[U-Boot] [PATCH v3 0/8] Fix CVE-2018-18440 and CVE-2018-18439

[Simon Glass]

Hi Simon,

On Mon, 3 Dec 2018 at 00:50, Simon Goldschmidt
<simon.k.r.goldschmidt at gmail.com> wrote:
>
> Simon,
>
> On Tue, Nov 27, 2018 at 6:45 AM Simon Goldschmidt
> <simon.k.r.goldschmidt at gmail.com> wrote:
> >
> > On Tue, Nov 27, 2018 at 2:02 AM Simon Glass <sjg at chromium.org> wrote:
> > >
> > > Hi Simon,
> > >
> > > On Sat, 17 Nov 2018 at 05:25, Simon Goldschmidt
> > > <simon.k.r.goldschmidt at gmail.com> wrote:
> > > >
> > > > This series fixes CVE-2018-18440 ("insufficient boundary checks in
> > > > filesystem image load") by adding restrictions to the 'load'
> > > > command and fixes CVE-2018-18439 ("insufficient boundary checks in
> > > > network image boot") by adding restrictions to the tftp code.
> > > >
> > > > The functions from lmb.c are used to setup regions of allowed and
> > > > reserved memory. Then, the file size to load is checked against these
> > > > addresses and loading the file is aborted if it would overwrite
> > > > reserved memory.
> > > >
> > > > The memory reservation code is reused from bootm/image.
> > > >
> > > > Changes in v3:
> > > > - No patch changes, but needed to resend since patman added too many cc
> > > >   addresses that gmail seemed to detect as spam :-(
> > > >
> > > > Changes in v2:
> > > > - added code to reserve devicetree reserved-memory in lmb
> > > > - added tftp fixes (patches 7 and 8)
> > > > - fixed a bug in new function lmb_alloc_addr
> > > >
> > > > Simon Goldschmidt (8):
> > > >   lib: lmb: reserving overlapping regions should fail
> > > >   fdt: parse "reserved-memory" for memory reservation
> > > >   lib: lmb: extend lmb for checks at load time
> > > >   fs: prevent overwriting reserved memory
> > > >   bootm: use new common function lmb_init_and_reserve
> > > >   lmb: remove unused extern declaration
> > > >   net: remove CONFIG_MCAST_TFTP
> > > >   tftp: prevent overwriting reserved memory
> > > >
> > > >  README                       |   9 --
> > > >  common/bootm.c               |   8 +-
> > > >  common/image-fdt.c           |  52 ++++++-
> > > >  drivers/net/rtl8139.c        |   9 --
> > > >  drivers/net/tsec.c           |  52 -------
> > > >  drivers/usb/gadget/ether.c   |   3 -
> > > >  fs/fs.c                      |  56 ++++++-
> > > >  include/lmb.h                |   7 +-
> > > >  include/net.h                |  17 ---
> > > >  lib/lmb.c                    |  69 +++++++++
> > > >  net/eth-uclass.c             |   4 -
> > > >  net/eth_legacy.c             |  46 ------
> > > >  net/net.c                    |   9 +-
> > > >  net/tftp.c                   | 289 +++++++----------------------------
> > > >  scripts/config_whitelist.txt |   1 -
> > > >  15 files changed, 232 insertions(+), 399 deletions(-)
> > >
> > > This is great work, but what is missing is a test for lmb.
> >
> > Yeah, well, the tests didn't work on my system and I figured it's
> > better to get the code fixed than to use my time on trying to get the
> > tests running.
> >
> > However, after searching for the required packages and fiddling around
> > some more, I guess I made them work so I could add tests now...
> >
> > I also have work-in-progress for compressing fit image contents (we
> > currently only support uncompressing the kernel). It will switch some
> > 'lmb_reserve' calls to the new 'lmb_alloc_addr' as this is more safe.
> > Maybe I can combine the tests in that series?
>
> After managing to get the tests to run via 'make qcheck' (and 'make
> tests'; had to install much more than listed in 'test/py/README.md'),
> I tried to add tests to 'test/lib/' (next to hexdump.c), but I failed
> to get them run. Even chaning 'test/lib/hexdump.c' to fail did not
> produce errors. Are these tests not included in 'make qcheck'?

That runs the Python tests which are in test/py/tests. Some of those
tests run compiled-in code (e.g. log_test.c and cmd_ut.c). Is your
test intended to be Python or C?
Regards,
Simon