[Lldb-commits] [lldb] r224698 - Add support for frameless function compact unwind encodings on x86_64/i386.
Jason Molenda
jmolenda at apple.com
Mon Dec 22 03:02:03 PST 2014
Author: jmolenda
Date: Mon Dec 22 05:02:02 2014
New Revision: 224698
URL: http://llvm.org/viewvc/llvm-project?rev=224698&view=rev
Log:
Add support for frameless function compact unwind encodings on x86_64/i386.
This completes the compact unwind support for x86 targets.
I'm still skipping the UNWIND_X86_64_MODE_STACK_IND encodings for
x86_64 right now because clang was emitting bad data for this form
until it was fixed in r217020 circa Sep 2014.
arm64 parsing still needs to be added.
Modified:
lldb/trunk/source/Symbol/CompactUnwindInfo.cpp
Modified: lldb/trunk/source/Symbol/CompactUnwindInfo.cpp
URL: http://llvm.org/viewvc/llvm-project/lldb/trunk/source/Symbol/CompactUnwindInfo.cpp?rev=224698&r1=224697&r2=224698&view=diff
==============================================================================
--- lldb/trunk/source/Symbol/CompactUnwindInfo.cpp (original)
+++ lldb/trunk/source/Symbol/CompactUnwindInfo.cpp Mon Dec 22 05:02:02 2014
@@ -759,46 +759,63 @@ CompactUnwindInfo::CreateUnwindPlan_x86_
case UNWIND_X86_64_MODE_STACK_IND:
{
// The clang in Xcode 6 is emitting incorrect compact unwind encodings for this
- // style of unwind. It was fixed in llvm r217020 although the algorith being
- // used to compute this style of unwind in generateCompactUnwindEncodingImpl()
- // isn't as foolproof as I'm comfortable with -- if any instructions other than
- // a push are scheduled before the subq, it will give bogus encoding results.
-
- // The target and pc_or_function_start arguments will be needed to handle this
- // encoding style correctly -- to find the start address of the function and
- // read memory offset from there.
+ // style of unwind. It was fixed in llvm r217020.
return false;
}
break;
-#if 0
case UNWIND_X86_64_MODE_STACK_IMMD:
{
- uint32_t stack_size = EXTRACT_BITS (encoding, UNWIND_X86_64_FRAMELESS_STACK_SIZE);
- uint32_t register_count = EXTRACT_BITS (encoding, UNWIND_X86_64_FRAMELESS_STACK_REG_COUNT);
- uint32_t permutation = EXTRACT_BITS (encoding, UNWIND_X86_64_FRAMELESS_STACK_REG_PERMUTATION);
+ uint32_t stack_size = EXTRACT_BITS (function_info.encoding, UNWIND_X86_64_FRAMELESS_STACK_SIZE);
+ uint32_t register_count = EXTRACT_BITS (function_info.encoding, UNWIND_X86_64_FRAMELESS_STACK_REG_COUNT);
+ uint32_t permutation = EXTRACT_BITS (function_info.encoding, UNWIND_X86_64_FRAMELESS_STACK_REG_PERMUTATION);
- if (mode == UNWIND_X86_64_MODE_STACK_IND && function_start)
+ if (mode == UNWIND_X86_64_MODE_STACK_IND && function_info.valid_range_offset_start != 0)
{
- uint32_t stack_adjust = EXTRACT_BITS (encoding, UNWIND_X86_64_FRAMELESS_STACK_ADJUST);
+ uint32_t stack_adjust = EXTRACT_BITS (function_info.encoding, UNWIND_X86_64_FRAMELESS_STACK_ADJUST);
// offset into the function instructions; 0 == beginning of first instruction
- uint32_t offset_to_subl_insn = EXTRACT_BITS (encoding, UNWIND_X86_64_FRAMELESS_STACK_SIZE);
+ uint32_t offset_to_subl_insn = EXTRACT_BITS (function_info.encoding, UNWIND_X86_64_FRAMELESS_STACK_SIZE);
- stack_size = *((uint32_t*) (function_start + offset_to_subl_insn));
-
- stack_size += stack_adjust * 8;
-
- printf ("large stack ");
+ SectionList *sl = m_objfile.GetSectionList ();
+ if (sl)
+ {
+ ProcessSP process_sp = target.GetProcessSP();
+ if (process_sp)
+ {
+ Address subl_payload_addr (function_info.valid_range_offset_start, sl);
+ subl_payload_addr.Slide (offset_to_subl_insn);
+ Error error;
+ uint64_t large_stack_size = process_sp->ReadUnsignedIntegerFromMemory (subl_payload_addr.GetLoadAddress (&target),
+ 4, 0, error);
+ if (large_stack_size != 0 && error.Success ())
+ {
+ // Got the large stack frame size correctly - use it
+ stack_size = large_stack_size + (stack_adjust * wordsize);
+ }
+ else
+ {
+ return false;
+ }
+ }
+ else
+ {
+ return false;
+ }
+ }
+ else
+ {
+ return false;
+ }
}
-
- printf ("frameless function: stack size %d, register count %d ", stack_size * 8, register_count);
- if (register_count == 0)
- {
- printf (" no registers saved");
- }
- else
+ row->SetCFARegister (x86_64_eh_regnum::rsp);
+ row->SetCFAOffset (stack_size * wordsize);
+ row->SetOffset (0);
+ row->SetRegisterLocationToAtCFAPlusOffset (x86_64_eh_regnum::rip, wordsize * -1, true);
+ row->SetRegisterLocationToIsCFAPlusOffset (x86_64_eh_regnum::rsp, 0, true);
+
+ if (register_count > 0)
{
// We need to include (up to) 6 registers in 10 bits.
@@ -887,11 +904,7 @@ CompactUnwindInfo::CreateUnwindPlan_x86_
}
}
-
- printf (" CFA is rsp+%d ", stack_size * 8);
-
uint32_t saved_registers_offset = 1;
- printf (" rip=[CFA-%d]", saved_registers_offset * 8);
saved_registers_offset++;
for (int i = (sizeof (registers) / sizeof (int)) - 1; i >= 0; i--)
@@ -901,32 +914,21 @@ CompactUnwindInfo::CreateUnwindPlan_x86_
case UNWIND_X86_64_REG_NONE:
break;
case UNWIND_X86_64_REG_RBX:
- printf (" rbx=[CFA-%d]", saved_registers_offset * 8);
- break;
case UNWIND_X86_64_REG_R12:
- printf (" r12=[CFA-%d]", saved_registers_offset * 8);
- break;
case UNWIND_X86_64_REG_R13:
- printf (" r13=[CFA-%d]", saved_registers_offset * 8);
- break;
case UNWIND_X86_64_REG_R14:
- printf (" r14=[CFA-%d]", saved_registers_offset * 8);
- break;
case UNWIND_X86_64_REG_R15:
- printf (" r15=[CFA-%d]", saved_registers_offset * 8);
- break;
case UNWIND_X86_64_REG_RBP:
- printf (" rbp=[CFA-%d]", saved_registers_offset * 8);
- break;
+ row->SetRegisterLocationToAtCFAPlusOffset (translate_to_eh_frame_regnum_x86_64 (registers[i]), wordsize * -saved_registers_offset, true);
+ break;
}
saved_registers_offset++;
}
-
}
-
+ unwind_plan.AppendRow (row);
+ return true;
}
break;
-#endif
case UNWIND_X86_64_MODE_DWARF:
{
@@ -1037,6 +1039,172 @@ CompactUnwindInfo::CreateUnwindPlan_i386
case UNWIND_X86_MODE_STACK_IND:
case UNWIND_X86_MODE_STACK_IMMD:
+ {
+ uint32_t stack_size = EXTRACT_BITS (function_info.encoding, UNWIND_X86_FRAMELESS_STACK_SIZE);
+ uint32_t register_count = EXTRACT_BITS (function_info.encoding, UNWIND_X86_FRAMELESS_STACK_REG_COUNT);
+ uint32_t permutation = EXTRACT_BITS (function_info.encoding, UNWIND_X86_FRAMELESS_STACK_REG_PERMUTATION);
+
+ if (mode == UNWIND_X86_MODE_STACK_IND && function_info.valid_range_offset_start != 0)
+ {
+ uint32_t stack_adjust = EXTRACT_BITS (function_info.encoding, UNWIND_X86_FRAMELESS_STACK_ADJUST);
+
+ // offset into the function instructions; 0 == beginning of first instruction
+ uint32_t offset_to_subl_insn = EXTRACT_BITS (function_info.encoding, UNWIND_X86_FRAMELESS_STACK_SIZE);
+
+ SectionList *sl = m_objfile.GetSectionList ();
+ if (sl)
+ {
+ ProcessSP process_sp = target.GetProcessSP();
+ if (process_sp)
+ {
+ Address subl_payload_addr (function_info.valid_range_offset_start, sl);
+ subl_payload_addr.Slide (offset_to_subl_insn);
+ Error error;
+ uint64_t large_stack_size = process_sp->ReadUnsignedIntegerFromMemory (subl_payload_addr.GetLoadAddress (&target),
+ 4, 0, error);
+ if (large_stack_size != 0 && error.Success ())
+ {
+ // Got the large stack frame size correctly - use it
+ stack_size = large_stack_size + (stack_adjust * wordsize);
+ }
+ else
+ {
+ return false;
+ }
+ }
+ else
+ {
+ return false;
+ }
+ }
+ else
+ {
+ return false;
+ }
+ }
+
+ row->SetCFARegister (i386_eh_regnum::esp);
+ row->SetCFAOffset (stack_size * wordsize);
+ row->SetOffset (0);
+ row->SetRegisterLocationToAtCFAPlusOffset (i386_eh_regnum::eip, wordsize * -1, true);
+ row->SetRegisterLocationToIsCFAPlusOffset (i386_eh_regnum::esp, 0, true);
+
+ if (register_count > 0)
+ {
+
+ // We need to include (up to) 6 registers in 10 bits.
+ // That would be 18 bits if we just used 3 bits per reg to indicate
+ // the order they're saved on the stack.
+ //
+ // This is done with Lehmer code permutation, e.g. see
+ // http://stackoverflow.com/questions/1506078/fast-permutation-number-permutation-mapping-algorithms
+ int permunreg[6];
+
+ // This decodes the variable-base number in the 10 bits
+ // and gives us the Lehmer code sequence which can then
+ // be decoded.
+
+ switch (register_count)
+ {
+ case 6:
+ permunreg[0] = permutation/120; // 120 == 5!
+ permutation -= (permunreg[0]*120);
+ permunreg[1] = permutation/24; // 24 == 4!
+ permutation -= (permunreg[1]*24);
+ permunreg[2] = permutation/6; // 6 == 3!
+ permutation -= (permunreg[2]*6);
+ permunreg[3] = permutation/2; // 2 == 2!
+ permutation -= (permunreg[3]*2);
+ permunreg[4] = permutation; // 1 == 1!
+ permunreg[5] = 0;
+ break;
+ case 5:
+ permunreg[0] = permutation/120;
+ permutation -= (permunreg[0]*120);
+ permunreg[1] = permutation/24;
+ permutation -= (permunreg[1]*24);
+ permunreg[2] = permutation/6;
+ permutation -= (permunreg[2]*6);
+ permunreg[3] = permutation/2;
+ permutation -= (permunreg[3]*2);
+ permunreg[4] = permutation;
+ break;
+ case 4:
+ permunreg[0] = permutation/60;
+ permutation -= (permunreg[0]*60);
+ permunreg[1] = permutation/12;
+ permutation -= (permunreg[1]*12);
+ permunreg[2] = permutation/3;
+ permutation -= (permunreg[2]*3);
+ permunreg[3] = permutation;
+ break;
+ case 3:
+ permunreg[0] = permutation/20;
+ permutation -= (permunreg[0]*20);
+ permunreg[1] = permutation/4;
+ permutation -= (permunreg[1]*4);
+ permunreg[2] = permutation;
+ break;
+ case 2:
+ permunreg[0] = permutation/5;
+ permutation -= (permunreg[0]*5);
+ permunreg[1] = permutation;
+ break;
+ case 1:
+ permunreg[0] = permutation;
+ break;
+ }
+
+ // Decode the Lehmer code for this permutation of
+ // the registers v. http://en.wikipedia.org/wiki/Lehmer_code
+
+ int registers[6];
+ bool used[7] = { false, false, false, false, false, false, false };
+ for (int i = 0; i < register_count; i++)
+ {
+ int renum = 0;
+ for (int j = 1; j < 7; j++)
+ {
+ if (used[j] == false)
+ {
+ if (renum == permunreg[i])
+ {
+ registers[i] = j;
+ used[j] = true;
+ break;
+ }
+ renum++;
+ }
+ }
+ }
+
+ uint32_t saved_registers_offset = 1;
+ saved_registers_offset++;
+
+ for (int i = (sizeof (registers) / sizeof (int)) - 1; i >= 0; i--)
+ {
+ switch (registers[i])
+ {
+ case UNWIND_X86_REG_NONE:
+ break;
+ case UNWIND_X86_REG_EBX:
+ case UNWIND_X86_REG_ECX:
+ case UNWIND_X86_REG_EDX:
+ case UNWIND_X86_REG_EDI:
+ case UNWIND_X86_REG_ESI:
+ case UNWIND_X86_REG_EBP:
+ row->SetRegisterLocationToAtCFAPlusOffset (translate_to_eh_frame_regnum_i386 (registers[i]), wordsize * -saved_registers_offset, true);
+ break;
+ }
+ saved_registers_offset++;
+ }
+ }
+
+ unwind_plan.AppendRow (row);
+ return true;
+ }
+ break;
+
case UNWIND_X86_MODE_DWARF:
{
return false;
More information about the lldb-commits
mailing list