i = 0; i < 8; i++) {
|
||||
/* Do not translate a page that has a valid lookup, as that is by definition valid
|
||||
and the whole purpose of the lookup is to avoid repeat identical translations. */
|
||||
if (!page_lookup[(addr + i) >> 12] || !page_lookup[(addr + i) >> 12]->write_b) {
|
||||
if ((i == 0) || !((addr + i) & 0xfff))
|
||||
addr64[i] = mmutranslate_write(addr + i);
|
||||
else
|
||||
addr64[i] = (addr64[i - 1] & ~0xfffLL) | ((uint64_t) ((addr + i) & 0xfff));
|
||||
|
||||
if (!page_lookup[(addr + 0) >> 12] || !page_lookup[(addr + 0) >> 12]->write_b)
|
||||
addr64[0] = mmutranslate_write(addr);
|
||||
else
|
||||
addr64[0] = 0xffffffff;
|
||||
|
||||
if (addr64[0] > 0xffffffffULL)
|
||||
return;
|
||||
|
||||
for (i = 1; i < wrap_i; i++)
|
||||
addr64[i] = (addr64[wrap_i] & ~0xfffLL) | ((uint64_t) ((addr + i) & 0xfff));
|
||||
|
||||
if (!page_lookup[(addr + wrap_i) >> 12] || !page_lookup[(addr + wrap_i) >> 12]->write_b)
|
||||
addr64[wrap_i] = mmutranslate_write(addr + wrap_i);
|
||||
else
|
||||
addr64[wrap_i] = 0xffffffff;
|
||||
|
||||
if (addr64[wrap_i] > 0xffffffffULL)
|
||||
return;
|
||||
|
||||
if (wrap_i != 7) {
|
||||
for (i = (wrap_i) + 1; i <= 7; i++)
|
||||
addr64[i] = (addr64[wrap_i] & ~0xfffLL) | ((uint64_t) ((addr + i) & 0xfff));
|
||||
if (addr64[i] > 0xffffffffULL)
|
||||
return;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
@@ -1535,20 +1480,19 @@ writememql(uint32_t addr, uint64_t val)
|
||||
void
|
||||
do_mmutranslate(uint32_t addr, uint64_t *addr64, int num, int write)
|
||||
{
|
||||
int i, wrap_i;
|
||||
int cond = 1;
|
||||
int i, cond = 1;
|
||||
|
||||
if (cr0 >> 31) {
|
||||
wrap_i = (num - (addr & (num - 1))) & (num - 1);
|
||||
for (i = 0; i < num; i++) {
|
||||
addr64[i] = (uint64_t) addr;
|
||||
|
||||
for (i = 0; i < num; i++) {
|
||||
if (write && ((i == 0) || (wrap_i && (i == wrap_i))))
|
||||
if (cr0 >> 31) {
|
||||
if (write && ((i == 0) || !(addr & 0xfff)))
|
||||
cond = (!page_lookup[addr >> 12] || !page_lookup[addr >> 12]->write_b);
|
||||
|
||||
if (cond) {
|
||||
/* If we are on the same page, there is no need to translate again, as we can just
|
||||
reuse the previous result. */
|
||||
if (((i == 0) || (wrap_i && (i == wrap_i)))) {
|
||||
if ((i == 0) || !(addr & 0xfff)) {
|
||||
addr64[i] = mmutranslatereal(addr, write);
|
||||
|
||||
if (addr64[i] > 0xffffffffULL)
|
||||
@@ -1556,9 +1500,9 @@ do_mmutranslate(uint32_t addr, uint64_t *addr64, int num, int write)
|
||||
} else
|
||||
addr64[i] = (addr64[i - 1] & ~0xfffLL) | ((uint64_t) (addr & 0xfff));
|
||||
}
|
||||
|
||||
addr++;
|
||||
}
|
||||
|
||||
addr++;
|
||||
}
|
||||
}
|
||||
|
||||
|
Reference in New Issue
Block a user