From: Orit Wasserman <owasserm@redhat.com>
To: "Michael R. Hines" <mrhines@linux.vnet.ibm.com>
Cc: pbonzini@redhat.com, aliguori@us.ibm.com, quintela@redhat.com,
qemu-devel@nongnu.org, mrhines@us.ibm.com
Subject: Re: [Qemu-devel] [PATCH] rdma: bugfix: make IPv6 support work
Date: Tue, 30 Jul 2013 18:31:48 +0300 [thread overview]
Message-ID: <51F7DC64.9060604@redhat.com> (raw)
In-Reply-To: <51F7D470.6050708@linux.vnet.ibm.com>
On 07/30/2013 05:57 PM, Michael R. Hines wrote:
> On 07/30/2013 04:14 AM, Orit Wasserman wrote:
>> On 07/27/2013 05:23 AM, mrhines@linux.vnet.ibm.com wrote:
>>> From: "Michael R. Hines" <mrhines@us.ibm.com>
>>>
>>> When testing with libvirt, a simple IPv6 migration test failed
>>> because we were not using getaddrinfo() properly.
>>> This makes IPv6 migration over RDMA work.
>>>
>>> Also, we forgot to turn the DPRINTF flag off =).
>>>
>>> Signed-off-by: Michael R. Hines <mrhines@us.ibm.com>
>>> ---
>>> migration-rdma.c | 35 ++++++++++++++++++++++-------------
>>> 1 file changed, 22 insertions(+), 13 deletions(-)
>>>
>>> diff --git a/migration-rdma.c b/migration-rdma.c
>>> index d044830..3256c9b 100644
>>> --- a/migration-rdma.c
>>> +++ b/migration-rdma.c
>>> @@ -27,7 +27,7 @@
>>> #include <string.h>
>>> #include <rdma/rdma_cma.h>
>>> -#define DEBUG_RDMA
>>> +//#define DEBUG_RDMA
>> Can you put this in a separate patch?
>
> Acknowledged.
>
>>> //#define DEBUG_RDMA_VERBOSE
>>> //#define DEBUG_RDMA_REALLY_VERBOSE
>>> @@ -392,6 +392,7 @@ typedef struct RDMAContext {
>>> uint64_t unregistrations[RDMA_SIGNALED_SEND_MAX];
>>> GHashTable *blockmap;
>>> + bool ipv6;
>>> } RDMAContext;
>>> /*
>>> @@ -744,6 +745,7 @@ static int qemu_rdma_resolve_host(RDMAContext *rdma, Error **errp)
>>> char port_str[16];
>>> struct rdma_cm_event *cm_event;
>>> char ip[40] = "unknown";
>>> + int af = rdma->ipv6 ? PF_INET6 : PF_INET;
>> We have code that handles ipv6 in utils/qemu-sockets.c, it also handle host
>> and port parsing please take a look at inet_parse_connect_opts.
>> I think it can be reused here and for the destination.
>
> RDMA cannot use that function - it creates a socket and RDMA does not use sockets.
>
> RDMA is *already*, however, using inet_parse() which does exactly what you said.
>
You can update the function so it can be used for RDMA also.
Orit
>> Regards,
>> Orit
>>> if (rdma->host == NULL || !strcmp(rdma->host, "")) {
>>> ERROR(errp, "RDMA hostname has not been set\n");
>>> @@ -773,7 +775,7 @@ static int qemu_rdma_resolve_host(RDMAContext *rdma, Error **errp)
>>> goto err_resolve_get_addr;
>>> }
>>> - inet_ntop(AF_INET, &((struct sockaddr_in *) res->ai_addr)->sin_addr,
>>> + inet_ntop(af, &((struct sockaddr_in *) res->ai_addr)->sin_addr,
>>> ip, sizeof ip);
>>> DPRINTF("%s => %s\n", rdma->host, ip);
>>> @@ -2236,9 +2238,12 @@ err_rdma_source_connect:
>>> static int qemu_rdma_dest_init(RDMAContext *rdma, Error **errp)
>>> {
>>> int ret = -EINVAL, idx;
>>> + int af = rdma->ipv6 ? PF_INET6 : PF_INET;
>>> struct sockaddr_in sin;
>>> struct rdma_cm_id *listen_id;
>>> char ip[40] = "unknown";
>>> + struct addrinfo *res;
>>> + char port_str[16];
>>> for (idx = 0; idx <= RDMA_WRID_MAX; idx++) {
>>> rdma->wr_data[idx].control_len = 0;
>>> @@ -2266,27 +2271,30 @@ static int qemu_rdma_dest_init(RDMAContext *rdma, Error **errp)
>>> }
>>> memset(&sin, 0, sizeof(sin));
>>> - sin.sin_family = AF_INET;
>>> + sin.sin_family = af;
>>> sin.sin_port = htons(rdma->port);
>>> + snprintf(port_str, 16, "%d", rdma->port);
>>> + port_str[15] = '\0';
>>> if (rdma->host && strcmp("", rdma->host)) {
>>> - struct hostent *dest_addr;
>>> - dest_addr = gethostbyname(rdma->host);
>>> - if (!dest_addr) {
>>> - ERROR(errp, "migration could not gethostbyname!\n");
>>> - ret = -EINVAL;
>>> + ret = getaddrinfo(rdma->host, port_str, NULL, &res);
>>> + if (ret < 0) {
>>> + ERROR(errp, "could not getaddrinfo address %s\n", rdma->host);
>>> goto err_dest_init_bind_addr;
>>> }
>>> - memcpy(&sin.sin_addr.s_addr, dest_addr->h_addr,
>>> - dest_addr->h_length);
>>> - inet_ntop(AF_INET, dest_addr->h_addr, ip, sizeof ip);
>>> +
>>> +
>>> + inet_ntop(af, &((struct sockaddr_in *) res->ai_addr)->sin_addr,
>>> + ip, sizeof ip);
>>> } else {
>>> - sin.sin_addr.s_addr = INADDR_ANY;
>>> + ERROR(errp, "migration host and port not specified!\n");
>>> + ret = -EINVAL;
>>> + goto err_dest_init_bind_addr;
>>> }
>>> DPRINTF("%s => %s\n", rdma->host, ip);
>>> - ret = rdma_bind_addr(listen_id, (struct sockaddr *)&sin);
>>> + ret = rdma_bind_addr(listen_id, res->ai_addr);
>>> if (ret) {
>>> ERROR(errp, "Error: could not rdma_bind_addr!\n");
>>> goto err_dest_init_bind_addr;
>>> @@ -2321,6 +2329,7 @@ static void *qemu_rdma_data_init(const char *host_port, Error **errp)
>>> if (addr != NULL) {
>>> rdma->port = atoi(addr->port);
>>> rdma->host = g_strdup(addr->host);
>>> + rdma->ipv6 = addr->ipv6;
>>> } else {
>>> ERROR(errp, "bad RDMA migration address '%s'", host_port);
>>> g_free(rdma);
>>>
>
next prev parent reply other threads:[~2013-07-30 15:31 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-07-27 2:23 [Qemu-devel] [PATCH] rdma: bugfix: make IPv6 support work mrhines
2013-07-27 2:30 ` Andreas Färber
2013-07-27 2:37 ` Michael R. Hines
2013-07-30 8:14 ` Orit Wasserman
2013-07-30 14:57 ` Michael R. Hines
2013-07-30 15:31 ` Orit Wasserman [this message]
2013-07-31 13:39 ` Michael R. Hines
2013-07-31 16:03 ` Orit Wasserman
2013-08-14 16:29 ` Anthony Liguori
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=51F7DC64.9060604@redhat.com \
--to=owasserm@redhat.com \
--cc=aliguori@us.ibm.com \
--cc=mrhines@linux.vnet.ibm.com \
--cc=mrhines@us.ibm.com \
--cc=pbonzini@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=quintela@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).