previously, LC_MESSAGES was treated specially as the only category
which could be set to a locale name without a definition file, in
order to facilitate gettext message translations when no libc locale
was available. LC_NUMERIC was completely un-settable, and LC_CTYPE
stored a flag intended to be used for a possible future byte-based C
locale, instead of storing a __locale_map pointer like the other
categories use.
this patch changes all categories to be represented by pointers to
__locale_map structures, and allows locale names without definition
files to be treated as valid locales with trivial definition when used
in any category. outwardly visible functional changes should be minor,
limited mainly to the strings read back from setlocale and the way
gettext handles translations in categories other than LC_MESSAGES.
various internal refactoring has also been performed, and improvements
in const correctness have been made.
this is part of a general program of removing direct use of atomics
where they are not necessary to meet correctness or performance needs,
but in this case it's also an optimization. only the global locale
needs synchronization; allocated locales referenced with locale_t
handles are immutable during their lifetimes, and using atomics to
initialize them increases their cost of setup.
commit f3ddd17380 introduced early
relocations and subsequent reprocessing as part of the dynamic linker
bootstrap overhaul, to allow use of arbitrary libc functions before
the main application and libraries are loaded, but only reprocessed
GOT/PLT relocation types.
commit c093e2e820 added reprocessing of
non-GOT/PLT relocations to fix an actual regression that was observed
on powerpc, but only for RELA format tables with out-of-line addends.
REL table (inline addends at the relocation address) reprocessing is
trickier because the first relocation pass clobbers the addends.
this patch extends symbolic relocation reprocessing for libc/ldso to
support all relocation types, whether REL or RELA format tables are
used. it is believed not to alter behavior on any existing archs for
the current dynamic linker and libc code. the motivations for this
change are consistency and future-proofing. it ensures that behavior
does not differ depending on whether REL or RELA tables are used,
which could lead to undetected arch-specific bugs. it also ensures
that, if in the future code depending on additional relocation types
is added to libc.so, either at the source level or as part of the
compiler runtime that gets pulled in (for example, soft-float with TLS
for fenv), the new code will work properly.
the implementation concept is simple: stage 2 of the dynamic linker
counts the number of symbolic relocations in the libc/ldso REL table
and allocates a VLA to save their addends into; stage 3 then uses the
saved addends in place of the inline ones which were clobbered. for
stack safety, a hard limit (currently 4k) is imposed on the number of
such addends; this should be a couple orders of magnitude larger than
the actual need. this number is not a runtime variable that could
break fail-safety; it is constant for a given libc.so build.
this move eliminates a duplicate "by-hand" symbol lookup loop from the
stage-1 code and replaces it with a call to find_sym, which can be
used once we're in stage 2. it reduces the size of the stage 1 code,
which is helpful because stage 1 will become the crt start file for
static-PIE executables, and it will allow stage 3 to access stage 2's
automatic storage, which will be important in an upcoming commit.
the outer-loop approach made sense when we were also processing
DT_JMPREL, which might be in REL or RELA form, to avoid major code
duplication. commit 09db855b35 removed
processing of DT_JMPREL, and in the remaining two tables, the format
(REL or RELA) is known by the name of the table. simply writing two
versions of the loop results in smaller and simpler code.
the DT_JMPREL relocation table necessarily consists entirely of
JMP_SLOT (REL_PLT in internal nomenclature) relocations, which are
symbolic; they cannot be resolved in stage 1, so there is no point in
processing them.
while not a requirement, it's common convention in other iconv
implementations to accept "CHAR" as an alias for nl_langinfo(CODESET),
meaning the encoding used for char[] strings in the current locale,
and also "" as an alternate form. supporting this is not costly and
improves compatibility.
this fixes a regression on powerpc that was introduced in commit
f3ddd17380. global data accesses on
powerpc seem to be using a translation-unit-local GOT filled via
R_PPC_ADDR32 relocations rather than R_PPC_GLOB_DAT. being a non-GOT
relocation type, these were not reprocessed after adding the main
application and its libraries to the chain, causing libc code not to
see copy relocations in the main program, and therefore to use the
pre-copy-relocation addresses for global data objects (like environ).
the motivation for the dynamic linker only reprocessing GOT/PLT
relocation types in stage 3 is that these types always have a zero
addend, making them safe to process again even if the storage for the
addend has been clobbered. other relocation types which can be used
for address constants in initialized data objects may have non-zero
addends which will be clobbered during the first pass of relocation
processing if they're stored inline (REL form) rather than out-of-line
(RELA form).
powerpc generally uses only RELA, so this patch is sufficient to fix
the regression in practice, but is not fully general, and would not
suffice if an alternate toolchain generated REL for powerpc.
if setlocale has not been called, the current locale's messages_name
may be a null pointer. the code path where it's assumed to be non-null
was only reachable if bindtextdomain had already been called, which is
normally not done in programs which do not call setlocale, so the
omitted check went unnoticed.
patch from Void Linux, with description rewritten.
the code being removed used atomics to track whether any threads might
be using a locale other than the current global locale, and whether
any threads might have abstract 8-bit (non-UTF-8) LC_CTYPE active, a
feature which was never committed (still pending). the motivations
were to support early execution prior to setup of the thread pointer,
to partially support systems (ancient kernels) where thread pointer
setup is not possible, and to avoid high performance cost on archs
where accessing the thread pointer may be very slow.
since commit 19a1fe670a, the thread
pointer is always available, so these hacks are no longer needed.
removing them greatly simplifies the affected code.
commit f630df09b1 added logic to handle
the case where __set_thread_area is called more than once by reusing
the GDT slot already in the %gs register, and only setting up a new
GDT slot when %gs is zero. this created a hidden assumption that %gs
is zero when a new process image starts, which is true in practice on
Linux, but does not seem to be documented ABI, and fails to hold under
qemu app-level emulation.
while it would in theory be possible to zero %gs in the entry point
code, this code is shared between static and dynamic binaries, and
dynamic binaries must not clobber the value of %gs already setup by
the dynamic linker.
the alternative solution implemented in this commit simply uses global
data to store the GDT index that's selected. __set_thread_area should
only be called in the initial thread anyway (subsequent threads get
their thread pointer setup by __clone), but even if it were called by
another thread, it would simply read and write back the same GDT index
that was already assigned to the initial thread, and thus (in the x86
memory model) there is no data race.
i386, x86_64, x32, and powerpc all use TLS for stack protector canary
values in the default stack protector ABI, but the location only
matched the ABI on i386 and x86_64. on x32, the expected location for
the canary contained the tid, thus producing spurious mismatches
(resulting in process termination) upon fork. on powerpc, the expected
location contained the stdio_locks list head, so returning from a
function after calling flockfile produced spurious mismatches. in both
cases, the random canary was not present, and a predictable value was
used instead, making the stack protector hardening much less effective
than it should be.
in the current fix, the thread structure has been expanded to have
canary fields at all three possible locations, and archs that use a
non-default location must define a macro in pthread_arch.h to choose
which location is used. for most archs (which lack TLS canary ABI) the
choice does not matter.
the 64-bit push reads not only the 32-bit return address but also the
first 32 signal mask bits. if any were nonzero, the return address
obtained will be invalid.
at some point storage of the return address should probably be moved
to follow the saved mask so that there's plenty room and the same code
can be used on x32 and regular x86_64, but for now I want a fix that
does not risk breaking x86_64, and this simple re-zeroing works.
due to an incorrect return statement in this error case, the
previously blocked cancellation state was not restored and no result
was stored. this could lead to invalid (read) accesses in the caller
resulting in crashes or nonsensical result data in the event of memory
exhaustion.
while the sh port is still experimental and subject to ABI
instability, this is not actually an application/libc boundary ABI
change. it only affects third-party APIs where jmp_buf is used in a
shared structure at the ABI boundary, because nothing anywhere near
the end of the jmp_buf object (which includes the oversized sigset_t)
is accessed by libc.
both glibc and uclibc have 15-slot jmp_buf for sh. presumably the
smaller version was used in musl because the slots for fpu status
register and thread pointer register (gbr) were incorrect and must not
be restored by longjmp, but the size should have been preserved, as
it's generally treated as a libc-agnostic ABI property for the arch,
and having extra slots free in case we ever need them for something is
useful anyway.
both static and dynamic linked versions of the __copy_tls function
have a hidden assumption that the alignment of the beginning or end of
the memory passed is suitable for storing an array of pointers for the
dtv. pthread_create satisfies this requirement except when
libc.tls_size is misaligned, which cannot happen with dynamic linking
due to way update_tls_size computes the total size, but could happen
with static linking and odd-sized TLS.
commit dab441aea2, which made thread
pointer init mandatory for all programs, rendered this store obsolete
by removing the early-return path for static programs with no TLS.
this slightly reduces the code size cost of TLS/thread-pointer for
static linking since __init_tp can be inlined into its only caller and
removed. this is analogous to the handling of __init_libc in
__libc_start_main, where the function only has external linkage when
it needs to be called from the dynamic linker.
the implicit-operand form of fucomip is rejected by binutils 2.19 and
perhaps other versions still in use. writing both operands explicitly
fixes the issue. there is no change to the resulting output.
commit a732e80d33 was the source of this
regression.
use CAS instead of swap since it's lighter for most archs, and keep
EBUSY in the lock value so that the old value obtained by CAS can be
used directly as the return value for pthread_spin_trylock.
the motivation for this change is that the extra declaration (with or
without visibility) using "struct _IO_FILE" instead of "FILE" seems to
trigger a bug in gcc 3.x where it considers the types mismatched.
however, this change also results in slightly better code and it is
valid because (1) these three objects are constant, and (2) applying
the & operator to any of them is invalid C, since they are not even
specified to be objects. thus it does not matter if the application
and libc see different addresses for them, as long as the (initial,
unchanging) value is seen the same by both.
these were hacks to work around toolchains that could not properly
optimize PIC accesses based on visibility and would generate GOT
lookups even for hidden data, which broke the old dynamic linker.
since commit f3ddd17380 it no longer
matters; the dynamic linker does not assume accessibility of this data
until stage 3.
when the non-stub duplocale code was added as part of the locale
framework in commit 0bc03091bb, the old
code to memcpy the old locale object to the new one was left behind.
the conditional for the memcpy no longer makes sense, because the
conditions are now always-true when it's reached, and the memcpy is
wrong because it clobbers the new->messages_name pointer setup just
above.
since the messages_name and ctype_utf8 members have already been
copied, all that remains is the cat[] array. these pointers are
volatile, so using memcpy to copy them is formally wrong; use a for
loop instead.
the first switch already returns in the F_SETLKW code path so it need
not be handled in the second switch. moreover the code in the second
switch is wrong for the F_SETLKW command: it's not cancellable.
the leak was found by static analysis (reported by Alexander Monakov),
not tested/observed, but seems to have occured both when failing due
to O_EXCL, and in a race condition with O_CREAT but not O_EXCL where a
semaphore by the same name was created concurrently.
the allocating path which can fail is for dynamic TLS, which can only
occur at runtime, and the check for runtime was already made in the
outer conditional.
commit 637dd2d383 introduced the checks
for RTLD_DEFAULT and RTLD_NEXT here, claiming they fixed a regression,
but the above conditional block clearly already covered these cases,
and removing the checks produces no difference in the generated code.
this is implemented via the build system and does not affect source
files. the idea is to use protected or hidden visibility to prevent
the compiler from pessimizing function calls within a shared (or
position-independent static) libc in the form of overhead setting up
for a call through the PLT. the ld-time symbol binding via the
-Bsymbolic-functions option already optimized out the PLT itself, but
not the code in the caller needed to support a call through the PLT.
on some archs this overhead can be substantial; on others it's
trivial.
these are perfectly fine with ld-time symbol binding, but otherwise
result in textrels. they cannot be replaced with @PLT jump targets
because the PLT thunks require a GOT register to be setup, so use a
hidden alias instead.
these are perfectly fine with ld-time symbol binding, but if the calls
go through a PLT thunk, they are invalid because the caller does not
setup a GOT register. use a hidden alias to bypass the issue.
none of these are actual textrels because of ld-time binding performed
by -Bsymbolic-functions, but I'm changing them with the goal of making
ld-time binding purely an optimization rather than relying on it for
semantic purposes.
in the case of memmove's call to memcpy, making it explicit that the
memmove asm is assuming the forward-copying behavior of the memcpy asm
is desirable anyway; in case memcpy is ever changed, the semantic
mismatch would be apparent while editing memmcpy.s.
this fixes truncation of error messages containing long pathnames or
symbol names.
the dlerror state was previously required by POSIX to be global. the
resolution of bug 97 relaxed the requirements to allow thread-safe
implementations of dlerror with thread-local state and message buffer.
these functions are never called directly; only their addresses are
used, so PLT indirections should never happen unless a broken
application tries to redefine them, but it's still best to make them
hidden.
Rich Felker
Szabolcs Nagy
Alexander Monakov