Bug 172652
| Summary: | bad dwarf location info for multi-register variables | ||||||
|---|---|---|---|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Roland McGrath <roland> | ||||
| Component: | gcc | Assignee: | Jakub Jelinek <jakub> | ||||
| Status: | CLOSED RAWHIDE | QA Contact: | |||||
| Severity: | medium | Docs Contact: | |||||
| Priority: | medium | ||||||
| Version: | rawhide | CC: | fche | ||||
| Target Milestone: | --- | ||||||
| Target Release: | --- | ||||||
| Hardware: | i386 | ||||||
| OS: | Linux | ||||||
| Whiteboard: | |||||||
| Fixed In Version: | 4.0.2-4 | Doc Type: | Bug Fix | ||||
| Doc Text: | Story Points: | --- | |||||
| Clone Of: | Environment: | ||||||
| Last Closed: | 2005-11-10 07:37:57 UTC | Type: | --- | ||||
| Regression: | --- | Mount Type: | --- | ||||
| Documentation: | --- | CRM: | |||||
| Verified Versions: | Category: | --- | |||||
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||
| Cloudforms Team: | --- | Target Upstream Version: | |||||
| Embargoed: | |||||||
| Attachments: |
|
||||||
Created attachment 120795 [details]
C code for test case
Should be fixed in gcc-4.0.2-4. |
Description of problem: Looks like wrong register numbers in location expressions for long long variable (parameter). Looking at multiple_reg_loc_descriptor, it seems like maybe it assumes that contiguous GCC register numbers corresponds to contiguous DWARF register numbers, which is not always so. Version-Release number of selected component (if applicable): gcc-4.0.2-3 used for cited details in this case. Note that FC4's gcc-4.0.1-4.fc4 has the same problems (and more). How reproducible: 100% Steps to Reproduce: 1. gcc -g -c -mregparm -O9 ll-debuginfo-loser.c 2. objdump -rd ll-debuginfo-loser.o; eu-readelf --debug-dump={info,loc} ll-debug-loser.o 3. Compare. Actual results: Disassembly of section .text: 00000000 <seek>: 0: 55 push %ebp 1: 89 e5 mov %esp,%ebp 3: 56 push %esi 4: 53 push %ebx 5: 89 d3 mov %edx,%ebx 7: 89 ce mov %ecx,%esi 9: a3 00 00 00 00 mov %eax,0x0 a: R_386_32 to_ptr e: 89 15 00 00 00 00 mov %edx,0x0 10: R_386_32 to_ll 14: 89 0d 04 00 00 00 mov %ecx,0x4 16: R_386_32 to_ll 1a: 8b 45 08 mov 0x8(%ebp),%eax 1d: a3 00 00 00 00 mov %eax,0x0 1e: R_386_32 to_x 22: 50 push %eax 23: 51 push %ecx 24: 52 push %edx 25: 68 00 00 00 00 push $0x0 26: R_386_32 .rodata.str1.1 2a: e8 fc ff ff ff call 2b <seek+0x2b> 2b: R_386_PC32 printf 2f: 01 db add %ebx,%ebx 31: 11 f6 adc %esi,%esi 33: 89 d8 mov %ebx,%eax 35: 89 f2 mov %esi,%edx 37: 8d 65 f8 lea 0xfffffff8(%ebp),%esp 3a: 5b pop %ebx 3b: 5e pop %esi 3c: c9 leave 3d: c3 ret [ b] compile_unit stmt_list 0 high_pc 0x0000003e low_pc 0000000000 producer "GNU C 4.0.2 20051007 (Red Hat 4.0.2-3)" language ISO C89 (1) name "/net/ma/home/roland/redhat/tmp/ll-debuginfo-loser.c" [ 21] subprogram sibling [ 6e] external name "seek" decl_file 1 decl_line 7 prototyped type [ 6e] low_pc 0000000000 high_pc 0x0000003e frame_base location list [ 0] [ 3e] structure_type name "foo" declaration [ 44] formal_parameter name "ptr" decl_file 1 decl_line 6 type [ 75] location location list [ 2c] [ 53] formal_parameter name "ll" decl_file 1 decl_line 6 type [ 6e] location location list [ 3f] [ 61] formal_parameter name "x" decl_file 1 decl_line 6 type [ 7b] location 2 byte block [ 0] fbreg 0 [ 6e] base_type name "long long int" byte_size 8 encoding signed (5) [ 75] pointer_type byte_size 4 type [ 3e] [ 7b] base_type name "int" byte_size 4 encoding signed (5) [ 82] variable name "to_ptr" decl_file 1 decl_line 1 type [ 75] external declaration [ 8f] variable name "to_ll" decl_file 1 decl_line 2 type [ 6e] external declaration [ 9c] variable name "to_x" decl_file 1 decl_line 3 type [ 7b] external declaration DWARF section '.debug_loc' at offset 0x240: [ 0] 0000000000..0x00000001 [ 0] breg4 4 0x00000001..0x00000003 [ 0] breg4 8 0x00000003..0x0000003e [ 0] breg5 8 [ 2c] 0000000000..0x0000001d [ 0] reg0 [ 3f] 0000000000..0x00000007 [ 0] reg2 [ 1] piece 4 [ 3] reg3 [ 4] piece 4 0x00000007..0x0000002f [ 0] reg3 [ 1] piece 4 [ 3] reg1 [ 4] piece 4 0x0000002f..0x0000003b [ 0] reg3 [ 1] piece 4 [ 3] reg4 [ 4] piece 4 Expected results: From 0..7, {ecx,edx} reg1 reg2 From 7..2f, {ecx,ebx} reg1 reg3 From 2f..3b, {esi,ebx} reg6 reg3 or something like that. Obviously reg4 (%esp) should never be involved. Additional info: Similar problems in 3.4.