From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <systemtap-return-20483-listarch-systemtap=sources.redhat.com@sourceware.org>
Received: (qmail 26938 invoked by alias); 4 Mar 2013 17:54:38 -0000
Received: (qmail 26382 invoked by uid 48); 4 Mar 2013 17:53:59 -0000
From: "dsmith at redhat dot com" <sourceware-bugzilla@sourceware.org>
To: systemtap@sourceware.org
Subject: [Bug tapsets/15219] syscall.exp failures on RHEL5, RHEL6, and rawhide
Date: Mon, 04 Mar 2013 17:54:00 -0000
X-Bugzilla-Reason: AssignedTo
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: systemtap
X-Bugzilla-Component: tapsets
X-Bugzilla-Keywords:
X-Bugzilla-Severity: normal
X-Bugzilla-Who: dsmith at redhat dot com
X-Bugzilla-Status: NEW
X-Bugzilla-Priority: P2
X-Bugzilla-Assigned-To: systemtap at sourceware dot org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Changed-Fields:
Message-ID: <bug-15219-6586-f8TpojVCTN@http.sourceware.org/bugzilla/>
In-Reply-To: <bug-15219-6586@http.sourceware.org/bugzilla/>
References: <bug-15219-6586@http.sourceware.org/bugzilla/>
X-Bugzilla-URL: http://sourceware.org/bugzilla/
Auto-Submitted: auto-generated
Content-Type: text/plain; charset="UTF-8"
MIME-Version: 1.0
Mailing-List: contact systemtap-help@sourceware.org; run by ezmlm
Precedence: bulk
List-Id: <systemtap.sourceware.org>
List-Subscribe: <mailto:systemtap-subscribe@sourceware.org>
List-Post: <mailto:systemtap@sourceware.org>
List-Help: <mailto:systemtap-help@sourceware.org>, <http://sourceware.org/lists.html#faqs>
Sender: systemtap-owner@sourceware.org
X-SW-Source: 2013-q1/txt/msg00206.txt.bz2

http://sourceware.org/bugzilla/show_bug.cgi?id=15219

--- Comment #1 from David Smith <dsmith at redhat dot com> 2013-03-04 17:53:58 UTC ---
Here's compat_sys_timer_settime() from kernel/compat.c. Notice it converts the
'struct compat_itimerspec' to a 'struct itimerspec' (which is in kernel
memory), calls 'set_fs(KERENL_DS), then calls the real syscall function,
'sys_timer_settime()'.

====
long compat_sys_timer_settime(timer_t timer_id, int flags,
              struct compat_itimerspec __user *new,
              struct compat_itimerspec __user *old)
{
    long err;
    mm_segment_t oldfs;
    struct itimerspec newts, oldts;

    if (!new)
        return -EINVAL;
    if (get_compat_itimerspec(&newts, new))
        return -EFAULT;
    oldfs = get_fs();
    set_fs(KERNEL_DS);
    err = sys_timer_settime(timer_id, flags,
                (struct itimerspec __user *) &newts,
                (struct itimerspec __user *) &oldts);
    set_fs(oldfs);
    if (!err && old && put_compat_itimerspec(old, &oldts))
        return -EFAULT;
    return err;
}
====

Here's our '_stp_copy_from_user()'. Notice we explicitly call
'set_fs(USER_DS)', which just overrode compat_sys_timer_settime() setting.

====
static unsigned long _stp_copy_from_user(char *dst, const char __user *src,
unsigned long count)
{
    if (count) {
                mm_segment_t _oldfs = get_fs();
                set_fs(USER_DS);
                pagefault_disable();
        if (access_ok(VERIFY_READ, src, count))
            count = __copy_from_user_inatomic(dst, src, count);
        else
            memset(dst, 0, count);
                pagefault_enable();
                set_fs(_oldfs);
    }
    return count;
}
====

I'm thinking we should no longer change the kernel's idea of what memory space
to use in _stp_copy_from_user().

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.