On 3/3/21 5:13 PM, Ian Jackson wrote:
> CAUTION: This email originated from outside of the organization. Do not click
> links or open attachments unless you can confirm the sender and know the
> content is safe.
>
>
>
> Norbert Manthey writes ("[PATCH XENSTORE v1 09/10] xs: handle daemon socket
> error"):
>> When starting the daemon, we might see a NULL pointer instead of the
>> path to the socket.
This first sentence could be more specific, i.e.:
When connecting to the deamon in xs_open, the functions that return the
socket or device location might return NULL in corner cases.
>>
>> Only relevant in case we start the process in a very deep directory
>> path, with a length close to 4096 so that appending "/socket" would
>> exceed the limit. Hence, such an error is unlikely, but should still be
>> fixed to not result in a NULL pointer dereference.
> This description talks about starting the daemon ...
>
>> ---
>> tools/libs/store/xs.c | 3 +++
>> 1 file changed, 3 insertions(+)
> But I think ...
>
>> + if (!connect_to)
>> + return NULL;
>> +
> ... this is client code ?
This is client code, yes. The patched 'get_handle' function receives the
parameter 'connect_to' in the function xs_open. There, the value of the
functions 'xs_deamon_socket_ro', 'xs_deamon_socket' and 'xs_domain_dev'
are passed to this function, without checking for the value NULL.
I agree that the description might be confusing, as the fix is applied
to a function that does not cause the actual problem. How about
rephrasing the first part of the commit message to the above proposal?
Best,
Norbert
>
> Apologies if I am confused.
>
> Ian.
Amazon Development Center Germany GmbH
Krausenstr. 38
10117 Berlin
Geschaeftsfuehrung: Christian Schlaeger, Jonathan Weiss
Eingetragen am Amtsgericht Charlottenburg unter HRB 149173 B
Sitz: Berlin
Ust-ID: DE 289 237 879
