Il 17/04/2013 17:59, Michael R. Hines ha scritto:
>>> Failure already happens for unknown capabilities.
>> Maybe I misread the code, but all I saw is:
>>
>> +if (cap.flags & RDMA_CAPABILITY_CHUNK_REGISTER) {
>> +rdma->chunk_register_destination = true;
>> +} else if (cap
On 04/17/2013 04:58 AM, Paolo Bonzini wrote:
No, I really mean it. No dead code. We included the one useful option
as a capability, and that should be it.
Acknowledged =)
Failure already happens for unknown capabilities.
Maybe I misread the code, but all I saw is:
+if (cap.flags
Il 17/04/2013 05:27, Michael R. Hines ha scritto:
> On 04/16/2013 12:49 AM, Paolo Bonzini wrote:
>> +#define RDMA_CHUNK_REGISTRATION
>> +#define RDMA_LAZY_CLIENT_REGISTRATION
>> Please drop these; no dead code.
>
> These are important sender-side-only debugging flags.
> I'll add an explicit "debug
On 04/16/2013 12:49 AM, Paolo Bonzini wrote:
+#define RDMA_CHUNK_REGISTRATION
+#define RDMA_LAZY_CLIENT_REGISTRATION
Please drop these; no dead code.
These are important sender-side-only debugging flags.
I'll add an explicit "debug-only" comment.
+/* Do not merge data if larger than this. */
Il 16/04/2013 06:49, Paolo Bonzini ha scritto:
>> > +remote_ram_blocks.num_blocks = remote_ram_blocks.remote_area;
>> > +remote_ram_blocks.block = (void *) (remote_ram_blocks.num_blocks + 1);
> You cannot do this, it doesn't guarantee that remote_ram_blocks.block is
> correctly aligned. Pl
