[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH v12 12/26] secondary vm suspend/resume/checkpoint code

On 03/24/2016 11:15 PM, Ian Jackson wrote:
Changlong Xie writes ("[PATCH v12 12/26] secondary vm suspend/resume/checkpoint 
From: Wen Congyang <wency@xxxxxxxxxxxxxx>


This is all fine except for one small thing:

+    /* COLO only supports HVM now because it does not work very
+     * well with pv drivers:
+     * 1. We need to resume vm in the slow path. In this case we
+     *    need to disconnect/reconnect backend and frontend. It
+     *    will take too much time and the performance is very slow.
+     * 2. PV disk cannot reuse block replication that is implemented
+     *    in QEMU.
+     */
+    if (info->type != LIBXL_DOMAIN_TYPE_HVM &&
+        checkpointed_stream == LIBXL_CHECKPOINTED_STREAM_COLO) {

Will add LOG(ERROR, "xxxx")

+        rc = ERROR_FAIL;
+        goto out;

In response to v11 I wrote:

Please log something here, or it may be very mysterious.

I now see that you replied:

OK. Will add some comments to explain why only support pure HVM now.

I'm sorry that I didn't spot that emailed response from you, because
it's not what I intended.

To "log something" means to add code to print a message to the
logfile.  In general all places in libxl where an ERROR_FAIL is
generated should be accompanied by an appropriate call of a LOG macro.

There are plenty of examples in (for example) libxl_dom.c.

Your log message should include the domid.



Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.