Skip to contentSkip to conversation

target/xtensa: convert to do_transaction_failed

PATCH Superseded Max Filippov 2 participants 4 messages Aug 20, 2018 – Aug 24, 2018

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore

Max Filippov

author
#0 lore raw

PATCH 1/1 target/xtensa: convert to do_transaction_failed

Signed-off-by Max Filippov <[email protected]>
---
 target/xtensa/cpu.c       |  2 +-
 target/xtensa/cpu.h       |  7 ++++---
 target/xtensa/op_helper.c | 12 +++++++-----
 3 files changed, 12 insertions(+), 9 deletions(-)
target/xtensa/cpu.c +1 -1
index 590813d4f7b9..a54dbe42602d 100644
@@ -186,7 +186,7 @@ static void xtensa_cpu_class_init(ObjectClass *oc, void *data)
 #else
     cc->do_unaligned_access = xtensa_cpu_do_unaligned_access;
     cc->get_phys_page_debug = xtensa_cpu_get_phys_page_debug;
-    cc->do_unassigned_access = xtensa_cpu_do_unassigned_access;
+    cc->do_transaction_failed = xtensa_cpu_do_transaction_failed;
 #endif
     cc->debug_excp_handler = xtensa_breakpoint_handler;
     cc->disas_set_info = xtensa_cpu_disas_set_info;
target/xtensa/cpu.h +4 -3
index 7472cf3ca32a..1362772617ea 100644
@@ -497,9 +497,10 @@ int xtensa_cpu_handle_mmu_fault(CPUState *cs, vaddr address, int rw, int size,
                                 int mmu_idx);
 void xtensa_cpu_do_interrupt(CPUState *cpu);
 bool xtensa_cpu_exec_interrupt(CPUState *cpu, int interrupt_request);
-void xtensa_cpu_do_unassigned_access(CPUState *cpu, hwaddr addr,
-                                     bool is_write, bool is_exec, int opaque,
-                                     unsigned size);
+void xtensa_cpu_do_transaction_failed(CPUState *cs, hwaddr physaddr, vaddr addr,
+                                      unsigned size, MMUAccessType access_type,
+                                      int mmu_idx, MemTxAttrs attrs,
+                                      MemTxResult response, uintptr_t retaddr);
 void xtensa_cpu_dump_state(CPUState *cpu, FILE *f,
                            fprintf_function cpu_fprintf, int flags);
 hwaddr xtensa_cpu_get_phys_page_debug(CPUState *cpu, vaddr addr);
target/xtensa/op_helper.c +7 -5
index d4c942d87980..06fe346f02ff 100644
@@ -78,18 +78,20 @@ void tlb_fill(CPUState *cs, target_ulong vaddr, int size,
     }
 }
 
-void xtensa_cpu_do_unassigned_access(CPUState *cs, hwaddr addr,
-                                     bool is_write, bool is_exec, int opaque,
-                                     unsigned size)
+void xtensa_cpu_do_transaction_failed(CPUState *cs, hwaddr physaddr, vaddr addr,
+                                      unsigned size, MMUAccessType access_type,
+                                      int mmu_idx, MemTxAttrs attrs,
+                                      MemTxResult response, uintptr_t retaddr)
 {
     XtensaCPU *cpu = XTENSA_CPU(cs);
     CPUXtensaState *env = &cpu->env;
 
+    cpu_restore_state(cs, retaddr, true);
     HELPER(exception_cause_vaddr)(env, env->pc,
-                                  is_exec ?
+                                  access_type == MMU_INST_FETCH ?
                                   INSTR_PIF_ADDR_ERROR_CAUSE :
                                   LOAD_STORE_PIF_ADDR_ERROR_CAUSE,
-                                  is_exec ? addr : cs->mem_io_vaddr);
+                                  addr);
 }
 
 static void tb_invalidate_virtual_addr(CPUXtensaState *env, uint32_t vaddr)
Signature · 1 line
2.11.0
#1 lore raw
Quoted 6 lines · Max Filippov
Signed-off-by Max Filippov <[email protected]>
---
 target/xtensa/cpu.c       |  2 +-
 target/xtensa/cpu.h       |  7 ++++---
 target/xtensa/op_helper.c | 12 +++++++-----
 3 files changed, 12 insertions(+), 9 deletions(-)

Hi -- thanks for doing this conversion. This change means that you'll no longer get guest exceptions for bus errors that happen during get_pte()'s page table entry fetch. Don't you need at least a TODO comment there indicating that there's missing behaviour?

(Obviously the ideal would be to switch to address_space_ldl_* and handle the case when they don't return MEMTX_OK.)

thanks
-- PMM

author
#2 lore raw

Hi Peter,

Quoted 8 lines · Peter Maydell
Quoted 6 lines · Max Filippov
Signed-off-by Max Filippov <[email protected]>
---
 target/xtensa/cpu.c       |  2 +-
 target/xtensa/cpu.h       |  7 ++++---
 target/xtensa/op_helper.c | 12 +++++++-----
 3 files changed, 12 insertions(+), 9 deletions(-)
Quoted 5 lines

Hi -- thanks for doing this conversion. This change means that you'll no longer get guest exceptions for bus errors that happen during get_pte()'s page table entry fetch. Don't you need at least a TODO comment there indicating that there's missing behaviour?

Thanks for taking a look and letting me know, I've missed that entirely.

Quoted 2 lines

(Obviously the ideal would be to switch to address_space_ldl_* and handle the case when they don't return MEMTX_OK.)

Yeah, let me do something about it.

Signature · 2 lines
Thanks.
-- Max
#3 lore raw
Quoted 12 lines · Max Filippov

Hi Peter,

Quoted 8 lines · Peter Maydell
Quoted 6 lines · Max Filippov
Signed-off-by Max Filippov <[email protected]>
---
 target/xtensa/cpu.c       |  2 +-
 target/xtensa/cpu.h       |  7 ++++---
 target/xtensa/op_helper.c | 12 +++++++-----
 3 files changed, 12 insertions(+), 9 deletions(-)
Quoted 7 lines
Quoted 5 lines

Hi -- thanks for doing this conversion. This change means that you'll no longer get guest exceptions for bus errors that happen during get_pte()'s page table entry fetch. Don't you need at least a TODO comment there indicating that there's missing behaviour?

Thanks for taking a look and letting me know, I've missed that entirely.

Basically the new hook only handles failures for code paths that go via the MMU; anything where the target code does a direct physical address lookup has to handle the failure itself. I think for xtensa that's just the page table reads.

(This is why it's a bit of a pain to convert a target, but the end result makes much more sense, because almost always the target code doing physical address accesses doesn't want to handle bus errors by randomly longjumping out.)

thanks
-- PMM