Bug 1265271

Summary:	Emacs is not able to be build up under ppc64le because of binutils
Product:	Red Hat Enterprise Linux 7	Reporter:	Petr Hracek <phracek>
Component:	emacs	Assignee:	Petr Hracek <phracek>
Status:	CLOSED DUPLICATE	QA Contact:	BaseOS QE Security Team <qe-baseos-security>
Severity:	urgent	Docs Contact:
Priority:	urgent
Version:	7.2	CC:	amodra, dhorak, jcapik, ksrot, law, mcermak, mhaicman, nickc, ohudlick, than
Target Milestone:	rc	Keywords:	BuildBlocker, Regression
Target Release:	---
Hardware:	ppc64le
OS:	Unspecified
Whiteboard:	blocker
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2015-10-07 12:45:51 UTC	Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:
Bug Depends On:
Bug Blocks:	1223033

Description Petr Hracek 2015-09-22 14:04:38 UTC

Description of problem:
the latest binutils-2.23.52.0.1-54 causes a problem with building emacs.

1) Build fails without any emacs.spec modification.
Finding pointers to doc strings...done
Dumping under the name emacs
emacs: Program segment above .bss in /builddir/build/BUILD/emacs-24.3/build-gtk/src/temacs

2) Build fails with commenting http://git.savannah.gnu.org/cgit/emacs.git/tree/src/unexelf.c?h=emacs-24#n869
Finding pointers to doc strings...done
Dumping under the name emacs
/bin/sh: line 7: 15764 Segmentation fault `/bin/pwd`/temacs --batch --load loadup bootstrap
make[2]: *** [bootstrap-emacs] Error 1

If emacs is build with binutils-2.23.52.0.1-50 and I comment out corresponging message 'Program segment above .bss' then emacs is buildable under ppc64le.

Version-Release number of selected component (if applicable):
binutils-2.23.52.0.1-54

How reproducible:
always

Steps to Reproduce:
1. rhpkg clone emacs
2. rhpkg scratch-build --srpm --arches ppc64le
3. build fails

Actual results:
emacs build fails.

Expected results:
emacs build is successfull

Additional info:
emacs upstream relevant part code is here:
http://git.savannah.gnu.org/cgit/emacs.git/tree/src/unexelf.c?h=emacs-24
and especially
http://git.savannah.gnu.org/cgit/emacs.git/tree/src/unexelf.c?h=emacs-24#n869

It seems that there is a regression in binutils from my point of view,
I have discussed it with secondary arches team.

emacs-24.5 on Fedora is buildable.
http://ppc.koji.fedoraproject.org/koji/taskinfo?taskID=2562249

Comment 4 Petr Hracek 2015-09-23 10:57:14 UTC

I will try to build up emacs-24.5 if the latest version works or not on PPC64LE.

Comment 8 Petr Hracek 2015-09-24 11:08:56 UTC

Only for sure I am adding sump dump information.

I have made some investigation and here are results:

When the dumping failed with message "Program segment above .bss" then I have add a some debug info. Hopefully It helps.
Dumping under the name emacs
Looking for .bss - found .interp
Looking for .bss - found .note.ABI-tag
Looking for .bss - found .note.gnu.build-id
Looking for .bss - found .gnu.hash
Looking for .bss - found .dynsym
Looking for .bss - found .dynstr
Looking for .bss - found .gnu.version
Looking for .bss - found .gnu.version_r
Looking for .bss - found .rela.toc
Looking for .bss - found .rela.plt
Looking for .bss - found .init
Looking for .bss - found .text
Looking for .bss - found .fini
Looking for .bss - found .rodata
Looking for .bss - found .eh_frame_hdr
Looking for .bss - found .eh_frame
Looking for .bss - found .data.rel.ro
Looking for .bss - found .dynamic
Looking for .bss - found .got
Looking for .bss - found .plt
Looking for .bss - found .data
Looking for .bss - found .bss
Looking for .sbss - found .interp
Looking for .sbss - found .note.ABI-tag
Looking for .sbss - found .note.gnu.build-id
Looking for .sbss - found .gnu.hash
Looking for .sbss - found .dynsym
Looking for .sbss - found .dynstr
Looking for .sbss - found .gnu.version
Looking for .sbss - found .gnu.version_r
Looking for .sbss - found .rela.toc
Looking for .sbss - found .rela.plt
Looking for .sbss - found .init
Looking for .sbss - found .text
Looking for .sbss - found .fini
Looking for .sbss - found .rodata
Looking for .sbss - found .eh_frame_hdr
Looking for .sbss - found .eh_frame
Looking for .sbss - found .data.rel.ro
Looking for .sbss - found .dynamic
Looking for .sbss - found .got
Looking for .sbss - found .plt
Looking for .sbss - found .data
Looking for .sbss - found .bss
Looking for .sbss - found .comment
Looking for .sbss - found .debug_aranges
Looking for .sbss - found .debug_info
Looking for .sbss - found .debug_abbrev
Looking for .sbss - found .debug_line
Looking for .sbss - found .debug_str
Looking for .sbss - found .debug_loc
Looking for .sbss - found .debug_ranges
Looking for .sbss - found .shstrtab
Looking for .sbss - found .symtab
Looking for .sbss - found .strtab
Looking for .plt - found .interp
Looking for .plt - found .note.ABI-tag
Looking for .plt - found .note.gnu.build-id
Looking for .plt - found .gnu.hash
Looking for .plt - found .dynsym
Looking for .plt - found .dynstr
Looking for .plt - found .gnu.version
Looking for .plt - found .gnu.version_r
Looking for .plt - found .rela.toc
Looking for .plt - found .rela.plt
Looking for .plt - found .init
Looking for .plt - found .text
Looking for .plt - found .fini
Looking for .plt - found .rodata
Looking for .plt - found .eh_frame_hdr
Looking for .plt - found .eh_frame
Looking for .plt - found .data.rel.ro
Looking for .plt - found .dynamic
Looking for .plt - found .got
Looking for .plt - found .plt
Looking for .data - found .interp
Looking for .data - found .note.ABI-tag
Looking for .data - found .note.gnu.build-id
Looking for .data - found .gnu.hash
Looking for .data - found .dynsym
Looking for .data - found .dynstr
Looking for .data - found .gnu.version
Looking for .data - found .gnu.version_r
Looking for .data - found .rela.toc
Looking for .data - found .rela.plt
Looking for .data - found .init
Looking for .data - found .text
Looking for .data - found .fini
Looking for .data - found .rodata
Looking for .data - found .eh_frame_hdr
Looking for .data - found .eh_frame
Looking for .data - found .data.rel.ro
Looking for .data - found .dynamic
Looking for .data - found .got
Looking for .data - found .plt
Looking for .data - found .data
old_bss_index 22
old_bss_addr 102b0000
old_bss_size 89798
old_bss_offset 2a0000
new_bss_addr 115d0000
new_data2_addr 102b0000
new_data2_size 1320000
new_data2_offset 2a0000
new_data2_incr 1320000
Old section offset da1ac8
Old section count 34
New section offset 20c1ac8
New section count 35
new_file_h->e_phnum 9
old_sbss_index ffffffff
old_bss_addr 102b0000
alignment 10
NEW_PROGRAM_H (n).p_vaddr 102af638
NEW_PROGRAM_H (n).p_filesz 9c8
(OLD_SECTION_H (old_bss_index)).sh_addralign 10
old_sbss_index ffffffff
old_bss_addr 102b0000
alignment 10
NEW_PROGRAM_H (n).p_vaddr 0
NEW_PROGRAM_H (n).p_filesz 0
(OLD_SECTION_H (old_bss_index)).sh_addralign 10
old_sbss_index ffffffff
old_bss_addr 102b0000
alignment 10
NEW_PROGRAM_H (n).p_vaddr 10269510
NEW_PROGRAM_H (n).p_filesz 73cc
(OLD_SECTION_H (old_bss_index)).sh_addralign 10
old_sbss_index ffffffff
old_bss_addr 102b0000
alignment 10
NEW_PROGRAM_H (n).p_vaddr 1000024c
NEW_PROGRAM_H (n).p_filesz 44
(OLD_SECTION_H (old_bss_index)).sh_addralign 10
old_sbss_index ffffffff
old_bss_addr 102b0000
alignment 10
NEW_PROGRAM_H (n).p_vaddr 102afa68
NEW_PROGRAM_H (n).p_filesz 440
(OLD_SECTION_H (old_bss_index)).sh_addralign 10
old_sbss_index ffffffff
old_bss_addr 102b0000
alignment 10000
NEW_PROGRAM_H (n).p_vaddr 102af638
NEW_PROGRAM_H (n).p_filesz 2c9756
(OLD_SECTION_H (old_bss_index)).sh_addralign 10
emacs: Program segment above .bss in /home/phracek/rpmbuild/BUILD/emacs-24.3/build-gtk/src/temacs


Some useful information taken by command readelf:
readelf -h ./temacs
ELF Header:
  Magic:   7f 45 4c 46 02 01 01 00 00 00 00 00 00 00 00 00
  Class:                             ELF64
  Data:                              2's complement, little endian
  Version:                           1 (current)
  OS/ABI:                            UNIX - System V
  ABI Version:                       0
  Type:                              EXEC (Executable file)
  Machine:                           PowerPC64
  Version:                           0x1
  Entry point address:               0x100184f0
  Start of program headers:          64 (bytes into file)
  Start of section headers:          14289272 (bytes into file)
  Flags:                             0x2, abiv2
  Size of this header:               64 (bytes)
  Size of program headers:           56 (bytes)
  Number of program headers:         9
  Size of section headers:           64 (bytes)
  Number of section headers:         34
  Section header string table index: 31

readelf -l ./temacs
Elf file type is EXEC (Executable file)
Entry point 0x100184f0
There are 9 program headers, starting at offset 64

Program Headers:
  Type           Offset             VirtAddr           PhysAddr
                 FileSiz            MemSiz              Flags  Align
  PHDR           0x0000000000000040 0x0000000010000040 0x0000000010000040
                 0x00000000000001f8 0x00000000000001f8  R E    8
  INTERP         0x0000000000000238 0x0000000010000238 0x0000000010000238
                 0x0000000000000011 0x0000000000000011  R      1
      [Requesting program interpreter: /lib64/ld64.so.2]
  LOAD           0x0000000000000000 0x0000000010000000 0x0000000010000000
                 0x000000000029e6d8 0x000000000029e6d8  R E    10000
  LOAD           0x000000000029f640 0x00000000102af640 0x00000000102af640
                 0x00000000002c974e 0x0000000000350e90  RW     10000
  DYNAMIC        0x000000000029fa70 0x00000000102afa70 0x00000000102afa70
                 0x0000000000000440 0x0000000000000440  RW     8
  NOTE           0x000000000000024c 0x000000001000024c 0x000000001000024c
                 0x0000000000000044 0x0000000000000044  R      4
  GNU_EH_FRAME   0x0000000000268fe0 0x0000000010268fe0 0x0000000010268fe0
                 0x00000000000073cc 0x00000000000073cc  R      4
  GNU_STACK      0x0000000000000000 0x0000000000000000 0x0000000000000000
                 0x0000000000000000 0x0000000000000000  RW     10
  GNU_RELRO      0x000000000029f640 0x00000000102af640 0x00000000102af640
                 0x00000000000009c0 0x00000000000009c0  R      1

 Section to Segment mapping:
  Segment Sections...
   00
   01     .interp
   02     .interp .note.ABI-tag .note.gnu.build-id .gnu.hash .dynsym .dynstr .gnu.version .gnu.version_r .rela.toc .rela.plt .init .text .fini .rodata .eh_frame_hdr .eh_frame
   03     .data.rel.ro .dynamic .got .plt .data .bss
   04     .dynamic
   05     .note.ABI-tag .note.gnu.build-id
   06     .eh_frame_hdr
   07
   08     .data.rel.ro .dynamic .got

readelf -S ./temacs
There are 34 section headers, starting at offset 0xda0978:

Section Headers:
  [Nr] Name              Type             Address           Offset
       Size              EntSize          Flags  Link  Info  Align
  [ 0]                   NULL             0000000000000000  00000000
       0000000000000000  0000000000000000           0     0     0
  [ 1] .interp           PROGBITS         0000000010000238  00000238
       0000000000000011  0000000000000000   A       0     0     1
  [ 2] .note.ABI-tag     NOTE             000000001000024c  0000024c
       0000000000000020  0000000000000000   A       0     0     4
  [ 3] .note.gnu.build-i NOTE             000000001000026c  0000026c
       0000000000000024  0000000000000000   A       0     0     4
  [ 4] .gnu.hash         GNU_HASH         0000000010000290  00000290
       00000000000000dc  0000000000000000   A       5     0     8
  [ 5] .dynsym           DYNSYM           0000000010000370  00000370
       00000000000063f0  0000000000000018   A       6     1     8
  [ 6] .dynstr           STRTAB           0000000010006760  00006760
       0000000000004ae9  0000000000000000   A       0     0     1
  [ 7] .gnu.version      VERSYM           000000001000b24a  0000b24a
       0000000000000854  0000000000000002   A       5     0     2
  [ 8] .gnu.version_r    VERNEED          000000001000baa0  0000baa0
       0000000000000160  0000000000000000   A       6    10     8
  [ 9] .rela.toc         RELA             000000001000bc00  0000bc00
       0000000000000330  0000000000000018   A       5     0     8
  [10] .rela.plt         RELA             000000001000bf30  0000bf30
       00000000000060d8  0000000000000018   A       5    20     8
  [11] .init             PROGBITS         0000000010012020  00012020
       000000000000004c  0000000000000000  AX       0     0     32
  [12] .text             PROGBITS         0000000010012080  00012080
       0000000000237140  0000000000000000  AX       0     0     32
  [13] .fini             PROGBITS         00000000102491c0  002491c0
       0000000000000024  0000000000000000  AX       0     0     4
  [14] .rodata           PROGBITS         00000000102491e8  002491e8
       000000000001fdf8  0000000000000000   A       0     0     8
  [15] .eh_frame_hdr     PROGBITS         0000000010268fe0  00268fe0
       00000000000073cc  0000000000000000   A       0     0     4
  [16] .eh_frame         PROGBITS         00000000102703b0  002703b0
       000000000002e328  0000000000000000   A       0     0     8
  [17] .data.rel.ro      PROGBITS         00000000102af640  0029f640
       0000000000000430  0000000000000000  WA       0     0     8
  [18] .dynamic          DYNAMIC          00000000102afa70  0029fa70
       0000000000000440  0000000000000010  WA       6     0     8
  [19] .got              PROGBITS         00000000102afeb0  0029feb0
       0000000000000150  0000000000000008  WA       0     0     8
  [20] .plt              NOBITS           00000000102b0000  002a0000
       0000000000002058  0000000000000008  WA       0     0     8
  [21] .data             PROGBITS         00000000102b2058  002a2058
       00000000002c6d36  0000000000000000  WA       0     0     8
  [22] .bss              NOBITS           0000000010578d90  00568d8e
       0000000000087740  0000000000000000  WA       0     0     16
  [23] .comment          PROGBITS         0000000000000000  00568d8e
       0000000000000058  0000000000000001  MS       0     0     1
  [24] .debug_aranges    PROGBITS         0000000000000000  00568de6
       0000000000001350  0000000000000000           0     0     1
  [25] .debug_info       PROGBITS         0000000000000000  0056a136
       00000000003aeadb  0000000000000000           0     0     1
  [26] .debug_abbrev     PROGBITS         0000000000000000  00918c11
       0000000000021b05  0000000000000000           0     0     1
  [27] .debug_line       PROGBITS         0000000000000000  0093a716
       000000000007d970  0000000000000000           0     0     1
  [28] .debug_str        PROGBITS         0000000000000000  009b8086
       000000000004911a  0000000000000001  MS       0     0     1
  [29] .debug_loc        PROGBITS         0000000000000000  00a011a0
       00000000003277c9  0000000000000000           0     0     1
  [30] .debug_ranges     PROGBITS         0000000000000000  00d28969
       0000000000077ec0  0000000000000000           0     0     1
  [31] .shstrtab         STRTAB           0000000000000000  00da0829
       0000000000000148  0000000000000000           0     0     1
  [32] .symtab           SYMTAB           0000000000000000  00da11f8
       0000000000037bc0  0000000000000018          33   5415     8
  [33] .strtab           STRTAB           0000000000000000  00dd8db8
       000000000002e23b  0000000000000000           0     0     1
Key to Flags:
  W (write), A (alloc), X (execute), M (merge), S (strings)
  I (info), L (link order), G (group), T (TLS), E (exclude), x (unknown)
  O (extra OS processing required) o (OS specific), p (processor specific)

Comment 11 Jaromír Cápík 2015-10-06 19:48:36 UTC

Hello guys.

I'm affraid it didn't work correctly even with the old binutils. The whole unexec logic is broken and needs a lot of love.

The relocation undoing is done the following way on x86_64:

.rela.plt -> .plt
.rela.bss -> .bss
.rela.got -> .got


... but on ppc64le I get the following with the old binutils:

.rela.plt -> .data
.rela.toc -> empty string (probably some zeros in memory)


... and with the new binutils it probably just attacks a different part of memory and segfaults.

Comment 12 Jaromír Cápík 2015-10-06 20:42:37 UTC

On ppc64be I also get wrong results ...


with binutils-2.23.52.0.1-50.el7.ppc64

.rela.plt -> .data
.rela.toc -> empty string
.rela.data.rel.ro -> .data.rel.ro


with binutils-2.23.52.0.1-54.el7.ppc64

.rela.plt -> .data
.rela.toc -> segfault

Comment 13 Jaromír Cápík 2015-10-06 21:05:23 UTC

Fedora 22 is affected too ...



with binutils-2.25-8.fc22.ppc64le

.rela.plt -> .data
.rela.toc -> empty string
.rela.data.rel.ro -> .data.rel.ro

Comment 14 Jaromír Cápík 2015-10-06 21:37:18 UTC

Fedora 22 for s390x seems ok ...

.rela.plt -> .plt
.rela.got -> .got


Fedora 22 for aarch64 seems ok too ..

.rela.plt -> .plt
.rela.got -> .got

Comment 15 Jaromír Cápík 2015-10-07 12:03:08 UTC

After enforcing old_plt_index=-1, the plt relocation undoing got fixed.

Comment 16 Jaromír Cápík 2015-10-07 12:26:17 UTC

.rela.plt -> .plt
.rela.toc -> empty string

the good thing is that it doesn't crash with the latest binutils

Comment 19 Alan Modra 2015-10-27 04:16:59 UTC

Why is bug #1223033 private?