| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <20191219102300.lqhxhalpa6ai4t5x@brauner.io>
Date: Thu, 19 Dec 2019 11:23:03 +0100
From: Christian Brauner <christian@...uner.io>
To: Sargun Dhillon <sargun@...gun.me>
Cc: linux-kernel@...r.kernel.org,
containers@...ts.linux-foundation.org, linux-api@...r.kernel.org,
linux-fsdevel@...r.kernel.org, tycho@...ho.ws, jannh@...gle.com,
cyphar@...har.com, christian.brauner@...ntu.com, oleg@...hat.com,
luto@...capital.net, viro@...iv.linux.org.uk,
gpascutto@...illa.com, ealvarez@...illa.com, fweimer@...hat.com,
jld@...illa.com, arnd@...db.de
Subject: Re: [PATCH v4 2/5] pid: Add PIDFD_IOCTL_GETFD to fetch file
descriptors from processes
On Wed, Dec 18, 2019 at 11:55:01PM +0000, Sargun Dhillon wrote:
> This adds an ioctl which allows file descriptors to be extracted
> from processes based on their pidfd.
>
> One reason to use this is to allow sandboxers to take actions on file
> descriptors on the behalf of another process. For example, this can be
> combined with seccomp-bpf's user notification to do on-demand fd
> extraction and take privileged actions. For example, it can be used
> to bind a socket to a privileged port. This is similar to ptrace, and
> using ptrace parasitic code injection to extract a file descriptor from a
> process, but without breaking debuggers, or paying the ptrace overhead
> cost.
>
> You must have the ability to ptrace the process in order to extract any
> file descriptors from it. ptrace can already be used to extract file
> descriptors based on parasitic code injections, so the permissions
> model is aligned.
>
> The ioctl takes a pointer to pidfd_getfd_args. pidfd_getfd_args contains
> a size, which allows for gradual evolution of the API. There is an options
> field, which can be used to state whether the fd should be opened with
> CLOEXEC, or not. An additional options field may be added in the future
> to include the ability to clear cgroup information about the file
> descriptor at a later point. If the structure is from a newer kernel, and
> includes members which make it larger than the structure that's known to
> this kernel version, E2BIG will be returned.
I think that whole part can go now. :)
The rest is just nits (see below).
>
> Signed-off-by: Sargun Dhillon <sargun@...gun.me>
> ---
> .../userspace-api/ioctl/ioctl-number.rst | 1 +
> MAINTAINERS | 1 +
> include/uapi/linux/pidfd.h | 10 +++
> kernel/fork.c | 77 +++++++++++++++++++
> 4 files changed, 89 insertions(+)
> create mode 100644 include/uapi/linux/pidfd.h
>
> diff --git a/Documentation/userspace-api/ioctl/ioctl-number.rst b/Documentation/userspace-api/ioctl/ioctl-number.rst
> index 4ef86433bd67..9f9be681662b 100644
> --- a/Documentation/userspace-api/ioctl/ioctl-number.rst
> +++ b/Documentation/userspace-api/ioctl/ioctl-number.rst
> @@ -273,6 +273,7 @@ Code Seq# Include File Comments
> <mailto:tim@...erelk.net>
> 'p' A1-A5 linux/pps.h LinuxPPS
> <mailto:giometti@...ux.it>
> +'p' B0-CF linux/pidfd.h
> 'q' 00-1F linux/serio.h
> 'q' 80-FF linux/telephony.h Internet PhoneJACK, Internet LineJACK
> linux/ixjuser.h <http://web.archive.org/web/%2A/http://www.quicknet.net>
> diff --git a/MAINTAINERS b/MAINTAINERS
> index cc0a4a8ae06a..bc370ff59dbf 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -13014,6 +13014,7 @@ M: Christian Brauner <christian@...uner.io>
> L: linux-kernel@...r.kernel.org
> S: Maintained
> T: git git://git.kernel.org/pub/scm/linux/kernel/git/brauner/linux.git
> +F: include/uapi/linux/pidfd.h
> F: samples/pidfd/
> F: tools/testing/selftests/pidfd/
> F: tools/testing/selftests/clone3/
> diff --git a/include/uapi/linux/pidfd.h b/include/uapi/linux/pidfd.h
> new file mode 100644
> index 000000000000..90ff535be048
> --- /dev/null
> +++ b/include/uapi/linux/pidfd.h
> @@ -0,0 +1,10 @@
> +/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
> +#ifndef _UAPI_LINUX_PID_H
> +#define _UAPI_LINUX_PID_H
I think that wants to be
#ifndef _UAPI_LINUX_PIDFD_H
> +
> +#include <linux/types.h>
> +#include <linux/ioctl.h>
> +
> +#define PIDFD_IOCTL_GETFD _IOWR('p', 0xb0, __u32)
As Arnd mentioned, this defines a read-write ioctl where it just needs
to be a read-ioctl.
> +
> +#endif /* _UAPI_LINUX_PID_H */
> diff --git a/kernel/fork.c b/kernel/fork.c
> index 2508a4f238a3..09b5f233b5a8 100644
> --- a/kernel/fork.c
> +++ b/kernel/fork.c
> @@ -94,6 +94,7 @@
> #include <linux/thread_info.h>
> #include <linux/stackleak.h>
> #include <linux/kasan.h>
> +#include <uapi/linux/pidfd.h>
>
> #include <asm/pgtable.h>
> #include <asm/pgalloc.h>
> @@ -1790,9 +1791,85 @@ static __poll_t pidfd_poll(struct file *file, struct poll_table_struct *pts)
> return poll_flags;
> }
>
> +static struct file *__pidfd_getfd_fget_task(struct task_struct *task, u32 fd)
> +{
> + struct file *file;
> + int ret;
> +
> + ret = mutex_lock_killable(&task->signal->cred_guard_mutex);
> + if (ret)
> + return ERR_PTR(ret);
> +
> + if (!ptrace_may_access(task, PTRACE_MODE_READ_REALCREDS)) {
> + file = ERR_PTR(-EPERM);
> + goto out;
> + }
> +
> + file = fget_task(task, fd);
> + if (!file)
> + file = ERR_PTR(-EBADF);
> +
> +out:
> + mutex_unlock(&task->signal->cred_guard_mutex);
> + return file;
> +}
> +
> +static long pidfd_getfd(struct pid *pid, u32 fd)
> +{
> + struct task_struct *task;
> + struct file *file;
> + int ret, retfd;
> +
> + task = get_pid_task(pid, PIDTYPE_PID);
> + if (!task)
> + return -ESRCH;
> +
> + file = __pidfd_getfd_fget_task(task, fd);
> + put_task_struct(task);
> + if (IS_ERR(file))
> + return PTR_ERR(file);
> +
> + retfd = get_unused_fd_flags(O_CLOEXEC);
> + if (retfd < 0) {
> + ret = retfd;
> + goto out;
> + }
> +
> + /*
> + * security_file_receive must come last since it may have side effects
> + * and cannot be reversed.
> + */
> + ret = security_file_receive(file);
> + if (ret)
> + goto out_put_fd;
> +
> + fd_install(retfd, file);
> + return retfd;
> +
> +out_put_fd:
> + put_unused_fd(retfd);
> +out:
> + fput(file);
> + return ret;
> +}
> +
> +static long pidfd_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
> +{
> + struct pid *pid = file->private_data;
> +
> + switch (cmd) {
> + case PIDFD_IOCTL_GETFD:
> + return pidfd_getfd(pid, arg);
> + default:
> + return -EINVAL;
> + }
> +}
> +
> const struct file_operations pidfd_fops = {
> .release = pidfd_release,
> .poll = pidfd_poll,
> + .unlocked_ioctl = pidfd_ioctl,
> + .compat_ioctl = compat_ptr_ioctl,
> #ifdef CONFIG_PROC_FS
> .show_fdinfo = pidfd_show_fdinfo,
> #endif
> --
> 2.20.1
>
Powered by blists - more mailing lists