From: Paul Holzinger <pholzing@redhat.com>
To: Stefano Brivio <sbrivio@redhat.com>, passt-dev@passt.top
Cc: David Gibson <david@gibson.dropbear.id.au>
Subject: Re: [PATCH v2] passt, util: Close any open file that the parent might have leaked
Date: Wed, 7 Aug 2024 11:34:34 +0200 [thread overview]
Message-ID: <59e8daad-cd5a-4723-8ae8-807b72ec10c2@redhat.com> (raw)
In-Reply-To: <20240807082745.1939437-1-sbrivio@redhat.com>
On 07/08/2024 10:27, Stefano Brivio wrote:
> If a parent accidentally or due to implementation reasons leaks any
> open file, we don't want to have access to them, except for the file
> passed via --fd, if any.
>
> This is the case for Podman when Podman's parent leaks files into
> Podman: it's not practical for Podman to close unrelated files before
> starting pasta, as reported by Paul.
>
> Use close_range(2) to close all open files except for standard streams
> and the one from --fd.
>
> Given that parts of conf() depend on other files to be already opened,
> such as the epoll file descriptor, we can't easily defer this to a
> more convenient point, where --fd was already parsed. Introduce a
> minimal, duplicate version of --fd parsing to keep this simple.
>
> Suggested-by: Paul Holzinger <pholzing@redhat.com>
> Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
> ---
> v2: Move call to close_open_files() to isolate_initial()
>
> conf.c | 1 +
> isolation.c | 12 +++++++++---
> isolation.h | 2 +-
> passt.c | 2 +-
> util.c | 36 ++++++++++++++++++++++++++++++++++++
> util.h | 1 +
> 6 files changed, 49 insertions(+), 5 deletions(-)
>
> diff --git a/conf.c b/conf.c
> index 14d8ece..89f5b3d 100644
> --- a/conf.c
> +++ b/conf.c
> @@ -1260,6 +1260,7 @@ void conf(struct ctx *c, int argc, char **argv)
> c->tcp.fwd_in.mode = c->tcp.fwd_out.mode = FWD_UNSET;
> c->udp.fwd_in.mode = c->udp.fwd_out.mode = FWD_UNSET;
>
> + optind = 1;
> do {
> name = getopt_long(argc, argv, optstring, options, NULL);
>
> diff --git a/isolation.c b/isolation.c
> index 4956d7e..45fba1e 100644
> --- a/isolation.c
> +++ b/isolation.c
> @@ -29,7 +29,8 @@
> *
> * Executed immediately after startup, drops capabilities we don't
> * need at any point during execution (or which we gain back when we
> - * need by joining other namespaces).
> + * need by joining other namespaces), and closes any leaked file we
> + * might have inherited from the parent process.
> *
> * 2. isolate_user()
> * =================
> @@ -166,14 +167,17 @@ static void clamp_caps(void)
> }
>
> /**
> - * isolate_initial() - Early, config independent self isolation
> + * isolate_initial() - Early, mostly config independent self isolation
> + * @argc: Argument count
> + * @argv: Command line options: only --fd (if present) is relevant here
> *
> * Should:
> * - drop unneeded capabilities
> + * - close all open files except for standard streams and the one from --fd
> * Musn't:
> * - remove filesytem access (we need to access files during setup)
> */
> -void isolate_initial(void)
> +void isolate_initial(int argc, char **argv)
> {
> uint64_t keep;
>
> @@ -207,6 +211,8 @@ void isolate_initial(void)
> keep |= BIT(CAP_SETFCAP) | BIT(CAP_SYS_PTRACE);
>
> drop_caps_ep_except(keep);
> +
> + close_open_files(argc, argv);
> }
>
> /**
> diff --git a/isolation.h b/isolation.h
> index 846b2af..80bb68d 100644
> --- a/isolation.h
> +++ b/isolation.h
> @@ -7,7 +7,7 @@
> #ifndef ISOLATION_H
> #define ISOLATION_H
>
> -void isolate_initial(void);
> +void isolate_initial(int argc, char **argv);
> void isolate_user(uid_t uid, gid_t gid, bool use_userns, const char *userns,
> enum passt_modes mode);
> int isolate_prefork(const struct ctx *c);
> diff --git a/passt.c b/passt.c
> index ea5bece..4b3c306 100644
> --- a/passt.c
> +++ b/passt.c
> @@ -211,7 +211,7 @@ int main(int argc, char **argv)
>
> arch_avx2_exec(argv);
>
> - isolate_initial();
> + isolate_initial(argc, argv);
>
> c.pasta_netns_fd = c.fd_tap = c.pidfile_fd = -1;
>
> diff --git a/util.c b/util.c
> index 07fb21c..bd65b5a 100644
> --- a/util.c
> +++ b/util.c
> @@ -26,6 +26,7 @@
> #include <errno.h>
> #include <stdbool.h>
> #include <linux/errqueue.h>
> +#include <getopt.h>
>
> #include "util.h"
> #include "iov.h"
> @@ -694,3 +695,38 @@ const char *str_ee_origin(const struct sock_extended_err *ee)
>
> return "<invalid>";
> }
> +
> +/**
> + * close_open_files() - Close leaked files, but not --fd, stdin, stdout, stderr
> + * @argc: Argument count
> + * @argv: Command line options, as we need to skip any file given via --fd
> + */
> +void close_open_files(int argc, char **argv)
> +{
> + const struct option optfd[] = { { "fd", required_argument, NULL, 'F' },
> + { 0 },
> + };
> + long fd = -1;
> + int name;
> +
> + do {
> + name = getopt_long(argc, argv, ":F", optfd, NULL);
> +
> + if (name == 'F') {
> + errno = 0;
> + fd = strtol(optarg, NULL, 0);
> +
> + if (fd < 0 || errno)
> + die("Invalid --fd: %s", optarg);
> + }
> + } while (name != -1);
> +
> + if (fd == -1) {
> + if (close_range(3, ~0U, CLOSE_RANGE_UNSHARE))
> + die_perror("Failed to close files leaked by parent");
> + } else {
> + if (close_range(3, fd - 1, CLOSE_RANGE_UNSHARE) ||
> + close_range(fd + 1, ~0U, CLOSE_RANGE_UNSHARE))
> + die_perror("Failed to close files leaked by parent");
This will fail for fd == 3 as the first range is 3 - 2 in that case
which causes EINVAL which could be a common choice for the extra passed fd.
On the other end of the range it is the same issue, you seem to parse
the fd as long so allows values > 4294967295 which then overflows when
passed to close_range as uint which causes issues and even allows us to
close stderr streams. Less likely that users would pass such a number
but no reason to allow it in the first place.
$ strace -e close_range ./pasta --config-net --fd 4294967296 echo test
close_range(3, 4294967295, CLOSE_RANGE_UNSHARE) = 0
close_range(1, 4294967295, CLOSE_RANGE_UNSHARE) = 0
> + }
> +}
> diff --git a/util.h b/util.h
> index 83d2b53..cb4d181 100644
> --- a/util.h
> +++ b/util.h
> @@ -183,6 +183,7 @@ int __daemon(int pidfile_fd, int devnull_fd);
> int fls(unsigned long x);
> int write_file(const char *path, const char *buf);
> int write_remainder(int fd, const struct iovec *iov, size_t iovcnt, size_t skip);
> +void close_open_files(int argc, char **argv);
>
> /**
> * af_name() - Return name of an address family
--
Paul Holzinger
next prev parent reply other threads:[~2024-08-07 9:34 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-08-07 8:27 [PATCH v2] passt, util: Close any open file that the parent might have leaked Stefano Brivio
2024-08-07 8:51 ` David Gibson
2024-08-07 9:34 ` Paul Holzinger [this message]
2024-08-07 11:10 ` Stefano Brivio
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=59e8daad-cd5a-4723-8ae8-807b72ec10c2@redhat.com \
--to=pholzing@redhat.com \
--cc=david@gibson.dropbear.id.au \
--cc=passt-dev@passt.top \
--cc=sbrivio@redhat.com \
/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.
Code repositories for project(s) associated with this public inbox
https://passt.top/passt
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for IMAP folder(s).