Siddhesh Poyarekar [Thu, 22 Jul 2021 13:08:06 +0000 (18:38 +0530)]
mtrace: Wean away from malloc hooks
Wean mtrace away from the malloc hooks and move them into the debug
DSO. Split the API away from the implementation so that we can add
the API to libc.so as well as libc_malloc_debug.so, with the libc
implementations being empty.
Update localplt data since memalign no longer has any callers after
this change.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Siddhesh Poyarekar [Thu, 22 Jul 2021 13:08:04 +0000 (18:38 +0530)]
Simplify __malloc_initialized
Now that mcheck no longer needs to check __malloc_initialized (and no
other third party hook can since the symbol is not exported), make the
variable boolean and static so that it is used strictly within malloc.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Siddhesh Poyarekar [Thu, 22 Jul 2021 13:08:02 +0000 (18:38 +0530)]
mcheck: Wean away from malloc hooks [BZ #23489]
Split the mcheck implementation into the debugging hooks and API so
that the API can be replicated in libc and libc_malloc_debug.so. The
libc APIs always result in failure.
The mcheck implementation has also been moved entirely into
libc_malloc_debug.so and with it, all of the hook initialization code
can now be moved into the debug library. Now the initialization can
be done independently of libc internals.
With this patch, libc_malloc_debug.so can no longer be used with older
libcs, which is not its goal anyway. tst-vfork3 breaks due to this
since it spawns shell scripts, which in turn execute using the system
glibc. Move the test to tests-container so that only the built glibc
is used.
This move also fixes bugs in the mcheck version of memalign and
realloc, thus allowing removal of the tests from tests-mcheck
exclusion list.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Siddhesh Poyarekar [Thu, 22 Jul 2021 13:07:59 +0000 (18:37 +0530)]
Move malloc hooks into a compat DSO
Remove all malloc hook uses from core malloc functions and move it
into a new library libc_malloc_debug.so. With this, the hooks now no
longer have any effect on the core library.
libc_malloc_debug.so is a malloc interposer that needs to be preloaded
to get hooks functionality back so that the debugging features that
depend on the hooks, i.e. malloc-check, mcheck and mtrace work again.
Without the preloaded DSO these debugging features will be nops.
These features will be ported away from hooks in subsequent patches.
Similarly, legacy applications that need hooks functionality need to
preload libc_malloc_debug.so.
The symbols exported by libc_malloc_debug.so are maintained at exactly
the same version as libc.so.
Finally, static binaries will no longer be able to use malloc
debugging features since they cannot preload the debugging DSO.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Siddhesh Poyarekar [Thu, 22 Jul 2021 13:07:57 +0000 (18:37 +0530)]
Remove __morecore and __default_morecore
Make the __morecore and __default_morecore symbols compat-only and
remove their declarations from the API. Also, include morecore.c
directly into malloc.c; this should ideally get merged into malloc in
a future cleanup.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Siddhesh Poyarekar [Thu, 22 Jul 2021 13:07:54 +0000 (18:37 +0530)]
Remove __after_morecore_hook
Remove __after_morecore_hook from the API and finalize the symbol so
that it can no longer be used in new applications. Old applications
using __after_morecore_hook will find that their hook is no longer
called.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Siddhesh Poyarekar [Thu, 22 Jul 2021 13:07:41 +0000 (18:37 +0530)]
Make mcheck tests conditional on GLIBC_2.23 or earlier
Targets with base versions of 2.24 or later won't have
__malloc_initialize_hook because of which the tests will essentially
be the same as the regular malloc tests. Avoid running them instead
and save time.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Samuel Thibault [Wed, 21 Jul 2021 23:23:02 +0000 (23:23 +0000)]
posix: Add sysconf(_SC_{MIN,}SIGSTKSZ) support
Samuel Thibault [Wed, 21 Jul 2021 22:40:57 +0000 (22:40 +0000)]
malloc: Fix tst-mallocfork3-malloc-check link
It uses pthread.
Vineet Gupta [Thu, 30 Jul 2020 05:49:14 +0000 (22:49 -0700)]
ARC: elf: make type safe
Signed-off-by: Vineet Gupta <vgupta@synopsys.com>
Vineet Gupta [Fri, 9 Jul 2021 21:56:18 +0000 (14:56 -0700)]
ARC: fp: (micro)optimize FPU_STATUS read by eliding FWE bit clearing
Any FPU_STATUS write needs setting the FWE bit (31) whcih just provides
a "control signal" to enable explicit write (vs. the side-effect of FPU
instructions). However this bit is RAZ and write-only, thus effectively
never stored in FPU_STATUS register. Thus when reading the register
there is no need to clear it. This shaves off a BCLR instruction from
the fe*exceptino family of functions and while no big deal still makes
sense to do.
This came up when debugging a race in math/test-fenv-tls [1]
[1]: https://github.com/foss-for-synopsys-dwc-arc-processors/linux/issues/54
Signed-off-by: Vineet Gupta <vgupta@synopsys.com>
Florian Weimer [Wed, 21 Jul 2021 09:42:31 +0000 (11:42 +0200)]
socket: Add time64 alias for sendmsg
Reviewed-by: Lukasz Majewski <lukma@denx.de>
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Florian Weimer [Wed, 21 Jul 2021 09:42:31 +0000 (11:42 +0200)]
socket: Add time64 alias for recvmsg
Reviewed-by: Lukasz Majewski <lukma@denx.de>
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Florian Weimer [Wed, 21 Jul 2021 09:42:31 +0000 (11:42 +0200)]
socket: Add time64 alias for sendmmsg
Reviewed-by: Lukasz Majewski <lukma@denx.de>
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Florian Weimer [Wed, 21 Jul 2021 09:42:31 +0000 (11:42 +0200)]
Linux: Add time64 alias for prctl
Reviewed-by: Lukasz Majewski <lukma@denx.de>
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Florian Weimer [Wed, 21 Jul 2021 09:42:31 +0000 (11:42 +0200)]
io: Add time64 alias for fcntl
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Florian Weimer [Wed, 21 Jul 2021 09:42:31 +0000 (11:42 +0200)]
misc: Add time64 alias for ioctl
Reviewed-by: Lukasz Majewski <lukma@denx.de>
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Darius Rad [Tue, 20 Jul 2021 00:45:59 +0000 (20:45 -0400)]
RISC-V: Update rv64 ULPs
Generated on a Microsemi Polarfire Icicle Kit running Linux version 5.6.18.
Same ULPs were also produced on QEMU 5.2.0 running Linux 5.10.46-1.
Florian Weimer [Tue, 20 Jul 2021 10:27:23 +0000 (12:27 +0200)]
resolv: Do not install libnss_dns.a, libnss_dns.so
Fixes commit
e1fcf21474c5b522f ("resolv: Move nss_dns into libc").
Samuel Thibault [Mon, 19 Jul 2021 21:45:29 +0000 (21:45 +0000)]
hurd: Add support for spawn_do_closefrom
Adhemerval Zanella [Thu, 15 Jul 2021 11:52:44 +0000 (08:52 -0300)]
elf: Fix tst-cpu-features-cpuinfo on some AMD systems (BZ #28090)
The SSBD feature is implemented in 2 different ways on AMD processors:
newer systems (Zen3) provides AMD_SSBD (function 8000_0008, EBX[24]),
while older system provides AMD_VIRT_SSBD (function 8000_0008, EBX[25]).
However for AMD_VIRT_SSBD, kernel shows both 'ssdb' and 'virt_ssdb' on
/proc/cpuinfo; while for AMD_SSBD only 'ssdb' is provided.
This now check is AMD_SSBD is set to check for 'ssbd', otherwise check
if AMD_VIRT_SSDB is set to check for 'virt_ssbd'.
Checked on x86_64-linux-gnu on a Ryzen 9 5900x.
Reviewed-by: H.J. Lu <hjl.tools@gmail.com>
H.J. Lu [Wed, 12 May 2021 18:02:47 +0000 (11:02 -0700)]
i386: Add the clone3 wrapper
extern int clone3 (struct clone_args *__cl_args, size_t __size,
int (*__func) (void *__arg), void *__arg);
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
nss: Directly load nss_dns, without going through dlsym/dlopen
This partially fixes static-only NSS support (bug 27959): The dns
module no longer needs dlopen. Support for disabling dlopen altogher
remains to be added.
This commit introduces module_load_builtin into nss/nss_module.c, which
handles the common parts of loading the built-in nss_files and nss_dns
modules.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Move nss_dns into libc
No abilist updates are needed because the symbols were GLIBC_PRIVATE.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Move res_query functions into libc
This switches to public symbols without __ prefixes, due to improved
namespace management in glibc.
The script was used with --no-new-version to move the symbols
__res_nquery, __res_nquerydomain, __res_nsearch, __res_query,
__res_querydomain, __res_search, res_query, res_querydomain,
res_search. The public symbols res_nquery, res_nquerydomain,
res_nsearch, res_ownok, res_query, res_querydomain, res_search
were added with make update-all-abi.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Move res_mkquery, res_nmkquery into libc
This switches to public symbols without __ prefixes, due to improved
namespace management in glibc.
The symbols res_mkquery, __res_mkquery, __res_nmkquery were
moved with the script (using --no-new-version).
res_mkquery@@GLIBC_2.34, res_nmkquery@@GLIBC_2.34 were added using
make update-all-abi.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Move res_send, res_nsend into libc
Switch to public symbols without __ prefix (due to improved
namespace management).
__res_send, __res_nsend were moved using the script (with
--no-new-version). res_send@@GLIBC_2.34 and res_nsend@@GLIBC_2.34
were added using make update-all-abi.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Move res_hostalias into its own file, along with hostalias
These deprecated symbols continue to be exported from libresolv.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Move __res_context_hostalias into its own file and into libc
And reformat it to GNU style. Remove the unecessary setbuf call.
Use __fgets_unlocked for PLT avoidance; no locking is required here.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Move res_queriesmatch to its own file and into libc
And reformat it to GNU style.
The treatment of this function matches res_nameinquery, for the
reasons stated there.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Move res_nameinquery to its own file and into libc
And reformat to GNU style.
This deprecated function is used in the implementation of the stub
resolver (for now). Keep the public symbol in libresolv for now
(so that no new symbol version is needed), and add a forwarder to
libresolv.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Move ns_samename into its own file, and into libc
But only as an internal symbol, __libc_ns_samename. The libresolv
ABI is preserved. This is because the function is deprecated, and
it does not make sense to add new symbol versions for deprecated
functions.
Also reformat the implementation to GNU style.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Move ns_makecanon into its own file, and into libc
But only as an internal symbol, __libc_ns_makecanon. The libresolv
ABI is preserved. This is because the function is deprecated, and
it does not make sense to add new symbol versions for deprecated
functions.
Also reformat the implementation to GNU style.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Move res_isourserver to its own file and reformat to GNU style
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Move __res_get_nsaddr to its own file and into libc
Eliminate the use of the EXT macro from it because it does not
add clarity. The function was added to res_send.c in 2015, and
the copyright year reflects that.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Rename res_comp.c to res-name-checking.c and move into libc
This reflects what the remaining functions in the file do.
The __res_dnok, __res_hnok, __res_mailok, __res_ownok were moved
with the script, using --no-new-version, and turned into compat
symbols. __libc_res_dnok@@GLIBC_PRIVATE and
__libc_res_hnok@@GLIBC_PRIVATE are added for internal use, to avoid
accidentally binding to compatibility symbols. The new public
symbols res_dnok, res_hnok, res_mailok, res_ownok were added using
make update-all-abi.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Move dn_skipname to its own file and into libc
And reformat it to GNU style.
dn_skipname is used outside glibc, so do not deprecate it,
and export it as dn_skipname (not __dn_skipname). Due to internal
users, provide a __libc_dn_skipname alias, and keep __dn_skipname
as a pure compatibility symbol.
__dn_skipname@GLIBC_2.0 was moved using the script, and
dn_skipname@@GLIBC_2.34 was added using make update-all-abi.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Move dn_comp to its own file and into libc
And reformat it to GNU style.
dn_comp is used in various programs, so keep it as a non-deprecated
symbol. Switch to dn_comp (not __dn_comp) for the ABI name. There
are no internal users, so interposition is not a problem.
The __dn_comp symbol was moved with scripts/move-symbol-to-libc.py
--no-new-version. dn_comp@@GLIBC_2.34 was added with
make update-all-abi.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Move _getlong, _getshort, __putlong, __putshort to res-putget
And reformat to GNU style.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Move dn_expand to its own file and into libc
And reformat to GNU style.
This switches back to the dn_expand name for the ABI symbol and turns
__dn_expand into a compatibility symbol. With the improved namespace
management in current glibc, it is no longer necessary to use a
private namespace symbol. To avoid old code binding to a
GLIBC_PRIVATE symbol by accident, use __libc_dn_expand for the
internal symbol name.
The symbols dn_expand, __dnexpand were moved using
scripts/move-symbol-to-libc.py, followed by an adjustment to make
dn_expand the only GLIBC_2.34 symbol.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Move ns_name_compress into its own file and into libc
And reformat to GNU style.
The symbol was moved using scripts/move-symbol-to-libc.py.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Move ns_name_pack into its own file and into libc
And reformat to GNU style, and eliminate the labellen function.
The symbol was moved using scripts/move-symbol-to-libc.py.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Move ns_name_pton into its own file and into libc
And reformat to GNU style, and eliminate the digits variable.
The symbol was moved using scripts/move-symbol-to-libc.py.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Move ns_name_uncompress into its own file and into libc
And reformat to GNU style. Check for negative error returns
(instead of -1).
The symbol was moved using scripts/move-symbol-to-libc.py.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Move ns_name_skip to its own file and into libc (bug 28091)
And reformat to GNU style. Avoid out-of-bounds pointer arithmetic.
This also results in a fix of bug 28091 due to the additional packet
length checks.
The symbol was moved using scripts/move-symbol-to-libc.py.
Reviewed-by: Carlos O'Donell <carlos@systemhalted.org>
Florian Weimer [Mon, 19 Jul 2021 05:55:27 +0000 (07:55 +0200)]
resolv: Deprecate legacy interfaces in libresolv
Debugging interfaces: p_*, fp_*, and sym_* could conceivably be
used to produce debug out, but these functions have not been
updated to parse more resource records, so they are not very useful
today. Likewise for ns_sprintrr and ns_sprintrrf. ns_format_ttl and
ns_parse_ttl are related to these.
Internal implementation details: res_isourserver is probably only
useful in the implementation of a stub resolver, and so is
res_nameinquery.
Unclear semantics and bad performance: ns_samedomain, ns_subdomain,
ns_makecanon, ns_samename do textual converions & copies instead of
checking equivalence of the wire format.
inet_neta cannot handle IPv6 addresses.
res_hostalias has been superseded by getaddrinfo with AI_CANONNAME.
hostalias is not thread-safe.
Some functions have int as size arguments instead of size_t, so they
do not follow current coding practices. However, dn_expand and
b64_ntop are somewhat widely used (to name just two examples), so
deprecating them seems problematic.
Reviewed-by: Carlos O'Donell <carlos@systemhalted.org>
Siddhesh Poyarekar [Mon, 19 Jul 2021 02:59:25 +0000 (08:29 +0530)]
tst-safe-linking: make false positives even more improbable
There is a 1 in 16 chance of a corruption escaping safe-linking and to
guard against spurious failures, tst-safe-linking runs each subtest 10
times to ensure that the chance is reduced to 1 in 2^40. However, in
the 1 in 16 chance that a corruption does escape safe linking, it
could well be caught by other sanity checks we do in malloc, which
then results in spurious test failures like below:
test test_fastbin_consolidate failed with a different error
expected: malloc_consolidate(): unaligned fastbin chunk detected
actual: malloc_consolidate(): invalid chunk size
This failure is seen more frequently on i686; I was able to reproduce
it in about 5 min of running it in a loop.
Guard against such failures by recording them and retrying the test.
Also, do not fail the test if we happened to get defeated by the 1 in
2^40 odds if in at least one of the instances it was detected by other
checks.
Finally, bolster the odds to 2^64 by running 16 times instead of 10.
The test still has a chance of failure so it is still flaky in theory.
However in practice if we see a failure here then it's more likely
that there's a bug than it being an issue with the test. Add more
printfs and also dump them to stdout so that in the event the test
actually fails, we will have some data to try and understand why it
may have failed.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Samuel Thibault [Sun, 18 Jul 2021 20:25:33 +0000 (20:25 +0000)]
htl: Do not expose pthread hidden proto outside libpthread
Only libpthread.so can access them.
Adhemerval Zanella [Fri, 16 Jul 2021 11:32:05 +0000 (08:32 -0300)]
elf: Fix a wrong array access on tst-tls20
Check on x86_64-linux-gnu with --enable-stack-protector=all.
Adhemerval Zanella [Thu, 15 Jul 2021 11:18:21 +0000 (08:18 -0300)]
elf: Add -Wl,--no-as-needed for tst-tls-manydynamic*mod-dep-bad.so (BZ #28089)
The tests explicit requires the dependencies and it is required for
the case the toolchain defaults to -Wl,--as-needed.
Checked on x86_64-linux-gnu.
Florian Weimer [Thu, 15 Jul 2021 06:28:50 +0000 (08:28 +0200)]
resolv: Move ns_name_unpack to its own file and into libc
Reformat to GNU style. Avoid out-of-bounds buffer arithmetic.
Eliminate the labellen function.
The symbol was moved using scripts/move-symbol-to-libc.py.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Thu, 15 Jul 2021 06:28:51 +0000 (08:28 +0200)]
resolv: Remove unnecessary res_isourserver_p call from send_dg
As the comment indicates, the check is unnecessary due to the way the
UDP socket is set up.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Thu, 15 Jul 2021 06:28:50 +0000 (08:28 +0200)]
resolv: Move ns_name_ntop to its own file and into libc
Reformat to GNU style. Avoid out-of-bounds pointer arithmetic
(e.g., use eom - dn < 2 instead of dn + 1 >= eom). Inline the
labellen function and fold the compression pointer check into
the length check (l >= 64). Assume ASCII encoding.
The symbol was moved using scripts/move-symbol-to-libc.py.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Thu, 15 Jul 2021 06:28:50 +0000 (08:28 +0200)]
nss_dns: Do not use deprecated packet parsing functions
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Thu, 15 Jul 2021 06:28:50 +0000 (08:28 +0200)]
resolv: Sort Makefile routines and Versions lexicographically
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Thu, 15 Jul 2021 06:28:50 +0000 (08:28 +0200)]
socket: Add hidden prototype for setsockopt
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Tested-by: Carlos O'Donell <carlos@redhat.com>
Adhemerval Zanella [Fri, 25 Jun 2021 13:54:12 +0000 (10:54 -0300)]
elf: Fix DTV gap reuse logic (BZ #27135)
This is updated version of the
572bd547d57a (reverted by
40ebfd016ad2)
that fixes the _dl_next_tls_modid issues.
This issue with
572bd547d57a patch is the DTV entry will be only
update on dl_open_worker() with the update_tls_slotinfo() call after
all dependencies are being processed by _dl_map_object_deps(). However
_dl_map_object_deps() itself might call _dl_next_tls_modid(), and since
the _dl_tls_dtv_slotinfo_list::map is not yet set the entry will be
wrongly reused.
This patch fixes by renaming the _dl_next_tls_modid() function to
_dl_assign_tls_modid() and by passing the link_map so it can set
the slotinfo value so a subsequente _dl_next_tls_modid() call will
see the entry as allocated.
The intermediary value is cleared up on remove_slotinfo() for the case
a library fails to load with RTLD_NOW.
This patch fixes BZ #27135.
Checked on x86_64-linux-gnu.
Reviewed-by: Szabolcs Nagy <szabolcs.nagy@arm.com>
Stefan Liebler [Wed, 14 Jul 2021 13:58:08 +0000 (15:58 +0200)]
Fix linknamespace errors and local-plt-usages in nss_files.
After commit
f9c8b11ed7726b858cd7b7cea0d3d7c5233d78cf
"nss: Access nss_files through direct references",
when building with -Os, multiple conform/.../linknamespace tests
and elf/check-localplt are failing:
Extra PLT reference: libc.so: fgetc_unlocked
Extra PLT reference: libc.so: getline
Or e.g.:
[initial] glob -> [libc.a(glob.o)] __getpwnam_r -> [libc.a(getpwnam_r.o)] __nss_database_custom -> [libc.a(nsswitch.o)] __nss_module_get_function -> [libc.a(nss_module.o)] __nss_files_functions -> [libc.a(nss_files_functions.o)] _nss_files_endaliasent -> [libc.a(files-alias.o)] feof_unlocked
[initial] glob -> [libc.a(glob.o)] __getpwnam_r -> [libc.a(getpwnam_r.o)] __nss_database_custom -> [libc.a(nsswitch.o)] __nss_module_get_function -> [libc.a(nss_module.o)] __nss_files_functions -> [libc.a(nss_files_functions.o)] _nss_files_endaliasent -> [libc.a(files-alias.o)] fgetc_unlocked
[initial] glob -> [libc.a(glob.o)] __getpwnam_r -> [libc.a(getpwnam_r.o)] __nss_database_custom -> [libc.a(nsswitch.o)] __nss_module_get_function -> [libc.a(nss_module.o)] __nss_files_functions -> [libc.a(nss_files_functions.o)] _nss_files_endnetgrent -> [libc.a(files-netgrp.o)] getline
This patch is using the hidden symbols where possible.
Instead of fputc_unlocked, __putc_unlocked is used.
(Compare to commit
eeaa19f75e52d2d48074ae0c423f2311d67c42c6
"mntent: Use __putc_unlocked instead of fputc_unlocked")
H.J. Lu [Fri, 14 May 2021 22:23:46 +0000 (15:23 -0700)]
Add static tests for __clone_internal
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
H.J. Lu [Thu, 13 May 2021 15:24:36 +0000 (08:24 -0700)]
x86-64: Add the clone3 wrapper
extern int clone3 (struct clone_args *__cl_args, size_t __size,
int (*__func) (void *__arg), void *__arg);
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
H.J. Lu [Sat, 13 Feb 2021 19:47:46 +0000 (11:47 -0800)]
Add an internal wrapper for clone, clone2 and clone3
The clone3 system call (since Linux 5.3) provides a superset of the
functionality of clone and clone2. It also provides a number of API
improvements, including the ability to specify the size of the child's
stack area which can be used by kernel to compute the shadow stack size
when allocating the shadow stack. Add:
extern int __clone_internal (struct clone_args *__cl_args,
int (*__func) (void *__arg), void *__arg);
to provide an abstract interface for clone, clone2 and clone3.
1. Simplify stack management for thread creation by passing both stack
base and size to create_thread.
2. Consolidate clone vs clone2 differences into a single file.
3. Call __clone3 if HAVE_CLONE3_WAPPER is defined. If __clone3 returns
-1 with ENOSYS, fall back to clone or clone2.
4. Use only __clone_internal to clone a thread. Since the stack size
argument for create_thread is now unconditional, always pass stack size
to create_thread.
5. Enable the public clone3 wrapper in the future after it has been
added to all targets.
NB: Sandbox will return ENOSYS on clone3 in both Chromium:
The following revision refers to this bug:
https://chromium.googlesource.com/chromium/src/+/
218438259dd795456f0a48f67cbe5b4e520db88b
commit
218438259dd795456f0a48f67cbe5b4e520db88b
Author: Matthew Denton <mpdenton@chromium.org>
Date: Thu Jun 03 20:06:13 2021
Linux sandbox: return ENOSYS for clone3
Because clone3 uses a pointer argument rather than a flags argument, we
cannot examine the contents with seccomp, which is essential to
preventing sandboxed processes from starting other processes. So, we
won't be able to support clone3 in Chromium. This CL modifies the
BPF policy to return ENOSYS for clone3 so glibc always uses the fallback
to clone.
Bug: 1213452
Change-Id: I7c7c585a319e0264eac5b1ebee1a45be2d782303
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2936184
Reviewed-by: Robert Sesek <rsesek@chromium.org>
Commit-Queue: Matthew Denton <mpdenton@chromium.org>
Cr-Commit-Position: refs/heads/master@{#888980}
[modify] https://crrev.com/
218438259dd795456f0a48f67cbe5b4e520db88b/sandbox/linux/seccomp-bpf-helpers/baseline_policy.cc
and Firefox:
https://hg.mozilla.org/integration/autoland/rev/
ecb4011a0c76
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Cooper Qu [Tue, 13 Jul 2021 12:50:40 +0000 (20:50 +0800)]
nss: Fix build error with --disable-nscd
The error is as follows:
nss_module.c: In function 'module_load_nss_files':
nss_module.c:117:7: error: 'is_nscd' undeclared (first use in this function)
117 | if (is_nscd)
| ^~~~~~~
nss_module.c:117:7: note: each undeclared identifier is reported only once for each function it appears in
nss_module.c:119:51: error: 'nscd_init_cb' undeclared (first use in this function); did you mean 'nscd_init'?
119 | void (*cb) (size_t, struct traced_file *) = nscd_init_cb;
| ^~~~~~~~~~~~
| nscd_init
Samuel Thibault [Tue, 13 Jul 2021 21:45:57 +0000 (21:45 +0000)]
htl: Fix linking static examples against libpthread
libpthread.a uses some mach and hurd RPCs so we need to link them in.
Samuel Thibault [Tue, 13 Jul 2021 21:33:39 +0000 (21:33 +0000)]
htl: Let libc call __pthread_mutex_{,try,un}lock
Now that NPTL was moved to libc, libc makes internal __pthread calls, so
htl has to expose them internally.
Adhemerval Zanella [Tue, 13 Jul 2021 13:46:23 +0000 (10:46 -0300)]
posix: Ignore non opened files on tst-spawn5
The make program might open a pipe for its job server, which triggers
an invalid check on the spawned process. This patch now passes the
lowest file descriptor as ithe first argument, so only the range
that was actually opened is checked.
Checked on x86_64-linux-gnu and i686-linux-gnu and centos7 (which
triggers the issue).
H.J. Lu [Mon, 12 Jul 2021 21:36:39 +0000 (14:36 -0700)]
mcheck: Align struct hdr to MALLOC_ALIGNMENT bytes [BZ #28068]
1. Align struct hdr to MALLOC_ALIGNMENT bytes so that malloc hooks in
libmcheck align memory to MALLOC_ALIGNMENT bytes.
2. Remove tst-mallocalign1 from tests-exclude-mcheck for i386 and x32.
3. Add tst-pvalloc-fortify and tst-reallocarray to tests-exclude-mcheck
since they use malloc_usable_size (see BZ #22057).
This fixed BZ #28068.
Reviewed-by: Siddhesh Poyarekar <siddhesh@sourceware.org>
Adhemerval Zanella [Sat, 10 Jul 2021 20:03:49 +0000 (17:03 -0300)]
Linux: Use 32-bit vDSO for clock_gettime, gettimeofday, time (BZ# 28071)
The previous approach defeats the vDSO optimization on older kernels
because a failing clock_gettime64 system call is performed on every
function call. It also results in a clobbered errno value, exposing
an OpenJDK bug (JDK-8270244).
This patch fixes by open-code INLINE_VSYSCALL macro and replace all
INLINE_SYSCALL_CALL with INTERNAL_SYSCALL_CALLS. Now for
__clock_gettime64x, the 64-bit vDSO is used and the 32-bit vDSO is
tried before falling back to 64-bit syscalls.
The previous code preferred 64-bit syscall for the case where the kernel
provides 64-bit time_t syscalls *and* also a 32-bit vDSO (in this case
the *64-bit* syscall should be preferable over the vDSO). All
architectures that provides 32-bit vDSO (i386, mips, powerpc, s390)
modulo sparc; but I am not sure if some kernels versions do provide
only 32-bit vDSO while still providing 64-bit time_t syscall.
Regardless, for such cases the 64-bit time_t syscall is used if the
vDSO returns overflowed 32-bit time_t.
Tested on i686-linux-gnu (with a time64 and non-time64 kernel),
x86_64-linux-gnu. Built with build-many-glibcs.py.
Co-authored-by: Florian Weimer <fweimer@redhat.com>
Florian Weimer [Mon, 12 Jul 2021 16:43:32 +0000 (18:43 +0200)]
Reduce <limits.h> pollution due to dynamic PTHREAD_STACK_MIN
<limits.h> used to be a header file with no declarations.
GCC's libgomp includes it in a #pragma GCC visibility hidden block.
Including <unistd.h> from <limits.h> (indirectly) declares everything
in <unistd.h> with hidden visibility, resulting in linker failures.
This commit avoids C declarations in assembler mode and only declares
__sysconf in <limits.h> (and not the entire contents of <unistd.h>).
The __sysconf symbol is already part of the ABI. PTHREAD_STACK_MIN
is no longer defined for __USE_DYNAMIC_STACK_SIZE && __ASSEMBLER__
because there is no possible definition.
Additionally, PTHREAD_STACK_MIN is now defined by <pthread.h> for
__USE_MISC because this is what developers expect based on the macro
name. It also helps to avoid libgomp linker failures in GCC because
libgomp includes <pthread.h> before its visibility hacks.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Stefan Liebler [Mon, 12 Jul 2021 09:00:53 +0000 (11:00 +0200)]
Fix failing nss/tst-nss-files-hosts-long.
Sometimes the test nss/tst-nss-files-hosts-long is failing as getent
fails with exit-code 2.
This happens if tst-reload1 was run just before this test:
make t=nss/tst-reload1 test
make t=nss/tst-nss-files-hosts-long test
Then the test fails as /etc/nsswitch.conf contains "hosts: test2"
and the hosts are not searched in /etc/hosts at all.
Thus this patch just requests a post cleanup after nss/tst-reload1
has run.
Florian Weimer [Mon, 12 Jul 2021 05:58:07 +0000 (07:58 +0200)]
nis: nis_local_group may read from __nisgroup[-1] (bug 28075)
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Samuel Thibault [Sun, 11 Jul 2021 17:51:12 +0000 (17:51 +0000)]
hurd _Fork: Drop duplicate malloc_fork_lock calls
This was put in __libc_fork by
c32c868ab8b2 ("posix: Add _Fork [BZ #4737]")
so we need to avoid locking them again in _Fork called by __libc_lock, otherwise
we deadlock.
H.J. Lu [Sat, 10 Jul 2021 17:56:50 +0000 (10:56 -0700)]
support: Replace _SC_MINSIGSTKSZ with _SC_SIGSTKSZ
Replace _SC_MINSIGSTKSZ with _SC_SIGSTKSZ since sysconf (_SC_MINSIGSTKSZ)
returns the minimum number of bytes of free stack space required in order
to guarantee successful, non-nested handling of a single signal whose
handler is an empty function while sysconf (_SC_SIGSTKSZ) returns the
suggested minimum number of bytes of stack space required for a signal
stack.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
H.J. Lu [Fri, 9 Jul 2021 21:17:04 +0000 (14:17 -0700)]
support: Replace MINSIGSTKSZ with sysconf (_SC_MINSIGSTKSZ)
Replace MINSIGSTKSZ with sysconf (_SC_MINSIGSTKSZ) since the constant
MINSIGSTKSZ used in glibc build may be too small.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
H.J. Lu [Mon, 21 Jun 2021 19:42:56 +0000 (12:42 -0700)]
Define PTHREAD_STACK_MIN to sysconf(_SC_THREAD_STACK_MIN)
The constant PTHREAD_STACK_MIN may be too small for some processors.
Rename _SC_SIGSTKSZ_SOURCE to _DYNAMIC_STACK_SIZE_SOURCE. When
_DYNAMIC_STACK_SIZE_SOURCE or _GNU_SOURCE are defined, define
PTHREAD_STACK_MIN to sysconf(_SC_THREAD_STACK_MIN) which is changed
to MIN (PTHREAD_STACK_MIN, sysconf(_SC_MINSIGSTKSZ)).
Consolidate <bits/local_lim.h> with <bits/pthread_stack_min.h> to
provide a constant target specific PTHREAD_STACK_MIN value.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
Florian Weimer [Fri, 9 Jul 2021 18:09:14 +0000 (20:09 +0200)]
Force building with -fno-common
As a result, is not necessary to specify __attribute__ ((nocommon))
on individual definitions.
GCC 10 defaults to -fno-common on all architectures except ARC,
but this change is compatible with older GCC versions and ARC, too.
Reviewed-by: Carlos O'Donell <carlos@redhat.com>
H.J. Lu [Fri, 9 Jul 2021 03:48:14 +0000 (20:48 -0700)]
Add a generic malloc test for MALLOC_ALIGNMENT
1. Add sysdeps/generic/malloc-size.h to define size related macros for
malloc.
2. Move x86_64/tst-mallocalign1.c to malloc and replace ALIGN_MASK with
MALLOC_ALIGN_MASK.
3. Add tst-mallocalign1 to tests-exclude-mcheck for i386 and x32 since
mcheck doesn't honor MALLOC_ALIGNMENT.
H.J. Lu [Fri, 9 Jul 2021 12:57:51 +0000 (05:57 -0700)]
Properly run tst-spawn5 directly [BZ #28067]
Change tst-spawn5.c to handle tst-spawn5 without optional path to ld.so,
--library-path nor the library path when glibc is configured with
--enable-hardcoded-path-in-tests. This fixes BZ #28067.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Florian Weimer [Fri, 9 Jul 2021 08:59:22 +0000 (10:59 +0200)]
build-many-glibcs.py: Add glibcs-arm-linux-gnueabihf-thumb
Previously, there was no thumb variant, despite that building
glibc in Thumb-2 mode is supported.
Florian Weimer [Fri, 9 Jul 2021 08:59:22 +0000 (10:59 +0200)]
nptl: Use out-of-line wake function in __libc_lock_unlock slow path
This slightly reduces code size, as can be seen below.
__libc_lock_unlock is usually used along with __libc_lock_lock in
the same function. __libc_lock_lock already has an out-of-line
slow path, so this change should not introduce many additional
non-leaf functions.
This change also fixes a link failure in 32-bit Arm thumb mode
because commit
1f9c804fbd699104adefbce9e56d2c8aa711b6b9
("nptl: Use internal low-level lock type for !IS_IN (libc)")
introduced __libc_do_syscall calls outside of libc.
Before x86-64:
text data bss dec hex filename
1937748 20456 54896 2013100 1eb7ac libc.so.6
25601 856 12768 39225 9939 nss/libnss_db.so.2
40310 952 25144 66406 10366 nss/libnss_files.so.2
After x86-64:
text data bss dec hex filename
1935312 20456 54896 2010664 1eae28 libc.so.6
25559 864 12768 39191 9917 nss/libnss_db.so.2
39764 960 25144 65868 1014c nss/libnss_files.so.2
Before i686:
2110961 11272 39144 2161377 20fae1 libc.so.6
27243 428 12652 40323 9d83 nss/libnss_db.so.2
43062 476 25028 68566 10bd6 nss/libnss_files.so.2
After i686:
2107347 11272 39144 2157763 20ecc3 libc.so.6
26929 432 12652 40013 9c4d nss/libnss_db.so.2
43132 480 25028 68640 10c20 nss/libnss_files.so.2
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Anton Blanchard [Tue, 6 Jul 2021 10:19:36 +0000 (20:19 +1000)]
powerpc64le: Fix typo in configure
The configure script checks for -mlong-double-128 but mentions -mlongdouble
when it fails.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Tulio Magno Quites Machado Filho [Fri, 2 Jul 2021 18:42:31 +0000 (15:42 -0300)]
powerpc64: Remove strcspn ifunc from the loader
5 years ago, commit
8f1b841e452dbb083112fd036033b7f4af506ba0
unintentionally added an ifunc to the loader.
That modification has not caused any harm so far, but it doesn't add any
value either, because the hwcap information is available later during
libc initialization.
Suggested-by: Anton Blanchard <anton@ozlabs.org>
Noah Goldstein [Thu, 8 Jul 2021 20:13:19 +0000 (16:13 -0400)]
x86: Remove wcsnlen-sse4_1 from wcslen ifunc-impl-list [BZ #28064]
The following commit
commit
6f573a27b6c8b4236445810a44660612323f5a73
Author: Noah Goldstein <goldstein.w.n@gmail.com>
Date: Wed Jun 23 01:19:34 2021 -0400
x86-64: Add wcslen optimize for sse4.1
Added wcsnlen-sse4.1 to the wcslen ifunc implementation list and did
not add wcslen-sse4.1 to wcslen ifunc implementation list. This commit
fixes that by removing wcsnlen-sse4.1 from the wcslen ifunc
implementation list and adding wcslen-sse4.1 to the ifunc
implementation list.
Testing:
test-wcslen.c, test-rsi-wcslen.c, and test-rsi-strlen.c are passing as
well as all other tests in wcsmbs and string.
Signed-off-by: Noah Goldstein <goldstein.w.n@gmail.com>
Reviewed-by: H.J. Lu <hjl.tools@gmail.com>
Reviewed-by: H.J. Lu <hjl.tools@gmail.com>
H.J. Lu [Thu, 8 Jul 2021 20:13:17 +0000 (16:13 -0400)]
x86-64: Test strlen and wcslen with 0 in the RSI register [BZ #28064]
commit
6f573a27b6c8b4236445810a44660612323f5a73
Author: Noah Goldstein <goldstein.w.n@gmail.com>
Date: Wed Jun 23 01:19:34 2021 -0400
x86-64: Add wcslen optimize for sse4.1
added wcsnlen-sse4.1 to the wcslen ifunc implementation list. Since the
random value in the the RSI register is larger than the wide-character
string length in the existing wcslen test, it didn't trigger the wcslen
test failure. Add a test to force 0 into the RSI register before calling
wcslen.
Fangrui Song [Thu, 8 Jul 2021 21:26:22 +0000 (14:26 -0700)]
x86_64: Remove unneeded static PIE check for undefined weak diagnostic
https://sourceware.org/bugzilla/show_bug.cgi?id=21782 dropped an ld
diagnostic for R_X86_64_PC32 referencing an undefined weak symbol in
-pie links. Arguably keeping the diagnostic like other ports is more
correct, since statically resolving movl foo(%rip), %eax to the
link-time zero address produces a corrupted output.
It turns out that --enable-static-pie builds do not depend on the ld
behavior. GCC generates GOT indirection for weak declarations for
-fPIE/-fPIC, so what ld does with the PC-relative relocation doesn't
really matter.
Reviewed-by: H.J. Lu <hjl.tools@gmail.com>
Joseph Myers [Thu, 8 Jul 2021 17:36:27 +0000 (17:36 +0000)]
Add NT_ARM_PAC_ENABLED_KEYS to elf.h
This patch adds the new NT_ARM_PAC_ENABLED_KEYS constant from Linux
5.13 to glibc's elf.h.
Tested for x86_64.
Adhemerval Zanella [Wed, 10 Mar 2021 15:26:33 +0000 (12:26 -0300)]
posix: Add posix_spawn_file_actions_addclosefrom_np
This patch adds a way to close a range of file descriptors on
posix_spawn as a new file action. The API is similar to the one
provided by Solaris 11 [1], where the file action causes the all open
file descriptors greater than or equal to input on to be closed when
the new process is spawned.
The function posix_spawn_file_actions_addclosefrom_np is safe to be
implemented by iterating over /proc/self/fd, since the Linux spawni.c
helper process does not use CLONE_FILES, so its has own file descriptor
table and any failure (in /proc operation) aborts the process creation
and returns an error to the caller.
I am aware that this file action might be redundant to the current
approach of POSIX in promoting O_CLOEXEC in more interfaces. However
O_CLOEXEC is still not the default and for some specific usages, the
caller needs to close all possible file descriptors to avoid them
leaking. Some examples are CPython (discussed in BZ#10353) and OpenJDK
jspawnhelper [2] (where OpenJDK spawns a helper process to exactly
closes all file descriptors). Most likely any environment which calls
functions that might open file descriptor under the hood and aim to use
posix_spawn might face the same requirement.
Checked on x86_64-linux-gnu and i686-linux-gnu on kernel 5.11 and 4.15.
[1] https://docs.oracle.com/cd/E36784_01/html/E36874/posix-spawn-file-actions-addclosefrom-np-3c.html
[2] https://github.com/openjdk/jdk/blob/master/src/java.base/unix/native/libjava/childproc.c#L82
Adhemerval Zanella [Wed, 10 Mar 2021 15:26:32 +0000 (12:26 -0300)]
io: Add closefrom [BZ #10353]
The function closes all open file descriptors greater than or equal to
input argument. Negative values are clamped to 0, i.e, it will close
all file descriptors.
As indicated by the bug report, this is a common symbol provided by
different systems (Solaris, OpenBSD, NetBSD, FreeBSD) and, although
its has inherent issues with not taking in consideration internal libc
file descriptors (such as syslog), this is also a common feature used
in multiple projects [1][2][3][4][5].
The Linux fallback implementation iterates over /proc and close all
file descriptors sequentially. Although it was raised the questioning
whether getdents on /proc/self/fd might return disjointed entries
when file descriptor are closed; it does not seems the case on my
testing on multiple kernel (v4.18, v5.4, v5.9) and the same strategy
is used on different projects [1][2][3][5].
Also, the interface is set a fail-safe meaning that a failure in the
fallback results in a process abort.
Checked on x86_64-linux-gnu and i686-linux-gnu on kernel 5.11 and 4.15.
[1] https://github.com/systemd/systemd/blob/
5238e9575906297608ff802a27e2ff9effa3b338/src/basic/fd-util.c#L217
[2] https://github.com/lxc/lxc/blob/
ddf4b77e11a4d08f09b7b9cd13e593f8c047edc5/src/lxc/start.c#L236
[3] https://github.com/python/cpython/blob/
9e4f2f3a6b8ee995c365e86d976937c141d867f8/Modules/_posixsubprocess.c#L220
[4] https://github.com/rust-lang/rust/blob/
5f47c0613ed4eb46fca3633c1297364c09e5e451/src/libstd/sys/unix/process2.rs#L303-L308
[5] https://github.com/openjdk/jdk/blob/master/src/java.base/unix/native/libjava/childproc.c#L82
Adhemerval Zanella [Wed, 10 Mar 2021 15:26:31 +0000 (12:26 -0300)]
linux: Add close_range
It was added on Linux 5.9 (
278a5fbaed89) with CLOSE_RANGE_CLOEXEC
added on 5.11 (
582f1fb6b721f). Although FreeBSD has added the same
syscall, this only adds the symbol on Linux ports. This syscall is
required to provided a fail-safe way to implement the closefrom
symbol (BZ #10353).
Checked on x86_64-linux-gnu and i686-linux-gnu on kernel 5.11 and 4.15.
Adhemerval Zanella [Wed, 10 Mar 2021 15:26:29 +0000 (12:26 -0300)]
support: Add support_stack_alloc
The code to allocate a stack from xsigstack is refactored so it can
be more generic. The new support_stack_alloc() also set PROT_EXEC
if DEFAULT_STACK_PERMS has PF_X. This is required on some
architectures (hppa for instance) and trying to access the rtld
global from testsuite will require more intrusive refactoring
in the ldsodefs.h header.
Checked on x86_64-linux-gnu and i686-linux-gnu. I also ran
tst-xsigstack on both hppa and ia64.
Siddhesh Poyarekar [Wed, 7 Jul 2021 09:31:20 +0000 (15:01 +0530)]
_int_realloc is static
_int_realloc is correctly declared at the top to be static, but
incorrectly defined without the static keyword. Fix that. The
generated binaries have identical code.
Siddhesh Poyarekar [Wed, 7 Jul 2021 13:44:00 +0000 (19:14 +0530)]
Move mcheck symbol from stdlib to malloc
It is defined in malloc, so it belongs there. Verified on x86_64 that
the built libraries are identical despite this change.
Florian Weimer [Thu, 8 Jul 2021 12:06:53 +0000 (14:06 +0200)]
nss: Do not install static linker input files for libnss_files
Siddhesh Poyarekar [Thu, 8 Jul 2021 08:09:19 +0000 (13:39 +0530)]
elf/tests: Make thrlock and noload depend on libm
Both tests try to dlopen libm.so at runtime, so make them depend on it
so that they're executed if libm.so has been updated.
Reviewed-by: Florian Weimer <fweimer@redhat.com>
Siddhesh Poyarekar [Wed, 7 Jul 2021 17:32:46 +0000 (23:02 +0530)]
Harden tcache double-free check
The tcache allocator layer uses the tcache pointer as a key to
identify a block that may be freed twice. Since this is in the
application data area, an attacker exploiting a use-after-free could
potentially get access to the entire tcache structure through this
key. A detailed write-up was provided by Awarau here:
https://awaraucom.wordpress.com/2020/07/19/house-of-io-remastered/
Replace this static pointer use for key checking with one that is
generated at malloc initialization. The first attempt is through
getrandom with a fallback to random_bits(), which is a simple
pseudo-random number generator based on the clock. The fallback ought
to be sufficient since the goal of the randomness is only to make the
key arbitrary enough that it is very unlikely to collide with user
data.
Co-authored-by: Eyal Itkin <eyalit@checkpoint.com>
Florian Weimer [Wed, 7 Jul 2021 16:33:52 +0000 (18:33 +0200)]
nss: Access nss_files through direct references
This partially fixes static-only NSS support (bug 27959): The files
module no longer needs dlopen. Support for the dns module remains
to be added, and also support for disabling dlopen altogether.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Florian Weimer [Wed, 7 Jul 2021 16:33:52 +0000 (18:33 +0200)]
nss_files: Move into libc
This is the first step towards fixing bug 27959.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Florian Weimer [Wed, 7 Jul 2021 16:33:52 +0000 (18:33 +0200)]
nss_files: Add generic code for set*ent, end*ent and file open
This reduces RSS usage if nss_files is not actually used, and can
be used later to make NSS data thread-specific. It also results in
a small code size reduction.
Before:
text data bss dec hex filename
2288 0 72 2360 938 nss/files-alias.os
1807 0 72 1879 757 nss/files-ethers.os
1371 0 72 1443 5a3 nss/files-grp.os
6246 0 72 6318 18ae nss/files-hosts.os
869 0 0 869 365 nss/files-initgroups.os
666 0 0 666 29a nss/files-init.os
1934 0 0 1934 78e nss/files-netgrp.os
2353 0 72 2425 979 nss/files-network.os
2130 0 72 2202 89a nss/files-proto.os
1372 0 72 1444 5a4 nss/files-pwd.os
2124 0 72 2196 894 nss/files-rpc.os
2265 0 72 2337 921 nss/files-service.os
1125 0 72 1197 4ad nss/files-sgrp.os
1124 0 72 1196 4ac nss/files-spwd.os
After:
text data bss dec hex filename
2040 0 0 2040 7f8 nss/files-alias.os
1599 0 0 1599 63f nss/files-ethers.os
1155 0 0 1155 483 nss/files-grp.os
6010 0 0 6010 177a nss/files-hosts.os
869 0 0 869 365 nss/files-initgroups.os
666 0 0 666 29a nss/files-init.os
1934 0 0 1934 78e nss/files-netgrp.os
2129 0 0 2129 851 nss/files-network.os
1914 0 0 1914 77a nss/files-proto.os
1156 0 0 1156 484 nss/files-pwd.os
1908 0 0 1908 774 nss/files-rpc.os
2057 0 0 2057 809 nss/files-service.os
909 0 0 909 38d nss/files-sgrp.os
908 0 0 908 38c nss/files-spwd.os
1090 0 8 1098 44a nss/nss_files_data.os
27674 code bytes before, 26344 code bytes after, so it is an overall
win despite the extra initialization code.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Florian Weimer [Wed, 7 Jul 2021 16:33:52 +0000 (18:33 +0200)]
nss_files: Allocate nscd file registration data on the heap
This is only needed if nss_files is loaded by nscd.
Before:
text data bss dec hex filename
767 0 24952 25719 6477 nss/files-init.os
After:
text data bss dec hex filename
666 0 0 666 29a nss/files-init.os
Using PATH_MAX bytes unconditionally for the directory name
is wasteful, but fixing that would constitute another break
of this semi-public ABI. (The other issue is that with
symbolic links, an arbitrary set of parent directories may need
watching, not just a single one.)
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Florian Weimer [Wed, 7 Jul 2021 16:33:52 +0000 (18:33 +0200)]
libio: Replace internal _IO_getdelim symbol with __getdelim
__getdelim is exported, _IO_getdelim is not. Add a hidden prototype
for __getdelim.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
Florian Weimer [Wed, 7 Jul 2021 16:33:52 +0000 (18:33 +0200)]
libio: Add hidden prototype for ungetc
And make ungetc the primary symbol, with _IO_ungetc as an alias.
Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>