Commit Graph

2060 Commits

Author SHA1 Message Date
Alexey Tourbin
a1a075caa2 rpmio: fixed Fwrite() truncated return value
This is the reason why files >= 2G could not be added to cpio, even on
64-bit architectures.  Fwrite was called with a buffer >= 2G, but the
return value was truncated to int (this narrowing conversion is probably
undefined behavior already; what happens next, though, is that after
casting to size_t it assumes a very big numeric value of >= 2^64-2^32).
So rpmbuild was failing with an unrelated errno, "cpio: write failed -
Bad file descriptor".

There are other reasons why this still won't work on 32-bit architectures.
For one, rpmbuild mmaps a file before feeding it to cpio, and 32-bit
architectures have a very limited ability, if any, to mmap 2G+ files.
2018-06-26 12:42:44 +03:00
Ivan Zakharyaschev
c58b3a24d0 4.0.4-alt101.M80P.1 - Merged rpm-build-4.0.4-alt102 into p8 (everything except for separating rpm & -build):
[4.0.4-alt102]
  + verify-elf: fixed passing of LD_PRELOAD in verify_unresolved.
  [4.0.4-alt101]
  + verify-elf: made verify_lfs check shared objects too.
  + Relaxed check for gcc package name in {cpp,pkgconfiglib}.req generators.
  [4.0.4-alt100.99]
  + Updated generator of non-LFS functions list.
  [4.0.4-alt100.97]
  + fixup-libraries: fixed recognition of PIEs (ldv@).
  + verify-elf: treat PIEs as executables in the check for unresolved symbols (ldv@).
2017-01-17 16:16:56 +03:00
cd09d0527e [rpm-build 4.0.4-alt102] cherry-picked (skipped the separation of rpm & -build)
- verify-elf: fixed passing of LD_PRELOAD in verify_unresolved.
2017-01-17 16:16:31 +03:00
88c1020672 verify-elf: fix passing of LD_PRELOAD in verify_unresolved
In verify_unresolved(), do not preload requested objects to $elf_ldd
directly because the latter is not an ELF interpreter but a shell
script.  Instead, pass the list of objects that have to be preloaded
to $elf_ldd script which in turn assigns it to LD_PRELOAD variable
in the ELF interpreter invocation.
2017-01-17 16:16:28 +03:00
9515fd6545 rpmio: fix potential buffer overflow in Fdopen
In file included from /usr/include/string.h:648:0,
                 from ../system.h:76,
                 from rpmio.c:6:
In function 'strncat',
    inlined from 'Fdopen' at rpmio.c:3473:5:
/usr/include/bits/string3.h:158:10: warning: call to __builtin___strncat_chk might overflow destination buffer
   return __builtin___strncat_chk (__dest, __src, __len, __bos (__dest));
2017-01-17 16:16:28 +03:00
Gleb Fotengauer-Malinovskiy
8605fa7393 [rpm-build 4.0.4-alt101] cherry-picked (skipped the separation of rpm & -build)
- verify-elf: made verify_lfs check shared objects too.
- Relaxed check for gcc package name in {cpp,pkgconfiglib}.req generators.
- [not for p8] rpm-build: add R: rpmspec.
2017-01-17 16:16:03 +03:00
Gleb Fotengauer-Malinovskiy
fc627b0ef0 Relax check for gcc package name in {cpp,pkgconfiglib}.req generators 2017-01-17 16:11:35 +03:00
Gleb Fotengauer-Malinovskiy
bea326ff93 verify-elf: make verify_lfs check libraries too
Previously, verify_lfs checked if object has valid interpreter, but it
doesn't make sense in case of shared objects.  verify_lfs check is valid
for all ELFs linked with libraries supporting both LFS and non-LFS API
(libc.so and libz.so in our case).
2017-01-17 16:11:27 +03:00
Gleb Fotengauer-Malinovskiy
e6eb33cf4b [rpm-build 4.0.4-alt100.99] cherry-picked (skipped the separation of rpm & -build)
- Fixed non-LFS check in verify-elf.
- [not for p8] Dropped fake provides made for rpm 4.13.0 bootstrap.
2017-01-17 16:10:06 +03:00
Gleb Fotengauer-Malinovskiy
1b41cd0741 Fix generation of verify-elf-non-lfs-funcs.list
If there were both `func' and `func'64, `func' was considered non-LFS.
Unfortunately, that approach missed functions like fts64_* and
readdir64_r.
2017-01-17 16:07:21 +03:00
Ivan Zakharyaschev
22eb7f0c27 4.0.4-alt100.96.M80P.1 - Build for p8. - verify-elf: correctly handle grep's status
...in practically impossible cases.
2016-12-09 03:27:48 +03:00
Ivan Zakharyaschev
f415265c95 Merge branch 'alt', remote-tracking branch 'ldv@ALT/maint' into altM/80P 2016-12-09 03:27:26 +03:00
Ivan Zakharyaschev
2ac0536ff6 verify_rpath(): distinguish grep's failure and a normal valid case
Previously (according to the exit statuses):

0. if something strange was found in $rpath, an error was reported;
1. if nothing strange was found, no error was reported;
2. if grep failed, it looked like the previous case and no error was
reported (as if $rpath was valid).

Now, the pipe (and the whole script) would abort on grep's failure.
2016-12-09 03:23:12 +03:00
Ivan Zakharyaschev
0445f481cc verify_stack(): distinguish failures and a normal status of grep
Now, if something fails, the pipe (and the function, and the whole
script) will fail.

If the exit codes are normal, they are handled according to our logic.

Previously, a failure (which is an exceptional rare case) would lead to:

1. skipping the further exe_stack check and error_strict STACK "$f" 'STACK entry not found'
2. silently skipping the exe_stack check (which was impossible anyway because of the failure)
2016-12-09 03:23:12 +03:00
Ivan Zakharyaschev
fc56b0be97 verify_unresolved(): correctly treat the status of grep in a very unusual case
The pipe (and the function, and the whole script) would fail when the
resulting constructed rpath was empty. But usually at least
$RPM_VERIFY_ELF_LDD_RPATH is not empty, so the bad behavior has not
been observed.
2016-12-09 03:23:12 +03:00
0e2f86b0fb 4.0.4-alt100.97
- fixup-libraries: fixed recognition of PIEs.
- verify-elf: treat PIEs as executables in the check for unresolved symbols.
2016-12-07 11:30:44 +00:00
a4fb4d42e7 verify-elf (verify_unresolved): treat PIEs as executables 2016-12-07 11:30:44 +00:00
8c1f9d589b Move "is this ELF shared object an executable" check to a separate script
* configure.ac (AC_OUTPUT): Add scripts/fixup-libraries.
* rpm-4_0.spec (%files build): Add %_rpmlibdir/is_elf_so_executable.
* scripts/is_elf_so_executable: New file.
* scripts/Makefile.am (EXTRA_DIST, config_SCRIPTS): Add it.
* scripts/fixup-libraries: Rename to scripts/fixup-libraries.in,
use is_elf_so_executable.
2016-12-07 11:30:42 +00:00
a51ea574e9 scripts/fixup-libraries: fix recognition of PIEs 2016-12-07 11:30:25 +00:00
Ivan Zakharyaschev
d9f7deba30 4.0.4-alt100.95.M80P.1 - Build for p8. 2016-12-01 13:40:39 +03:00
Ivan Zakharyaschev
401d9520ba 4.0.4-alt100.96
- verify-elf: don't confuse the initial verify_rpath() in case
  of two RUNPATH/RPATHs (ALT#32826).
2016-11-30 17:50:04 +03:00
Ivan Zakharyaschev
d362bae600 verify-elf made a bit safer: pipes or command substitutions make it fail
1. set -o pipefail
2. don't use command substitutions inside other complex expressions; just assign to a var -- then it would make the script fail
2016-11-30 17:49:59 +03:00
Ivan Zakharyaschev
00b00565ea verify-elf: don't call verify_rpath() under a pipe because the errors are saved in a var
(A pipe makes a subshell and the var is modified there.)
2016-11-30 17:48:24 +03:00
Ivan Zakharyaschev
c0f11587a5 verify-elf: don't confuse verify_rpath() in case of two RUNPATH/RPATHs (ALT#32826) 2016-11-30 17:30:49 +03:00
Ivan Zakharyaschev
91c923f251 verify-elf: tiny cleanup: unify the awk expression for RPATH extraction
Rewritten equivalently, like in lib.req.
2016-11-30 17:09:09 +03:00
Ivan Zakharyaschev
70369c1a8e verify-elf: tiny cleanup: local var declaration moved to the place of use
The other occurence of "rpath" is in the second part of a pipe and is
local to that subshell anyway.
2016-11-30 17:08:09 +03:00
Ivan Zakharyaschev
00635cb1ac 4.0.4-alt100.95
- verify-elf: honor RUNPATH in addition to RPATH, too (like in lib.req).
- (unnoticeable) shell.req: generalize the shebang regexp w.r.t. other
  hardcoded locations of /usr/bin/env. (No need to be too strict here:
  shebang.req should catch bad locations. It's not our job.)
- %distribution: ALT Linux --> ALT (ALT#32707).
2016-11-29 15:12:34 +03:00
Ivan Zakharyaschev
3172abb8c1 %%distribution: ALT Linux --> ALT (ALT#32707) and some minor corrections of old @altlinux.ru addresses 2016-11-29 15:12:34 +03:00
Ivan Zakharyaschev
66f1a4cb8d verify-elf: RUNPATH overrides RPATH for verify_unresolved
From ld.so's documentation about the search order:

Using the directories specified in the DT_RPATH dynamic section
attribute of the binary if present and DT_RUNPATH attribute does not
exist. Use of DT_RPATH is deprecated.
2016-11-29 15:11:46 +03:00
Ivan Zakharyaschev
65509e2763 verify-elf: honor RUNPATH, too (like in lib.req.in:95 and debuginfo.req.in:76)
in addition to RPATH
2016-11-28 22:37:18 +03:00
Ivan Zakharyaschev
e2810bda5c added me to "Copyright (C)" 2016-11-28 22:37:13 +03:00
Ivan Zakharyaschev
d2e5d7dce7 shell.req: generalize the shebang regexp w.r.t. other locations of env
No need to be too strict here: shebang.req would (or should) catch bad
locations. It's not our work.
2016-05-27 16:02:37 +03:00
Ivan Zakharyaschev
ef66de16d5 shell.req: more idiomatic regexp for reading the shebang (use ?)
Compared with a similar regexp in Emacs sources --
http://git.altlinux.org/gears/e/emacs24.git?p=emacs24.git;a=blob;f=lisp/progmodes/sh-script.el;h=4700324d376db79e27d05e5352fff2f7e841a603;hb=HEAD#l1637
: "#![ \t]?\\([^ \t\n]*/bin/env[ \t]\\)?\\([^ \t\n]+\\)"

Conclusion: mostly identical meaning, with the exception of a more
liberal path of env in Emacs.

But ALT Sisyphus has /usr/bin/env only, so I don't see a good reason
to make the regexp more general. (If a different path is written
there, it won't work in ALT Sisyphus.)
2016-05-27 15:19:10 +03:00
Ivan Zakharyaschev
06f844cd6f 4.0.4-alt100.94
- shell.req: use the version of (ba|)sh from the shebang.
- find-{requires,provides}: run all scripts even for empty lists of files.
2016-05-24 23:08:16 +03:00
Ivan Zakharyaschev
f9bf39c2aa shell.req: use the version of (ba|)sh from the shebang 2016-05-24 23:08:04 +03:00
Ivan Zakharyaschev
d13f11fe4f find-provides: removed optimization for empty lists of files
The same reasons as for find-requires:

1. can be buggy in future;
2. give a chance to the methods to generate some special Provides
2016-04-07 13:58:27 +03:00
Ivan Zakharyaschev
0ebfa8f191 find-requires: removed optimization for empty lists of files
Reasons:

1. code clarity (and more robust w.r.t. to future changes which might
lead to bugs, as before)

2. to give all the special methods a chance to print something in the
case of an empty list, too (when needed)
2016-04-06 17:09:38 +03:00
Ivan Zakharyaschev
edf305d42c 4.0.4-alt100.93
- find-requires fixed for packages consisting completely of symlinks.
2016-04-06 03:30:21 +03:00
Ivan Zakharyaschev
4d23ad4702 percolate: removed loop optimization (for code clarity), result won't change 2016-04-06 03:26:17 +03:00
Ivan Zakharyaschev
95ffed16a6 find-requires: it was wrong to check "$tmpdir"/files for emptiness
Since there appeared a special handling for just some kind of
symlinks, "$tmpdir"/files gave only a part of "$tmpdir"/files+types --
the final list of what to analyze.
2016-04-06 03:22:49 +03:00
Ivan Zakharyaschev
e6e202e522 4.0.4-alt100.92
- find-requires: treat symlinks leading to some real content all way
  through the same package as that content.
- find-{requires,provides}: in case of an error in it, the pipe to it
  will be broken (must not change the visible behavior of rpm-build).
2016-04-05 23:26:03 +03:00
Ivan Zakharyaschev
6b43518d50 find-{requires,provides}: switched to the equivalent use of tmpdir.sh 2016-04-05 23:24:45 +03:00
Ivan Zakharyaschev
816c140ae7 find-{requires,provides}: in case of an error in it, the pipe to it will be broken (must not change the visible behavior of rpm-build)
We move "cat > /dev/null" to the only place where consuming the whole
input is useful for there to be no error.

This is done in preparation for a switch to the usage of tmpdir.sh.
2016-04-05 23:22:24 +03:00
Ivan Zakharyaschev
5630438416 debuginfo.req.files: ignore everything under /usr/lib/debug/.build-id/
The special directory structure there doesn't correspond to the
directory structure where debuginfo.req expects to find and process
*.debug files.
2016-04-05 22:57:42 +03:00
Ivan Zakharyaschev
7f43b629e2 find-requires: a TODO notice on directory symlinks vs percolation 2016-04-05 22:57:42 +03:00
Ivan Zakharyaschev
6412a7f161 find-requires: treat symlinks leading to some real content inside the same package as that content 2016-04-05 22:57:42 +03:00
Ivan Zakharyaschev
dc0def4a3b percolate: hopefully, it will work just with /bin/sh (not to create an extra dep)
Conflicts:
	scripts/percolate
2016-04-05 22:57:42 +03:00
Ivan Zakharyaschev
0f969acf44 percolate: register it in Makefile etc. 2016-04-05 22:57:42 +03:00
Ivan Zakharyaschev
70bf25fc08 percolate: a helper script (will be used to follow symlinks only through known files) 2016-04-05 22:57:35 +03:00
Ivan Zakharyaschev
484d6f523f 4.0.4-alt100.91
- brp-verify_elf: handle all rules for setting LD_PRELOAD in the form
  of pairs of RPM_LD_PRELOAD_@ and RPM_FILES_TO_LD_PRELOAD_@ env vars.
- verify-elf: substitute LD_PRELOAD with "$VERIFY_ELF_LD_PRELOAD"
  before ldd (i.e., any prior value of LD_PRELOAD will be cleared!)
- examples of usage of the above facilities:
  rpm-build-python3-0.1.9.3 and rpm-build-python.
2016-03-31 21:23:29 +03:00