Differential D19058 Diff 53521 lib/libc/sys/sigaction.c

Changeset View

Standalone View

lib/libc/sys/sigaction.c

	Show All 28 Lines
	* OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,			* OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
	* EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.			* EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
	*/			*/

	#include <sys/cdefs.h>			#include <sys/cdefs.h>
	__FBSDID("$FreeBSD$");			__FBSDID("$FreeBSD$");

	#include <sys/types.h>			#include <sys/types.h>
				#include <machine/atomic.h>
	#include <signal.h>			#include <signal.h>
				#include <stdlib.h>
	#include "libc_private.h"			#include "libc_private.h"

	__weak_reference(__sys_sigaction, __sigaction);			__weak_reference(__sys_sigaction, __sigaction);
	__weak_reference(sigaction, __libc_sigaction);			__weak_reference(sigaction, __libc_sigaction);

				/*
				* Used to track signals that are given a non-ignored handler so that other
				* parts of libc, notably posix_spawn, can easily and safely reset these to
				* default handler before unblocking signals again and proceeding without
				* the possibility of a signal handler altering process state in the middle of
				* the vfork.
				*
				* This set of used signals should not be treated as an ultimate source of
				* truth, but rather as a 'best guess'.
				*/
				static uint32_t _used_sigs[_SIG_WORDS];
				kibUnsubmitted Not Done Inline Actions There is no need in the underscore prefix. The symbol is static. Why you cannot use sigset_t ? SIGCANCEL should be added unconditionally to the set. kib: There is no need in the underscore prefix. The symbol is static. Why you cannot use sigset_t ?
				kevansAuthorUnsubmitted Done Inline Actions I hadn't spent any time considering any concurrency issues that may happen here, so I opted for the bare minimum uint32[_SIG_WORDS] that I can operate on atomically -- a luxury not provided by sigset operations last I had looked. kevans: I hadn't spent any time considering any concurrency issues that may happen here, so I opted for…

				void
				__libc_used_signal_reset(void)
				{
				int sig;
				struct sigaction sa;

				sa.sa_handler = SIG_DFL;
				jillesUnsubmitted Not Done Inline Actions `sa_flags` and `sa_mask` should be initialized as well. jilles: `sa_flags` and `sa_mask` should be initialized as well.
				kevansAuthorUnsubmitted Done Inline Actions I thought I had a `memset` here -- must have been lost in transit across machines. =( Will address this in another revision. kevans: I thought I had a `memset` here -- must have been lost in transit across machines. =( Will…
				for (sig = 1; sig < _SIG_MAXSIG; ++sig) {
				if ((_used_sigs[_SIG_WORD(sig)] & _SIG_BIT(sig)) != 0)
				__libc_sigaction(sig, &sa, NULL);
				jillesUnsubmitted Not Done Inline Actions Calling libthr's version prevents resetting SIGCANCEL/SIGTHR and does additional sigprocmask and rwlock calls. jilles: Calling libthr's version prevents resetting SIGCANCEL/SIGTHR and does additional sigprocmask…
				}
				}

	#pragma weak sigaction			#pragma weak sigaction
	int			int
	sigaction(int sig, const struct sigaction act, struct sigaction oact)			sigaction(int sig, const struct sigaction act, struct sigaction oact)
	{			{

				if (act != NULL && _SIG_VALID(sig) &&
				((act->sa_flags & SA_SIGINFO) != 0 \|\| act->sa_handler != SIG_IGN))
				kibUnsubmitted Not Done Inline Actions SIG_DFL can be ignored as well. You may also clear bits when SIG_IGN and SIG_DFL are specified. kib: SIG_DFL can be ignored as well. You may also clear bits when SIG_IGN and SIG_DFL are specified.
				jillesUnsubmitted Not Done Inline Actions Clearing bits may be incorrect when two threads issue sigaction for the same signal at the same time, one setting a handler and one setting the default action. libthr does not guarantee that the previously installed handler cannot be called after a sigaction with SIG_DFL returns, but does appear to lock the change properly otherwise. jilles: Clearing bits may be incorrect when two threads issue sigaction for the same signal at the same…
				kevansAuthorUnsubmitted Done Inline Actions The initial concern I had is invalid when I address your concerns about using libthr versions of sig* operations, but I did not want to clear these initially because I didn't want to clobber `_used_sigs` in the vfork parent when the child resets everything. kevans: The initial concern I had is invalid when I address your concerns about using libthr versions…
				atomic_set_32(&_used_sigs[_SIG_WORD(sig)], _SIG_BIT(sig));
				jillesUnsubmitted Not Done Inline Actions Since libthr's setting of SIGCANCEL/SIGTHR does not pass through here, it needs to be accounted for differently. jilles: Since libthr's setting of SIGCANCEL/SIGTHR does not pass through here, it needs to be accounted…
				kevansAuthorUnsubmitted Done Inline Actions Is there any reason I shouldn't have `__libc_used_signal_reset` just reset these ones unconditionally? kevans: Is there any reason I shouldn't have `__libc_used_signal_reset` just reset these ones…
	return (((int ()(int, const struct sigaction , struct sigaction *))			return (((int ()(int, const struct sigaction , struct sigaction *))
	__libc_interposing[INTERPOS_sigaction])(sig, act, oact));			__libc_interposing[INTERPOS_sigaction])(sig, act, oact));
	}			}