Re: [Qemu-devel] [PATCH v1 5/9] test-string-input-visitor: add more tests

[Markus Armbruster <armbru@redhat.com>]

David Hildenbrand <david@redhat.com> writes:

> On 15.11.18 18:13, Eric Blake wrote:
>> On 11/15/18 8:04 AM, David Hildenbrand wrote:
>>> Test that very big/small values are not accepted and that ranges with
>>> only one element work.
>>>
>>> Rename expect4 to expect5, as we will be moving that to a separate ulist
>>> test after the rework.
>>>
>>> Signed-off-by: David Hildenbrand <david@redhat.com>
>>> ---
>>>   tests/test-string-input-visitor.c | 22 ++++++++++++++++++++--
>>>   1 file changed, 20 insertions(+), 2 deletions(-)
>>>
>> 
>> I don't see a test for a range that wraps around (such as UINT_MAX-0); 
>> that's worth testing (whether it happens to work or is rejected as 
>> invalid).  Do we require ranges to be ascending, or does 6-5 result in 
>> the sequence 5, 6?  I also recall that our range code imposes a limit on 
>> the maximum elements present in a single range, in order to prevent 
>> denial-of-service attacks where a caller could request 0-INT_MAX to 
>> exhaust resources enumerating everything in the range; does our 
>> testsuite cover those limits?
>>
> Ranges have to be ascending and old code enforced that. New code still
> enforces it. Wrapping ranges are AFAIC also not supported - not
> ascending. I can add a test.

Good.

> The limit is a good point. It is neither in the tests nor in the new
> code. But now we finally have an explanation on the 65000-somewhat
> thingy. I assume that we need such a limit?

Yes, we do.  We don't expect untrusted input here, but typo in the
monitor (say 0--1 parsed as uint64_t 0, UINT_MAX) killing the VM by
eating all memory is not a happy user experience.