[RFC] Clear SEC_HAS_CONTENTS for read-only sections in gcore files?

[RFC] Clear SEC_HAS_CONTENTS for read-only sections in gcore files?

[Mark Kettenis]

The discussion starting with:

   http://sources.redhat.com/ml/gdb/2004-09/msg00202.html

raised the question why we're clearing SEC_LOAD, but don't clear
SEC_HAS_CONTENTS when generating a core file with the gcore command.
This results in core files with zero-filled read-only sections that
just waste space on disk.  I tried the attached patch and it seems to
work fine on i386-unknown-freebsd4.7; no regressions in the testsuite.
Here is some readelf output on the resulting core file:


There are 22 section headers, starting at offset 0xc748c:

Section Headers:
  [Nr] Name          Type            Addr     Off    Size   ES Flg Lk Inf Al
  [ 0]               NULL            00000000 000000 000000 00      0   0  0
  [ 1] note0         PROGBITS        00000000 0002b4 0001c0 00   A  0   0  1
  [ 2] load          NOBITS          08048000 000474 1c3000 00  AX  0   0  1
  [ 3] load          PROGBITS        0820b000 000474 00c000 00  WA  0   0  1
  [ 4] load          PROGBITS        08217000 00c474 01a000 00  WA  0   0  1
  [ 5] load          PROGBITS        08231000 026474 04d000 00 WAX  0   0  1
  [ 6] load          NOBITS          2820b000 073474 012000 00  AX  0   0  1
  [ 7] load          PROGBITS        2821d000 073474 001000 00  WA  0   0  1
  [ 8] load          PROGBITS        2821e000 074474 002000 00  WA  0   0  1
  [ 9] load          PROGBITS        28220000 076474 008000 00 WAX  0   0  1
  [10] load          NOBITS          28228000 07e474 035000 00  AX  0   0  1
  [11] load          PROGBITS        2825d000 07e474 009000 00 WAX  0   0  1
  [12] load          PROGBITS        28266000 087474 003000 00 WAX  0   0  1
  [13] load          NOBITS          28269000 08a474 017000 00  AX  0   0  1
  [14] load          PROGBITS        28280000 08a474 004000 00 WAX  0   0  1
  [15] load          NOBITS          28284000 08e474 005000 00  AX  0   0  1
  [16] load          PROGBITS        28289000 08e474 001000 00 WAX  0   0  1
  [17] load          NOBITS          2828a000 08f474 081000 00  AX  0   0  1
  [18] load          PROGBITS        2830b000 08f474 005000 00 WAX  0   0  1
  [19] load          PROGBITS        28310000 094474 013000 00 WAX  0   0  1
  [20] load          PROGBITS        bfbe0000 0a7474 020000 00 WAX  0   0  1
  [21] .shstrtab     STRTAB          00000000 0c7474 000016 00      0   0  1
Key to Flags:
  W (write), A (alloc), X (execute), M (merge), S (strings)
  I (info), L (link order), G (group), x (unknown)
  O (extra OS processing required) o (OS specific), p (processor specific)

What do people think about this change?

Mark


Index: gcore.c
===================================================================
RCS file: /cvs/src/src/gdb/gcore.c,v
retrieving revision 1.15
diff -u -p -r1.15 gcore.c
--- gcore.c 15 Jun 2004 01:04:19 -0000 1.15
+++ gcore.c 26 Sep 2004 20:45:53 -0000
@@ -354,7 +354,7 @@ gcore_create_callback (CORE_ADDR vaddr, 
 	       || (start >= vaddr && end <= vaddr + size))
 	      && !(bfd_get_file_flags (abfd) & BFD_IN_MEMORY))
 	    {
-	      flags &= ~SEC_LOAD;
+	      flags &= ~(SEC_LOAD | SEC_HAS_CONTENTS);
 	      goto keep;	/* break out of two nested for loops */
 	    }
 	}