From: Cyril Hrubis <chrubis@suse.cz>
To: ltp@lists.linux.it
Subject: [LTP] [PATCH v2] syscalls/recvmsg03.c: add new testcase
Date: Wed, 2 Nov 2016 14:06:47 +0100 [thread overview]
Message-ID: <20161102130647.GB22840@rei.lan> (raw)
In-Reply-To: <1478064173-5704-1-git-send-email-yangx.jy@cn.fujitsu.com>
Hi!
> +static void server(void)
> +{
> + int sock_fd, sock_fd2;
> + static char recv_buf[128];
> + struct sockaddr_in server_addr;
> + struct sockaddr_in from_addr;
> + struct msghdr msg;
> + struct iovec iov;
> +
> + sock_fd2 = SAFE_SOCKET(AF_RDS, SOCK_SEQPACKET, 0);
> + sock_fd = sock_fd2;
> +
> + memset(&server_addr, 0, sizeof(server_addr));
> + server_addr.sin_family = AF_INET;
> + server_addr.sin_addr.s_addr = inet_addr("127.0.0.1");
> + server_addr.sin_port = htons(4000);
> +
> + SAFE_BIND(sock_fd2, (struct sockaddr *) &server_addr, sizeof(server_addr));
> +
> + msg.msg_name = &from_addr;
> + msg.msg_namelen = sizeof(from_addr) + 16;
> + msg.msg_iov = &iov;
> + msg.msg_iovlen = 1;
> + msg.msg_iov->iov_base = recv_buf;
> + msg.msg_iov->iov_len = 128;
> + msg.msg_control = 0;
> + msg.msg_controllen = 0;
> + msg.msg_flags = 0;
> +
> + TST_CHECKPOINT_WAKE(0);
> +
> + TEST(recvmsg(sock_fd2, &msg, 0));
> + if (TEST_RETURN == -1) {
> + tst_res(TFAIL | TERRNO,
> + "recvmsg() failed to recvice data from client");
> + goto end;
> + }
> +
> + if (msg.msg_namelen != sizeof(from_addr)) {
> + tst_res(TFAIL, "msg_namelen was set to %u incorrectly, "
> + "expected %lu", msg.msg_namelen, sizeof(from_addr));
> + goto end;
> + }
> +
> + if (sock_fd2 != sock_fd) {
> + tst_res(TFAIL, "sock_fd was destroyed");
> + goto end;
> + }
> +
> + tst_res(TPASS, "msg_namelen was set to %u correctly and sock_fd was "
> + "not destroyed", msg.msg_namelen);
> +
> +end:
> + SAFE_CLOSE(sock_fd2);
I'm a bit confused here, which one of the sock_fd/sock_fd2 is destroyed?
Looking at the original code in the kernel commit the sock_fd there is
stored on the stack directly after the sockaddr_in from_addr so I guess
that the kernel will actually write a few bytes after the end of
from_addr structure in this case, which will rewrite the msghrd msg in
your code. Does the test actually fail on kernel without the fix?
> +}
> +
> +static void verify_recvmsg(void)
> +{
> + pid_t pid;
> +
> + pid = SAFE_FORK();
> + if (pid == 0) {
> + TST_CHECKPOINT_WAIT(0);
> + client();
> + } else {
> + server();
> + SAFE_WAIT(NULL);
We should rather call tst_reap_children() in this case instead of the
WAIT since otherwise TBROK from the client() function will not get
propagated.
> + }
> +}
> +
> +static struct tst_test test = {
> + .tid = "recvmsg03",
> + .forks_child = 1,
> + .needs_checkpoints = 1,
> + .setup = setup,
> + .test_all = verify_recvmsg
> +};
> --
> 1.8.3.1
>
>
>
--
Cyril Hrubis
chrubis@suse.cz
next prev parent reply other threads:[~2016-11-02 13:06 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-10-31 11:23 [LTP] [PATCH] syscalls/recvmsg03.c: add new testcase Xiao Yang
2016-10-31 13:39 ` Cyril Hrubis
2016-11-01 2:24 ` [LTP] [PATCH v2] " Xiao Yang
2016-11-02 5:34 ` Xiao Yang
2016-11-02 5:22 ` Xiao Yang
2016-11-02 13:06 ` Cyril Hrubis [this message]
2016-11-03 7:49 ` [LTP] [PATCH v3] " Xiao Yang
2016-11-11 1:30 ` Xiao Yang
2016-11-15 14:04 ` Cyril Hrubis
2016-11-16 4:34 ` Xiao Yang
2016-11-16 5:37 ` Xiao Yang
2016-12-13 7:50 ` Xiao Yang
2017-03-15 15:35 ` Cyril Hrubis
2016-11-07 10:52 ` [LTP] [PATCH v2] " Xiao Yang
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 \
--in-reply-to=20161102130647.GB22840@rei.lan \
--to=chrubis@suse.cz \
--cc=ltp@lists.linux.it \
/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