this patch improves the correctness, simplicity, and size of
cancellation-related code. modulo any small errors, it should now be
completely conformant, safe, and resource-leak free.
the notion of entering and exiting cancellation-point context has been
completely eliminated and replaced with alternative syscall assembly
code for cancellable syscalls. the assembly is responsible for setting
up execution context information (stack pointer and address of the
syscall instruction) which the cancellation signal handler can use to
determine whether the interrupted code was in a cancellable state.
these changes eliminate race conditions in the previous generation of
cancellation handling code (whereby a cancellation request received
just prior to the syscall would not be processed, leaving the syscall
to block, potentially indefinitely), and remedy an issue where
non-cancellable syscalls made from signal handlers became cancellable
if the signal handler interrupted a cancellation point.
x86_64 asm is untested and may need a second try to get it right.
some of these definitions were just plain wrong, others based on
outdated ancient "non-64" versions of the kernel interface.
as much as possible has now been moved out of bits/*
these changes break abi (the old abi for these functions was wrong),
but since they were not working anyway it can hardly matter.
the basic idea is that the only things in alltypes.h should be types
that either vary from system to system (in practice, not just in
theoretical la-la land - this is the implementation so we choose what
constraints we want to impose on ports) or which are needed by
multiple system headers.
actually FLT_ROUNDS needs to expand to a static inline function that
obtains the current rounding mode and returns it, but that will be
added later with fenv.h stuff.
instead of allocating a userspace structure for signal-based timers,
simply use the kernel timer id. we use the fact that thread pointers
will always be zero in the low bit (actually more) to encode integer
timerid values as pointers.
also, this change ensures that the timer_destroy syscall has completed
before the library timer_destroy function returns, in case it matters.
glibc made the ridiculous choice to use pass-by-register calling
convention for these functions, which is impossible to duplicate
directly on non-gcc compilers. instead, we use ugly asm to wrap and
convert the calling convention. presumably this works with every
compiler anyone could potentially want to use.
this commit addresses two issues:
1. a race condition, whereby a cancellation request occurring after a
syscall returned from kernelspace but before the subsequent
CANCELPT_END would cause cancellable resource-allocating syscalls
(like open) to leak resources.
2. signal handlers invoked while the thread was blocked at a
cancellation point behaved as if asynchronous cancellation mode wer in
effect, resulting in potentially dangerous state corruption if a
cancellation request occurs.
the glibc/nptl implementation of threads shares both of these issues.
with this commit, both are fixed. however, cancellation points
encountered in a signal handler will not be acted upon if the signal
was received while the thread was already at a cancellation point.
they will of course be acted upon after the signal handler returns, so
in real-world usage where signal handlers quickly return, it should
not be a problem. it's possible to solve this problem too by having
sigaction() wrap all signal handlers with a function that uses a
pthread_cleanup handler to catch cancellation, patch up the saved
context, and return into the cancellable function that will catch and
act upon the cancellation. however that would be a lot of complexity
for minimal if any benefit...
with this patch, the syscallN() functions are no longer needed; a
variadic syscall() macro allows syscalls with anywhere from 0 to 6
arguments to be made with a single macro name. also, manually casting
each non-integer argument with (long) is no longer necessary; the
casts are hidden in the macros.
some source files which depended on being able to define the old macro
SYSCALL_RETURNS_ERRNO have been modified to directly use __syscall()
instead of syscall(). references to SYSCALL_SIGSET_SIZE and SYSCALL_LL
have also been changed.
x86_64 has not been tested, and may need a follow-up commit to fix any
minor bugs/oversights.
this commit shuffles around the location of syscall definitions so
that we can make a syscall() library function with both SYS_* and
__NR_* style syscall names available to user applications, provides
the syscall() library function, and optimizes the code that performs
the actual inline syscalls in the library itself.
previously on i386 when built as PIC (shared library), syscalls were
incurring bus lock (lock prefix) overhead at entry and exit, due to
the way the ebx register was being loaded (xchg instruction with a
memory operand). now the xchg takes place between two registers.
further cleanup to arch/$(ARCH)/syscall.h is planned.
only the structures, not the functions from ucontext.h, are supported
at this point. the main goal of this commit is to make modern gcc with
dwarf2 unwinding build without errors.
honestly, it probably doesn't matter how we define these as long as
they have members with the right names to prevent errors while
compiling libgcc. the only time they will be used is for propagating
exceptions across signal-handler boundaries, which invokes undefined
behavior anyway. but as-is, they're probably correct and may be useful
to various low-level applications dealing with virtualization, jit
code generation, and so on...
this allows sys/types.h to provide the pthread types, as required by
POSIX. this design also facilitates forcing ABI-compatible sizes in
the arch-specific alltypes.h, while eliminating the need for
developers changing the internals of the pthread types to poke around
with arch-specific headers they may not be able to test.
this was originally written for an early draft of the library where
non-standard functions would reside in a static library separate from
the shared libc.so, which would implement a pure standard. the idea
was not to depend on an implementation-dependent __syscall_ret
function in the main libc. but it turned out to be better to put
everything in a single library for both static and dynamic linking
uses, and thus the (incomplete) remnants of this feature were just
enlarging the source and binary.
Rich Felker