IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
E2K arch have old elfutils-0.159 which doesn't have eu-elfcompress.
Error:
/usr/lib/rpm/process-debuginfo: line 68: eu-elfcompress: command not found
Reported-by: Andrey Savchenko <bircoph@altlinux.org>
`vmlinux' on ppc appears to have `Dynamic Section', this triggers its
processing with `/usr/lib/rpm/ldd' which returns error:
ldd: ERROR: /usr/src/tmp/kernel-image-std-test-buildroot/usr/lib/debug/lib/modules/5.4.49-std-test-alt4/vmlinux: failed to find the program interpreter
find-requires: ERROR: /usr/lib/rpm/debuginfo.req failed
error: /bin/sh failed
error: Failed to find Requires
Reported-by: Anton V. Boyarshinov <boyarsh@altlinux.org>
Do not try to process ELF relocatable objects that are not kernel
modules, this fixes the following regression:
/usr/lib/rpm/debugedit: ./usr/lib64/gcc/x86_64-alt-linux/9/crtfastmath.o: DWARF version 0 unhandled
objdump: ./usr/lib64/gcc/x86_64-alt-linux/9/crtfastmath.o: bad value
verify-elf: ERROR: ./usr/lib64/gcc/x86_64-alt-linux/9/crtfastmath.o: objdump failed
Fixes: 7aa048df38 ("Generate debuginfo for kernel packages")
Allow `crash` (gdb-7.6 based) to work out of the box.
Otherwise it will fail with the error:
Dwarf Error: wrong version in compilation unit header (is 0, should be 2, 3, or 4)
Alas, it will not be able to load debuginfo for the ko modules, but, advanced
user can `eu-elfcompress -t none` them manually.
We can't keep modules uncompressed due to `cpio archive too big - 4136M'
RPM build error.
Avoid `DWARF version 0 unhandled' for compressed ELFs in
find-debuginfo-files.
Based on 6e9fd97f6 ("debugedit: Add -n, --no-recompute-build-id.")
by Mark Wielaard <mark@klomp.org>.
- Support uncompressed (std-def) and compressed (un-def) modules. Also,
support xz(1) compression if we'd switch to it.
- If module is uncompressed it stays in the kernel-image rpm, if it's
compressed then uncompressed module goes into /usr/lib/debug tree and
will be in debuginfo rpm. Compressed modules will be re-compressed after
stripping.
- vmlinuz binary triggers extraction of vmlinux from the kernel-source
tree, and it's then placed into /usr/lib/debug tree, then processed and
stripped as usual.
- Logic of processing of vmlinux binary for domU packages is not changed.
Some flavours do not even pack it. Thus, do not remove your
`%brp_strip_none'.
- Output of file(1) for vmlinuz is only parsed for filename, because, on
different architectures it detects completely different things.
- Debug files are compressed with `eu-elfcompress'.
- You still need to enable `CONFIG_DEBUG_INFO=y' in for `.config'.
No other changes to spec should be needed.
Thanks to Nikita Ermakov for starting this work.
glibc starting with commit glibc-2.30~85
(elf: Refuse to dlopen PIE objects [BZ #24323])
no longer supports preloading of PIE objects.
Workaround this by swapping the target object and the preloading
PIE object when possible.
Reported-by: ALT beekeeper
Reported-by: Gleb Fotengauer-Malinovskiy <glebfm@altlinux.org>
The patterns used were OK for the output of "file -NF$'\t'": it would
put a space after the separator (the default separator being ":").
But not for the output of "file -b": we need to pad the result with a
space to use the same patterns.
Putting a space at the beggining is convenient, because it allows to
match independently for "* sh" or "* bash", otherwise "*sh" would
consume "bash", too.
Other uses of "file -b" in scripts/ don't suffer from this problem.
ProvidedSymbols() used to be defined both in lib.prov.in and lib.req.in,
fix this code duplication by moving ProvidedSymbols() to separate
provided_symbols executable script.
Likewise, SuggestBPP() used to be defined both in lib.prov.in and
lib.req.in, fix this code duplication by moving SuggestBPP() to separate
suggest_bpp executable script.
lld linked binaries differ by structure from GNU ld, so eu-lint will
always complaint about. Allow to skip those checks by passing
%set_verify_elf_method lint=(skip|no)
Signed-off-by: Konstantin A. Lepikhov <lakostis@altlinux.ru>
This enhances ProvidedSymbols() in two ways:
* fixes st_shndx check for the case when it is not $7 in
"readelf --wide --dyn-syms" output as it happens on some architectures
like ppc64le;
* filters out special symbols (__bss_start, _edata, _end, _fini, _init)
after stripping of versioning.
Co-Authored-by: Dmitry V. Levin <ldv@altlinux.org>
Most of the binaries on mipsel do not contain the stack section,
so this check is not relevant for that architecture. Morover,
it fails, since it can't find the stack section, which is
(rightfully) not present.
eu-elflint does not understand mipsel binaries currently,
so here we disable that check, too.
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 $rpath was valid).
Now, the pipe (and the whole script) would abort on grep's failure.
Now, if something fails, the pipe (and the function, and the whole
script) will fail.
If the exit codes are normal, they are handled according to our logic.
Previously, a failure (which is an exceptional rare case) would lead to:
1. skipping the further exe_stack check and error_strict STACK "$f" 'STACK entry not found'
2. silently skipping the exe_stack check (which was impossible anyway because of the failure)
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.
However, when verify-elf is used standalone, $RPM_VERIFY_ELF_LDD_RPATH
is empty indeed, and this improvement becomes important.
Before this improvement:
$ RPM_BUILD_ROOT=/home /usr/lib/rpm/verify-elf /bin/pwd; echo $?
1
After this improvement:
$ VERIFY_ELF_UNRESOLVED=strict RPM_BUILD_ROOT=/home /usr/lib/rpm/verify-elf /bin/pwd; echo $?
0
$ VERIFY_ELF_UNRESOLVED=strict RPM_BUILD_ROOT=/home /usr/lib/rpm/verify-elf /usr/lib64/python3/site-packages/_cffi_backend.cpython-35m.so; echo $?
verify-elf: ERROR: /usr/lib64/python3/site-packages/_cffi_backend.cpython-35m.so: undefined symbol: PyExc_SystemError
....
1
When the given ELF file doesn't specify PT_INTERP, e.g. if it's a shared
library, we have to guess it. We used to try the program interpreter
of dump_ld_config, but it doesn't work well enough in multilib
environments, so extend the guess by trying interpreters listed
in RTLDLIST from /usr/bin/ldd script.
Add a brp interface for checking contents of files in buildroot.
The initial revision of brp-check_contents scripts checks
for suspicious path names like /var/lib/run, complementing
the change introduced in sisyphus_check-0.8.43-alt1.
When scanning files for deps, debuginfo, fixup, and verify_elf,
ignore files with type that matches "ELF * shared object, no machine, *"
pattern.
These are ELF files generated by GNU Guile, ignore them for now.
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.
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).
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.
