From: Jason Wang <jasowang@redhat.com>
To: Yong Wang <wang.yong155@zte.com.cn>,
zhang.zhanghailiang@huawei.com, zhangchen.fnst@cn.fujitsu.com
Cc: lizhijian@cn.fujitsu.com, qemu-devel@nongnu.org, wang.guang55@zte.com.cn
Subject: Re: [Qemu-devel] [PATCHv2 02/04] colo-compare: Process pactkets in the IOThread of the primary
Date: Wed, 7 Jun 2017 16:34:36 +0800 [thread overview]
Message-ID: <586d53fb-aed0-aa25-64d9-a2761a382efa@redhat.com> (raw)
In-Reply-To: <1496659493-1105-3-git-send-email-wang.yong155@zte.com.cn>
On 2017年06月05日 18:44, Yong Wang wrote:
> From: Wang Yong <wang.yong155@zte.com.cn>
>
> Process pactkets in the IOThread which arrived over the socket.
> we use qio_channel_set_aio_fd_handler to set the handlers on the
> IOThread AioContext.then the packets from the primary and the secondary
> are processed in the IOThread.
> Finally remove the colo-compare thread using the IOThread instead.
>
> Signed-off-by: Wang Yong<wang.yong155@zte.com.cn>
> Signed-off-by: Wang Guang<wang.guang55@zte.com.cn>
> ---
> net/colo-compare.c | 133 ++++++++++++++++++++++++++++++++++++-----------------
> net/colo.h | 1 +
> 2 files changed, 91 insertions(+), 43 deletions(-)
>
> diff --git a/net/colo-compare.c b/net/colo-compare.c
> index b0942a4..e3af791 100644
> --- a/net/colo-compare.c
> +++ b/net/colo-compare.c
> @@ -29,6 +29,7 @@
> #include "qemu/sockets.h"
> #include "qapi-visit.h"
> #include "net/colo.h"
> +#include "io/channel.h"
> #include "sysemu/iothread.h"
>
> #define TYPE_COLO_COMPARE "colo-compare"
> @@ -82,11 +83,6 @@ typedef struct CompareState {
> GQueue conn_list;
> /* hashtable to save connection */
> GHashTable *connection_track_table;
> - /* compare thread, a thread for each NIC */
> - QemuThread thread;
> -
> - GMainContext *worker_context;
> - GMainLoop *compare_loop;
>
> /*compare iothread*/
> IOThread *iothread;
> @@ -95,6 +91,14 @@ typedef struct CompareState {
> QEMUTimer *packet_check_timer;
> } CompareState;
>
> +typedef struct {
> + Chardev parent;
> + QIOChannel *ioc; /*I/O channel */
We probably don't want to manipulate char backend's internal io channel.
All need here is to access the frontend API (char-fe.c) I believe, and
hide the internal implementation.
> +} CompareChardev;
> +
> +#define COMPARE_CHARDEV(obj) \
> + OBJECT_CHECK(CompareChardev, (obj), TYPE_CHARDEV_SOCKET)
> +
> typedef struct CompareClass {
> ObjectClass parent_class;
> } CompareClass;
> @@ -107,6 +111,12 @@ enum {
> static int compare_chr_send(CharBackend *out,
> const uint8_t *buf,
> uint32_t size);
> +static void compare_chr_set_aio_fd_handlers(CharBackend *b,
> + AioContext *ctx,
> + IOCanReadHandler *fd_can_read,
> + IOReadHandler *fd_read,
> + IOEventHandler *fd_event,
> + void *opaque);
>
> static gint seq_sorter(Packet *a, Packet *b, gpointer data)
> {
> @@ -534,6 +544,30 @@ err:
> return ret < 0 ? ret : -EIO;
> }
>
> +static void compare_chr_read(void *opaque)
> +{
> + Chardev *chr = opaque;
> + uint8_t buf[CHR_READ_BUF_LEN];
> + int len, size;
> + int max_size;
> +
> + max_size = qemu_chr_be_can_write(chr);
> + if (max_size <= 0) {
> + return;
> + }
> +
> + len = sizeof(buf);
> + if (len > max_size) {
> + len = max_size;
> + }
> + size = CHARDEV_GET_CLASS(chr)->chr_sync_read(chr, (void *)buf, len);
> + if (size == 0) {
> + return;
> + } else if (size > 0) {
> + qemu_chr_be_write(chr, buf, size);
> + }
> +}
> +
> static int compare_chr_can_read(void *opaque)
> {
> return COMPARE_READ_LEN_MAX;
> @@ -550,8 +584,8 @@ static void compare_pri_chr_in(void *opaque, const uint8_t *buf, int size)
>
> ret = net_fill_rstate(&s->pri_rs, buf, size);
> if (ret == -1) {
> - qemu_chr_fe_set_handlers(&s->chr_pri_in, NULL, NULL, NULL,
> - NULL, NULL, true);
> + compare_chr_set_aio_fd_handlers(&s->chr_pri_in, s->ctx,
> + NULL, NULL, NULL, NULL);
> error_report("colo-compare primary_in error");
> }
> }
> @@ -567,8 +601,8 @@ static void compare_sec_chr_in(void *opaque, const uint8_t *buf, int size)
>
> ret = net_fill_rstate(&s->sec_rs, buf, size);
> if (ret == -1) {
> - qemu_chr_fe_set_handlers(&s->chr_sec_in, NULL, NULL, NULL,
> - NULL, NULL, true);
> + compare_chr_set_aio_fd_handlers(&s->chr_sec_in, s->ctx,
> + NULL, NULL, NULL, NULL);
> error_report("colo-compare secondary_in error");
> }
> }
> @@ -605,34 +639,57 @@ static void colo_compare_timer_del(CompareState *s)
> }
> }
>
> -static void *colo_compare_thread(void *opaque)
> -{
> - CompareState *s = opaque;
> -
> - s->worker_context = g_main_context_new();
> -
> - qemu_chr_fe_set_handlers(&s->chr_pri_in, compare_chr_can_read,
> - compare_pri_chr_in, NULL, s, s->worker_context, true);
> - qemu_chr_fe_set_handlers(&s->chr_sec_in, compare_chr_can_read,
> - compare_sec_chr_in, NULL, s, s->worker_context, true);
> -
> - s->compare_loop = g_main_loop_new(s->worker_context, FALSE);
> -
> - g_main_loop_run(s->compare_loop);
> -
> - g_main_loop_unref(s->compare_loop);
> - g_main_context_unref(s->worker_context);
> - return NULL;
> -}
>
> static void colo_compare_iothread(CompareState *s)
> {
> object_ref(OBJECT(s->iothread));
> s->ctx = iothread_get_aio_context(s->iothread);
>
> + compare_chr_set_aio_fd_handlers(&s->chr_pri_in, s->ctx,
> + compare_chr_can_read,
> + compare_pri_chr_in,
> + NULL,
> + s);
> + compare_chr_set_aio_fd_handlers(&s->chr_sec_in, s->ctx,
> + compare_chr_can_read,
> + compare_sec_chr_in,
> + NULL,
> + s);
> +
> colo_compare_timer_init(s);
> }
>
> +static void compare_chr_set_aio_fd_handlers(CharBackend *b,
> + AioContext *ctx,
> + IOCanReadHandler *fd_can_read,
> + IOReadHandler *fd_read,
> + IOEventHandler *fd_event,
> + void *opaque)
> +{
> + CompareChardev *s;
> +
> + if (!b->chr) {
> + return;
> + }
> + s = COMPARE_CHARDEV(b->chr);
> + if (!s->ioc) {
> + return;
> + }
So this is hacky, you can refer how vhost-user validate udp socket char
backend.
> +
> + b->chr_can_read = fd_can_read;
> + b->chr_read = fd_read;
> + b->chr_event = fd_event;
> + b->opaque = opaque;
> + remove_fd_in_watch(b->chr);
> +
> + if (b->chr_read) {
> + qio_channel_set_aio_fd_handler(s->ioc, ctx,
> + compare_chr_read, NULL, b->chr);
> + } else {
> + qio_channel_set_aio_fd_handler(s->ioc, ctx, NULL, NULL, NULL);
So instead of doing such hack, how about passing a AioContext * instead
of GMainContext * to qemu_chr_fe_set_handlers?
Thanks
> + }
> +}
> +
> static char *compare_get_pri_indev(Object *obj, Error **errp)
> {
> CompareState *s = COLO_COMPARE(obj);
> @@ -736,8 +793,6 @@ static void colo_compare_complete(UserCreatable *uc, Error **errp)
> {
> CompareState *s = COLO_COMPARE(uc);
> Chardev *chr;
> - char thread_name[64];
> - static int compare_id;
>
> if (!s->pri_indev || !s->sec_indev || !s->outdev || !s->iothread) {
> error_setg(errp, "colo compare needs 'primary_in' ,"
> @@ -776,12 +831,6 @@ static void colo_compare_complete(UserCreatable *uc, Error **errp)
> g_free,
> connection_destroy);
>
> - sprintf(thread_name, "colo-compare %d", compare_id);
> - qemu_thread_create(&s->thread, thread_name,
> - colo_compare_thread, s,
> - QEMU_THREAD_JOINABLE);
> - compare_id++;
> -
> colo_compare_iothread(s);
>
> return;
> @@ -834,16 +883,14 @@ static void colo_compare_finalize(Object *obj)
> {
> CompareState *s = COLO_COMPARE(obj);
>
> - qemu_chr_fe_set_handlers(&s->chr_pri_in, NULL, NULL, NULL, NULL,
> - s->worker_context, true);
> - qemu_chr_fe_set_handlers(&s->chr_sec_in, NULL, NULL, NULL, NULL,
> - s->worker_context, true);
> + compare_chr_set_aio_fd_handlers(&s->chr_pri_in, s->ctx,
> + NULL, NULL, NULL, NULL);
> + compare_chr_set_aio_fd_handlers(&s->chr_sec_in, s->ctx,
> + NULL, NULL, NULL, NULL);
> +
> qemu_chr_fe_deinit(&s->chr_out);
> colo_compare_timer_del(s);
>
> - g_main_loop_quit(s->compare_loop);
> - qemu_thread_join(&s->thread);
> -
> /* Release all unhandled packets after compare thead exited */
> g_queue_foreach(&s->conn_list, colo_flush_packets, s);
>
> diff --git a/net/colo.h b/net/colo.h
> index 7c524f3..936dea1 100644
> --- a/net/colo.h
> +++ b/net/colo.h
> @@ -84,5 +84,6 @@ Connection *connection_get(GHashTable *connection_track_table,
> void connection_hashtable_reset(GHashTable *connection_track_table);
> Packet *packet_new(const void *data, int size);
> void packet_destroy(void *opaque, void *user_data);
> +void remove_fd_in_watch(Chardev *chr);
>
> #endif /* QEMU_COLO_PROXY_H */
next prev parent reply other threads:[~2017-06-07 8:34 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-06-05 10:44 [Qemu-devel] [PATCHv2 00/04] Replace the COLO comparing thread with IOThread Yong Wang
2017-06-05 10:44 ` [Qemu-devel] [PATCHv2 01/04] colo-compare: Use IOThread context timer to Check old packet regularly Yong Wang
2017-06-07 8:30 ` Jason Wang
2017-06-05 10:44 ` [Qemu-devel] [PATCHv2 02/04] colo-compare: Process pactkets in the IOThread of the primary Yong Wang
2017-06-07 8:34 ` Jason Wang [this message]
2017-06-05 10:44 ` [Qemu-devel] [PATCHv2 03/04] colo-compare: Update the COLO document to add the IOThread configuration Yong Wang
2017-06-05 10:44 ` [Qemu-devel] [PATCHv2 04/04] colo-compare: Update the COLO document to fix the processing of secondary packets in the main thread Yong Wang
2017-06-07 8:37 ` Jason Wang
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=586d53fb-aed0-aa25-64d9-a2761a382efa@redhat.com \
--to=jasowang@redhat.com \
--cc=lizhijian@cn.fujitsu.com \
--cc=qemu-devel@nongnu.org \
--cc=wang.guang55@zte.com.cn \
--cc=wang.yong155@zte.com.cn \
--cc=zhang.zhanghailiang@huawei.com \
--cc=zhangchen.fnst@cn.fujitsu.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).