From: Janis Danisevskis <jdanis@google.com>
To: Kees Cook <keescook@chromium.org>
Cc: LKML <linux-kernel@vger.kernel.org>,
Andrew Morton <akpm@linux-foundation.org>,
Al Viro <viro@zeniv.linux.org.uk>,
Cyrill Gorcunov <gorcunov@openvz.org>,
Alexey Dobriyan <adobriyan@gmail.com>,
Colin Ian King <colin.king@canonical.com>,
David Rientjes <rientjes@google.com>,
Minfei Huang <mnfhuang@gmail.com>,
John Stultz <john.stultz@linaro.org>,
Calvin Owens <calvinowens@fb.com>, Jann Horn <jann@thejh.net>
Subject: Re: [PATCH] procfs: fixes pthread cross-thread naming if !PR_DUMPABLE
Date: Tue, 3 May 2016 18:25:25 +0100 [thread overview]
Message-ID: <5728DF05.70906@google.com> (raw)
In-Reply-To: <CAGXu5jJ7XKdePkq0pQOe6NS6+8w3dH4REw8NOFXXbJ+rX7vP5A@mail.gmail.com>
On 26/04/16 21:14, Kees Cook wrote:
> On Tue, Apr 26, 2016 at 10:20 AM, Janis Danisevskis <jdanis@google.com> wrote:
>> The PR_DUMPABLE flag causes the pid related paths of the
>> proc file system to be owned by ROOT. The implementation
>> of pthread_set/getname_np however needs access to
>> /proc/<pid>/task/<tid>/comm.
>> If PR_DUMPABLE is false this implementation is locked out.
>>
>> This patch installs a special permission function for
>> the file "comm" that grants read and write access to
>> all threads of the same group regardless of the ownership
>> of the inode. For all other threads the function falls back
>> to the generic inode permission check.
>>
>> Signed-off-by: Janis Danisevskis <jdanis@google.com>
>
> Instead of a permissions function, perhaps this should be handled in
> the open() of proc_pid_set_comm_operations (and the REG permissions
> loosened)? I'm concerned there's a race here between the perm check
> and the resulting open. I'd rather have the open doing the check to
> eliminate the race.
I kind of thought that the permission check is on the open path
could you elaborate on the race that you are expecting?
Also, in what way would you loosen the permissions on the REG?
If the DUMPABLE flag is cleared this node is owned by ROOT.
So the only way to make it writable to a user process would be
to make it world writable. This cannot be your intention.
Janis
>
> -Kees
>
>> ---
>> fs/proc/base.c | 42 +++++++++++++++++++++++++++++++++++++++++-
>> 1 file changed, 41 insertions(+), 1 deletion(-)
>>
>> diff --git a/fs/proc/base.c b/fs/proc/base.c
>> index b1755b2..c8ceb3c8 100644
>> --- a/fs/proc/base.c
>> +++ b/fs/proc/base.c
>> @@ -3157,6 +3157,44 @@ int proc_pid_readdir(struct file *file, struct dir_context *ctx)
>> }
>>
>> /*
>> + * proc_tid_comm_permission is a special permission function exclusively
>> + * used for the node /proc/<pid>/task/<tid>/comm.
>> + * It bypasses generic permission checks in the case where a task of the same
>> + * task group attempts to access the node.
>> + * The rational behind this is that glibc and bionic access this node for
>> + * cross thread naming (pthread_set/getname_np(!self)). However, if
>> + * PR_SET_DUMPABLE gets set to 0 this node among others becomes uid=0 gid=0,
>> + * which locks out the cross thread naming implementation.
>> + * This function makes sure that the node is always accessible for members of
>> + * same thread group.
>> + */
>> +static int proc_tid_comm_permission(struct inode *inode, int mask)
>> +{
>> + bool is_same_tgroup;
>> + struct task_struct *task;
>> +
>> + task = get_proc_task(inode);
>> + if (!task)
>> + return -ESRCH;
>> + is_same_tgroup = same_thread_group(current, task);
>> + put_task_struct(task);
>> +
>> + if (likely(is_same_tgroup && !(mask & MAY_EXEC))) {
>> + /* This file (/proc/<pid>/task/<tid>/comm) can always be
>> + * read or written by the members of the corresponding
>> + * thread group.
>> + */
>> + return 0;
>> + }
>> +
>> + return generic_permission(inode, mask);
>> +}
>> +
>> +static const struct inode_operations proc_tid_comm_inode_operations = {
>> + .permission = proc_tid_comm_permission,
>> +};
>> +
>> +/*
>> * Tasks
>> */
>> static const struct pid_entry tid_base_stuff[] = {
>> @@ -3174,7 +3212,9 @@ static const struct pid_entry tid_base_stuff[] = {
>> #ifdef CONFIG_SCHED_DEBUG
>> REG("sched", S_IRUGO|S_IWUSR, proc_pid_sched_operations),
>> #endif
>> - REG("comm", S_IRUGO|S_IWUSR, proc_pid_set_comm_operations),
>> + NOD("comm", S_IFREG|S_IRUGO|S_IWUSR,
>> + &proc_tid_comm_inode_operations,
>> + &proc_pid_set_comm_operations, {}),
>> #ifdef CONFIG_HAVE_ARCH_TRACEHOOK
>> ONE("syscall", S_IRUSR, proc_pid_syscall),
>> #endif
>> --
>> 2.8.0.rc3.226.g39d4020
>>
>
>
>
next prev parent reply other threads:[~2016-05-03 17:25 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-04-26 17:20 [PATCH] procfs: fixes pthread cross-thread naming if !PR_DUMPABLE Janis Danisevskis
2016-04-26 20:14 ` Kees Cook
2016-05-03 17:25 ` Janis Danisevskis [this message]
2016-05-03 17:42 ` Kees Cook
2016-05-03 18:16 ` Janis Danisevskis
2016-05-03 19:02 ` Kees Cook
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=5728DF05.70906@google.com \
--to=jdanis@google.com \
--cc=adobriyan@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=calvinowens@fb.com \
--cc=colin.king@canonical.com \
--cc=gorcunov@openvz.org \
--cc=jann@thejh.net \
--cc=john.stultz@linaro.org \
--cc=keescook@chromium.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mnfhuang@gmail.com \
--cc=rientjes@google.com \
--cc=viro@zeniv.linux.org.uk \
/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