On 27/04/16 09:57, Roger Pau Monne wrote:
> On Tue, Apr 26, 2016 at 04:16:51PM +0100, Wei Liu wrote:
>> On Tue, Apr 26, 2016 at 04:52:16PM +0200, Roger Pau Monne wrote:
>>> Signed-off-by: Roger Pau Monné <roger....@citrix.com>
>>> ---
>>> Cc: Ian Jackson <ian.jack...@eu.citrix.com>
>>> Cc: Wei Liu <wei.l...@citrix.com>
>>> ---
>>> tools/libxc/xc_dom_bzimageloader.c | 2 +-
>>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/tools/libxc/xc_dom_bzimageloader.c
>>> b/tools/libxc/xc_dom_bzimageloader.c
>>> index 7fde42a..0a4041c 100644
>>> --- a/tools/libxc/xc_dom_bzimageloader.c
>>> +++ b/tools/libxc/xc_dom_bzimageloader.c
>>> @@ -482,7 +482,7 @@ static int xc_try_lzo1x_decode(
>>> if ( !dst_len )
>>> {
>>> msg = "Error registering stream output";
>>> - if ( xc_dom_register_external(dom, out_buf, out_len) )
>>> + if ( xc_dom_register_external(dom, out_buf, 0) )
>> I fail to figure out why this is correct.
>>
>> I think the out_len should be moved out of the loop and initialise as 0.
>> We still need to use out_len here.
> I'm not following here. AFAICT out_len is always uninitialized at this
> point (from loop 0 to N), so I assume 0 is what was intended to be passed
> here. Or it this supposed to be passing in the last out_len value from the
> previous iteration of the loop?
xc_dom_register_external() is completely mad, and it is not obvious as
to its purpose; tt is doing some kind of allocation tracking for
malloc()-allocated blocks. It looks like the output length is only
uses for stats purposes, as the memory in out_buf will be freed properly
by free().
So I think the correct value should be the actual size of the allocation
of out_buf, but I also don't think it makes any real difference either way.
~Andrew
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
