stage1/xgcc -Bstage1/ -B/home/dave/opt/gnu/gcc/gcc-4.1.0/hppa-linux/bin/ -c -g - O2 -gnatpg -gnata -I- -I. -Iada -I../../gcc/gcc/ada ../../gcc/gcc/ada/ada.a ds -o ada/ada.o +===========================GNAT BUG DETECTED==============================+ | 4.1.0 20050410 (experimental) (hppa-unknown-linux-gnu) Segmentation fault| | Error detected at ada.ads:17:1 | | Please submit a bug report; see http://gcc.gnu.org/bugs.html. | | Use a subject line meaningful to you and us to track the bug. | | Include the entire contents of this bug box in the report. | | Include the exact gcc or gnatmake command that you entered. | | Also include sources listed below in gnatchop format | | (concatenated together with no headers between files). | +==========================================================================+ This is the backtrace: Program received signal SIGSEGV, Segmentation fault. 0x008bdd88 in in_unlikely_text_section () at ../../gcc/gcc/varasm.c:309 309 ret_val = ((in_section == in_unlikely_executed_text) (gdb) bt #0 0x008bdd88 in in_unlikely_text_section () at ../../gcc/gcc/varasm.c:309 #1 0x008be220 in named_section_real (name=0xaa8900 ".debug_info", flags=1024, decl=0x0) at ../../gcc/gcc/varasm.c:444 #2 0x0060a640 in output_comp_unit (die=0x4000b150, output_if_empty=0) at ../../gcc/gcc/dwarf2out.c:7108 #3 0x0062229c in dwarf2out_finish ( filename=0xbff0007c "../../gcc/gcc/ada/ada.ads") at ../../gcc/gcc/dwarf2out.c:13916 #4 0x00880d78 in compile_file () at ../../gcc/gcc/toplev.c:1038 #5 0x00883c1c in do_compile () at ../../gcc/gcc/toplev.c:2120 #6 0x00883cc0 in toplev_main (argc=15, argv=0xbff005d4) at ../../gcc/gcc/toplev.c:2152 #7 0x003ff3dc in main (argc=15, argv=0xbff005d4) at ../../gcc/gcc/main.c:35 0x008bdd84 <in_unlikely_text_section+160>: ldw c(,r3),r19 0x008bdd88 <in_unlikely_text_section+164>: ldw ec(,r19),r19 0x008bdd8c <in_unlikely_text_section+168>: cmpib,=,n 0,r19,0x8bddbc < in_unlikely_text_section+216> (gdb) p/x $r19 $4 = 0x0 (gdb) p cfun $5 = (struct function *) 0x0 (gdb) p in_section $6 = in_named It looks as if we need a check on cfun: 309 ret_val = ((in_section == in_unlikely_executed_text) 310 || (in_section == in_named 311 && cfun->unlikely_text_section_name 312 && strcmp (in_named_name, 313 cfun->unlikely_text_section_name) == 0));
This might also be the failure which Diego is seeing with libjava building: http://gcc.gnu.org/ml/gcc-patches/2005-04/msg01050.html
Subject: Re: [4.1 Regression] Segmentation fault in gnat1 > This might also be the failure which Diego is seeing with libjava building: > http://gcc.gnu.org/ml/gcc-patches/2005-04/msg01050.html I think this started yesterday. In any event, a build yesterday evening failed the same way. Dave
Created attachment 8582 [details] Hot/Cold patch Patch by David which was attached to the wrong bug.
This patch fixes bootstrap failure on i686-pc-linux-gnu.
*** Bug 20942 has been marked as a duplicate of this bug. ***
Created attachment 8594 [details] patch which Caroline sent me to attach "I mistakenly assumed that if current_function_decl was defined, then the decl would always have a function struct attached to it; also that when we were writing out debug information, current_function_decl would be defined."
Created attachment 8595 [details] new patch which Caroline sent me to attach Apparently there were two typo's in the patch I sent to you, both in dwarf2out.c: dwarf2out.c line 6816: 'internal_eror' should be 'internal_error' dwarf2out.c line 7671: 'cfun = = NULL' should be 'cfun = NULL' Attached is a patch with these two lines corrected.
Fixed by reverting the patch which caused this.