linux: improve exception-signal handling

First, we use an alternate stack for signal handling now. The stack is
shared among all threads of the component, which is okay as we only
handle exceptions with log output and pass on to the default handler
(that terminates the execution). The primary motivation for the
alternate stack is the detection of SIGSEGV due to stack overflows.

Also, hybrid components now handle exception signals by logging and the
support for multi-threaded applications was improved.

Fixes #1935
This commit is contained in:
Christian Helmuth
2016-04-08 18:24:15 +02:00
parent 051e84c4b4
commit 14f1ac497e
5 changed files with 40 additions and 7 deletions

View File

@ -258,10 +258,10 @@ inline int lx_sigaction(int signum, void (*handler)(int))
* when leaving the signal handler and it should call the rt_sigreturn syscall.
*/
enum { SA_RESTORER = 0x04000000 };
act.flags = SA_RESTORER;
act.flags = SA_RESTORER | SA_ONSTACK;
act.restorer = lx_restore_rt;
#else
act.flags = 0;
act.flags = SA_ONSTACK;
act.restorer = 0;
#endif
lx_sigemptyset(&act.mask);
@ -282,6 +282,17 @@ inline int lx_tgkill(int pid, int tid, int signal)
}
/**
* Alternate signal stack (handles also SIGSEGV in a safe way)
*/
inline int lx_sigaltstack(void *signal_stack, Genode::size_t stack_size)
{
stack_t stack { signal_stack, 0, stack_size };
return lx_syscall(SYS_sigaltstack, &stack, nullptr);
}
inline int lx_create_thread(void (*entry)(), void *stack, void *arg)
{
int flags = CLONE_VM | CLONE_FS | CLONE_FILES | CLONE_SIGHAND