I think this has already been addressed. But forwarding for sure.

Begin forwarded message:

Date: Mon, 23 Apr 2018 01:26:46 +0000
From: bugzilla-dae...@bugzilla.kernel.org
To: step...@networkplumber.org
Subject: [Bug 199469] New: Regression in 32-bit-compat dev_ioctl due to 
bf4405737f9f85a06db2b0ce5d76a818b61992e2


https://bugzilla.kernel.org/show_bug.cgi?id=199469

            Bug ID: 199469
           Summary: Regression in 32-bit-compat dev_ioctl due to
                    bf4405737f9f85a06db2b0ce5d76a818b61992e2
           Product: Networking
           Version: 2.5
    Kernel Version: 4.16.0-0.rc4.git0.1.fc28.x86_64
          Hardware: All
                OS: Linux
              Tree: Mainline
            Status: NEW
          Severity: normal
          Priority: P1
         Component: Other
          Assignee: step...@networkplumber.org
          Reporter: rob...@ocallahan.org
        Regression: No

Created attachment 275501
  --> https://bugzilla.kernel.org/attachment.cgi?id=275501&action=edit  
Test program

The commit says

    Once upon a time net/socket.c:dev_ifsioc() used to handle SIOCSHWTSTAMP and
    SIOCSIFMAP.  These have different native and compat layout, so the format
    conversion had been needed.  In 2009 these two cases had been taken out,
    turning the rest into a convoluted way to calling sock_do_ioctl().  We copy
    compat structure into native one, call sock_do_ioctl() on that and copy
    the result back for the in/out ioctls.  No layout transformation anywhere,
    so we might as well just call sock_do_ioctl() and skip all the headache
with
    copying.

However there is one problem: 32-bit 'struct ifreq' and 64-bit 'struct ifreq'
are not the same size. The former is 32 bytes and the latter is 40 bytes. Thus,
if you place a 32-bit 'struct ifreq' immediately before an unmapped page and
try to pass it to one of these ioctls, the syscall fails with EFAULT due to
this commit.

Steps to reproduce:
Copy attached file to /tmp/test.c, then:
[roc@localhost-live ~]$ gcc -o /tmp/test /tmp/test.c && /tmp/test
Index: 1
[roc@localhost-live ~]$ gcc -m32 -o /tmp/test /tmp/test.c && /tmp/test
Failed SIOCGIFINDEX: Bad address

-- 
You are receiving this mail because:
You are the assignee for the bug.

Reply via email to