GNU/Weeb Mailing List <[email protected]>
 help / color / mirror / Atom feed
From: Ammar Faizi <[email protected]>
To: Willy Tarreau <[email protected]>
Cc: Ammar Faizi <[email protected]>,
	Shuah Khan <[email protected]>,
	"Paul E. McKenney" <[email protected]>,
	Gilang Fachrezy <[email protected]>,
	Alviro Iskandar Setiawan <[email protected]>,
	GNU/Weeb Mailing List <[email protected]>,
	Linux Kernel Mailing List <[email protected]>,
	Linux Kselftest Mailing List <[email protected]>
Subject: [PATCH v3 1/5] nolibc/sys: Implement `sigaction(2)` function
Date: Sun,  8 Jan 2023 20:59:00 +0700	[thread overview]
Message-ID: <[email protected]> (raw)
In-Reply-To: <[email protected]>

From: Ammar Faizi <[email protected]>

This commit adds the initial implementation of nolibc `sigaction()`
function. Currently, this implementation is only available on the
x86-64 arch.

`sigaction()` needs an architecture-dependent "signal trampoline"
function that invokes __rt_sigreturn syscall to resume the process
after a signal gets handled.

The "signal trampoline" function is called `__restore_rt` in this
implementation. The naming `__restore_rt` is important for GDB. It also
has to be given a special optimization attribute "omit-frame-pointer"
to prevent the compiler from creating a stack frame that makes the
`%rsp` value no longer points to the `struct rt_sigframe` that the
kernel constructed.

Link: https://lore.kernel.org/lkml/[email protected]
Suggested-by: Willy Tarreau <[email protected]>
Signed-off-by: Ammar Faizi <[email protected]>
---

Side note:
This has been tested on x86-64 arch and `__restore_rt` generates the
correct code. The `__restore_rt` codegen correctness on other
architectures need to be evaluated as well. If it can't generate the
correct code, it has to be written in inline Assembly.

 tools/include/nolibc/sys.h | 72 ++++++++++++++++++++++++++++++++++++++
 1 file changed, 72 insertions(+)

diff --git a/tools/include/nolibc/sys.h b/tools/include/nolibc/sys.h
index acf7cf438010..7d594155e77f 100644
--- a/tools/include/nolibc/sys.h
+++ b/tools/include/nolibc/sys.h
@@ -1047,6 +1047,78 @@ pid_t setsid(void)
 	return ret;
 }
 
+typedef void (*sighandler_t)(int sig);
+
+/*
+ * int sigaction(int signum, const struct sigaction *act, struct sigaction *oldact);
+ */
+
+static __attribute__((unused))
+int sys_sigaction(int signum, const struct sigaction *act,
+		  struct sigaction *oldact)
+{
+	return my_syscall4(__NR_rt_sigaction, signum, act, oldact,
+			   sizeof(sigset_t));
+}
+
+__attribute__((weak,unused,noreturn,optimize("omit-frame-pointer"),section(".text.__restore_rt")))
+void __restore_rt(void)
+{
+	my_syscall0(__NR_rt_sigreturn);
+	__builtin_unreachable();
+}
+
+static __attribute__((unused))
+int sigaction(int signum, const struct sigaction *act, struct sigaction *oldact)
+{
+	struct sigaction act2 = *act;
+	int ret;
+
+	/*
+	 * On Linux x86-64, libc's sigaction() always sets the
+	 * @act->sa_restorer when the caller passes a NULL.
+	 *
+	 * @act->sa_restorer is an arch-specific function used
+	 * as a "signal trampoline".
+	 *
+	 * @act->sa_handler is a signal handler provided by the
+	 * user.
+	 *
+	 * When the handled signal is caught, the %rip jumps to
+	 * @act->sa_handler with user stack already set by the
+	 * kernel as below:
+	 *
+	 *         |--------------------|
+	 * %rsp -> |  act->sa_restorer  | (return address)
+	 *         |--------------------|
+	 *         | struct rt_sigframe | (process context info)
+	 *         |                    |
+	 *         |                    |
+	 *          ....................
+	 *
+	 * Once this signal handler executes the "ret" instruction,
+	 * the %rip jumps to @act->sa_restorer. The sa_restorer
+	 * function has to invoke the __rt_sigreturn syscall with
+	 * %rsp pointing to the `struct rt_sigframe` that the kernel
+	 * constructed previously to resume the process.
+	 *
+	 * `struct rt_sigframe` contains the registers' value before
+	 * the signal is caught.
+	 *
+	 */
+	if (!act2.sa_restorer) {
+		act2.sa_flags |= SA_RESTORER;
+		act2.sa_restorer = __restore_rt;
+	}
+
+	ret = sys_sigaction(signum, &act2, oldact);
+	if (ret < 0) {
+		SET_ERRNO(-ret);
+		ret = -1;
+	}
+	return ret;
+}
+
 
 /*
  * int stat(const char *path, struct stat *buf);
-- 
Ammar Faizi


  reply	other threads:[~2023-01-08 13:59 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-01-08 13:58 [PATCH v3 0/5] nolibc signal handling support Ammar Faizi
2023-01-08 13:59 ` Ammar Faizi [this message]
2023-01-08 13:59 ` [PATCH v3 2/5] nolibc/sys: Implement `signal(2)` function Ammar Faizi
2023-01-08 13:59 ` [PATCH v3 3/5] selftests/nolibc: Add `fork(2)` selftest Ammar Faizi
2023-01-08 13:59 ` [PATCH v3 4/5] selftests/nolibc: Add `sigaction(2)` selftest Ammar Faizi
2023-01-08 13:59 ` [PATCH v3 5/5] selftests/nolibc: Add `signal(2)` selftest Ammar Faizi
2023-01-08 17:58 ` [PATCH v3 0/5] nolibc signal handling support Willy Tarreau
2023-01-08 18:31   ` Ammar Faizi
2023-01-08 18:49     ` Willy Tarreau
2023-01-15 16:01       ` Ammar Faizi
2023-01-15 17:06         ` Alviro Iskandar Setiawan
2023-01-17 23:17           ` Ammar Faizi

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox