nexoniarz/SchOS
nexoniarz/SchOS
Watch
1
0
Fork
You've already forked SchOS
 0
Code Issues Pull requests Projects Releases 1 Packages Wiki Activity Actions

kernel: HTTPS through wget, half-close, and /proc/<pid>/fd

Browse source 
busybox wget hands TLS to ssl_client over a socketpair and then does
shutdown(SHUT_WR). That cut the AF_UNIX link both ways, so ssl_client's
reply was a SIGPIPE and every https:// download hung. SHUT_WR is now a
half-close: the peer reads EOF and can still answer. TCP sockets get
shutdown() at all (tcp_shutdown for the FIN); it used to do nothing.

- exec keeps ignored signals and the signal mask, as execve(2) says.
- /proc/<pid>/fd: a link per descriptor -- the path it was opened by,
  socket:[n], pipe:[n], anon_inode:[eventpoll]. /proc/self too. The
  global inode bit moves above pid << 8; at 0x10000 it was /proc/256.
- fstat on a socket or pipe answers S_IFSOCK / S_IFIFO instead of EINVAL.
- poll, select and epoll wake as soon as a pipe, socket or input device
  changes, not on the next tick.
- Pipes are 64 KB of lazily allocated pages, copied in blocks, with
  PIPE_BUF atomicity: 300 MB through `yes | head | wc` in 7 s.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01S1giNvrHTrPMPw7V6x1kQ4
This commit is contained in:
Nexoniarz 2026-10-24 20:28:03 +02:00
commit 8351e0323b
15 changed files with 434 additions and 71 deletions
Download patch file Download diff file Expand all files Collapse all files

32
kernel/sys/sys_event.c
View file

@ -113,19 +113,13 @@ clock_ns(int clock)
return sec * 1000000000ULL + usec * 1000ULL;
}
// One tick, or until a signal or kill makes waiting pointless.
// One tick, a poll_kick, or a signal or kill that makes waiting pointless:
// poll's wait, so data arriving wakes an epoll_wait as it wakes a poll.
// Returns 1 for a tick, 0 for a kick, -EINTR.
static int
ev_block(void)
{
struct proc *p = myproc();
if (p->killed || (p->pending_signals & ~p->signal_mask))
return -EINTR;
acquire(&tickslock);
uint t0 = ticks;
while (ticks == t0 && !p->killed && !(p->pending_signals & ~p->signal_mask))
sleep(&ticks, &tickslock);
release(&tickslock);
return 0;
return poll_wait_tick();
}
static int
@ -178,6 +172,19 @@ ev_arg(int n, int kind, struct file **pf)
return f->ev;
}
// The name Linux gives each kind in /proc/<pid>/fd.
const char *
ev_kind_name(struct evobj *o)
{
switch (o ? o->kind : 0) {
case EV_EVENTFD: return "anon_inode:[eventfd]";
case EV_TIMERFD: return "anon_inode:[timerfd]";
case EV_SIGNALFD: return "anon_inode:[signalfd]";
case EV_EPOLL: return "anon_inode:[eventpoll]";
}
return "anon_inode:[event]";
}
// Called by fileclose() for the last reference.
void
ev_close(struct evobj *o)
@ -714,11 +721,12 @@ epoll_wait_common(int timeout_ms)
got = ep_ready(o, buf, max);
if (got > 0 || ticks_left == 0)
break;
if (ev_block() < 0) {
int w = ev_block();
if (w < 0) {
got = -EINTR;
break;
}
if (ticks_left > 0)
if (w > 0 && ticks_left > 0)
ticks_left--;
}
if (got > 0 && copy_to_user(uevents, buf, got * sizeof(*buf)) < 0)