kernel: HTTPS through wget, half-close, and /proc/<pid>/fd
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:
parent
906318f2b6
commit
8351e0323b
15 changed files with 434 additions and 71 deletions
|
|
@ -574,11 +574,15 @@ exec_elf(char *path, struct argsrc *av, struct argsrc *ev)
|
|||
// first delivered signal jump to a stale address in the new image.
|
||||
// mmap_next likewise belongs to the old address space; left alone it
|
||||
// could hand out mmap ranges overlapping the new image's layout.
|
||||
// Ignored signals stay ignored, and the mask is inherited (execve(2)):
|
||||
// neither points into the old image. `nohup` and every program that ignores
|
||||
// SIGPIPE before running a helper depend on it.
|
||||
memset(myproc()->vma, 0, sizeof(myproc()->vma));
|
||||
memset(myproc()->sig_handler, 0, sizeof(myproc()->sig_handler));
|
||||
for (int i = 0; i < 32; i++)
|
||||
if (myproc()->sig_handler[i] != 1)
|
||||
myproc()->sig_handler[i] = 0;
|
||||
memset(myproc()->sig_restorer, 0, sizeof(myproc()->sig_restorer));
|
||||
myproc()->fs_base = 0;
|
||||
myproc()->signal_mask = 0;
|
||||
myproc()->pending_signals = 0;
|
||||
myproc()->in_signal = 0;
|
||||
myproc()->sigsusp_active = 0;
|
||||
|
|
|
|||
Loading…
Reference in a new issue