[Bug corefiles/11608] New: gcore does not support build-id

[Bug corefiles/11608] New: gcore does not support build-id

[jan dot kratochvil at redhat dot com]

gcore omits whole readonly-executable-code segments.

It should dump their first page so that:
eu-unstrip -n --core=corefile
works as for the kernel-dumped core files.

GDB:
  Type           Offset   VirtAddr           PhysAddr           FileSiz  MemSiz
  Flg Align
  LOAD           0x000b10 0x0000000000400000 0x0000000000000000 0x000000
0x006000 R E 0x1

Linux kernel:
  Type           Offset   VirtAddr           PhysAddr           FileSiz  MemSiz
  Flg Align
  LOAD           0x001000 0x0000000000400000 0x0000000000000000 0x001000
0x006000 R E 0x1000

-- 
           Summary: gcore does not support build-id
           Product: gdb
           Version: unknown
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: corefiles
        AssignedTo: unassigned at sourceware dot org
        ReportedBy: jan dot kratochvil at redhat dot com
                CC: gdb-prs at sourceware dot org
GCC target triplet: x86_64-fedora12-linux-gnu


http://sourceware.org/bugzilla/show_bug.cgi?id=11608

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug corefiles/11608] gcore does not support build-id

[mjw at redhat dot com]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |mjw at redhat dot com


http://sourceware.org/bugzilla/show_bug.cgi?id=11608

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug corefiles/11608] gcore does not support build-id

[ppluzhnikov at google dot com]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ppluzhnikov at google dot
                   |                            |com


http://sourceware.org/bugzilla/show_bug.cgi?id=11608

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug corefiles/11608] gcore does not support build-id

[jan.kratochvil at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=11608

Jan Kratochvil <jan.kratochvil at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2013-11-13
             Blocks|                            |16092
     Ever confirmed|0                           |1

--- Comment #2 from Jan Kratochvil <jan.kratochvil at redhat dot com> ---
At least with Fedora 19 x86_64 /usr/bin/sleep it does not work:

Program Headers:
  Type           Offset             VirtAddr           PhysAddr
                 FileSiz            MemSiz              Flags  Align
  LOAD           0x0000000000000fc4 0x0000000000400000 0x0000000000000000
                 0x0000000000000000 0x0000000000007000  R E    1

Shared libraries also do not have the first page dumped:
/lib64/libc.so.6:
  LOAD           0x0000000000047fc4 0x0000003347200000 0x0000000000000000
                 0x0000000000000000 0x00000000001b6000  R E    1

Although some shared libraries have it dumped:
/lib64/ld-linux-x86-64.so.2
  LOAD           0x0000000000023fc4 0x0000003346e00000 0x0000000000000000
                 0x0000000000021000 0x0000000000021000  R E    1

Bug 16092 is related.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug corefiles/11608] gcore does not support build-id

[jan.kratochvil at redhat dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=11608

Jan Kratochvil <jan.kratochvil at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jan.kratochvil at redhat
                   |                            |dot com
            Version|unknown                     |7.5

--- Comment #1 from Jan Kratochvil <jan.kratochvil at redhat dot com> 2012-08-09 08:41:34 UTC ---
In practice it works now although it depends on something accidentally
happening in Linux kernel:

echo 'const int i[2000]={0};void _start(void){}'|gcc -Wall -nostdlib
-fno-asynchronous-unwind-tables -Wl,--build-id -x c -;gdb -nx ./a.out -ex 'b
*_start' -ex r -ex 'gcore core' -ex 'set confirm no' -ex q;eu-unstrip -n
--core=core
[...]
0x400000+0x400000 2fb48d92cfa19eb24524f14211565853e3da3deb@0x400284 - - [exe]

It works since:
http://sourceware.org/ml/gdb-patches/2012-08/msg00225.html

But it is more accidental:
$ cat /proc/22440/smaps
00400000-00401000 r-xp 00000000 fd:02 15079087 /home/jkratoch/t/a.out
Shared_Dirty:          0 kB
Private_Dirty:         4 kB
Anonymous:             4 kB
Swap:                  0 kB

Despite it is r-x Linux kernel had to write there some data - see Private_Dirty
and Anonymous.  I do not know why, Linux kernel hackers could advice.

gcc (GCC) 4.7.2 20120809 (prerelease)
GNU gdb (GDB) 7.5.50.20120809-cvs
binutils-2.22.52.0.4-8.fc18.x86_64
kernel-3.4.6-1.fc16.x86_64

The right fix would be to see the '[exe]' line even with gdb-7.5 or earlier
(not FSF GDB HEAD where is committed the patch for PR 11804 above).  Earlier
GDBs did not pay attention to the 4 lines in smaps and thus they did not dump
the build-id page so that the '[exe]' line was not visible.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.