qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
To: Liang Li <liang.z.li@intel.com>
Cc: quintela@redhat.com, qemu-devel@nongnu.org, armbru@redhat.com,
	lcapitulino@redhat.com, yang.z.zhang@intel.com
Subject: Re: [Qemu-devel] [v3 10/13] migration: Add the core code of multi-thread decompression
Date: Fri, 23 Jan 2015 13:42:40 +0000	[thread overview]
Message-ID: <20150123134239.GJ2370@work-vm> (raw)
In-Reply-To: <1418347746-15829-11-git-send-email-liang.z.li@intel.com>

* Liang Li (liang.z.li@intel.com) wrote:
> Signed-off-by: Liang Li <liang.z.li@intel.com>
> Signed-off-by: Yang Zhang <yang.z.zhang@intel.com>
> ---
>  arch_init.c | 48 ++++++++++++++++++++++++++++++++++++++++++++++--
>  1 file changed, 46 insertions(+), 2 deletions(-)
> 
> diff --git a/arch_init.c b/arch_init.c
> index 14bc486..7103f4f 100644
> --- a/arch_init.c
> +++ b/arch_init.c
> @@ -24,6 +24,7 @@
>  #include <stdint.h>
>  #include <stdarg.h>
>  #include <stdlib.h>
> +#include <zlib.h>
>  #ifndef _WIN32
>  #include <sys/types.h>
>  #include <sys/mman.h>
> @@ -820,6 +821,14 @@ static inline void start_compression(compress_param *param)
>      qemu_mutex_unlock(&param->mutex);
>  }
>  
> +static inline void start_decompression(decompress_param *param)
> +{
> +    qemu_mutex_lock(&param->mutex);
> +    param->state = START;
> +    qemu_cond_signal(&param->cond);
> +    qemu_mutex_unlock(&param->mutex);
> +}
> +
>  
>  static uint64_t bytes_transferred;
>  
> @@ -1351,8 +1360,24 @@ void ram_handle_compressed(void *host, uint8_t ch, uint64_t size)
>  
>  static void *do_data_decompress(void *opaque)
>  {
> +    decompress_param *param = opaque;
>      while (!quit_thread) {
> -        /* To be done */
> +        qemu_mutex_lock(&param->mutex);
> +        while (param->state != START) {
> +            qemu_cond_wait(&param->cond, &param->mutex);
> +            if (quit_thread) {
> +                break;
> +            }
> +            size_t pagesize = TARGET_PAGE_SIZE;
> +            /* uncompress() will return failed in some case,
> +             * especially when the page is dirted when doing
> +             * the compression, ignore the return value because
> +             * the dirty page will be retransferred. */
> +            uncompress((Bytef *)param->des, &pagesize,
> +                    (const Bytef *)param->compbuf, param->len);

That's kind of a scary comment!  It looks like 'uncompress' is supposed
to be safe, so shouldn't damage any other data; it's worrying me
might not find real problems though.

However,

Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

> +            param->state = DONE;
> +        }
> +        qemu_mutex_unlock(&param->mutex);
>      }
>      return NULL;
>  }
> @@ -1379,6 +1404,9 @@ void migrate_decompress_threads_join(void)
>      quit_thread = true;
>      thread_count = migrate_decompress_threads();
>      for (i = 0; i < thread_count; i++) {
> +        qemu_cond_signal(&decomp_param[i].cond);
> +    }
> +    for (i = 0; i < thread_count; i++) {
>          qemu_thread_join(decompress_threads + i);
>          qemu_mutex_destroy(&decomp_param[i].mutex);
>          qemu_cond_destroy(&decomp_param[i].cond);
> @@ -1392,7 +1420,23 @@ void migrate_decompress_threads_join(void)
>  static void decompress_data_with_multi_threads(uint8_t *compbuf,
>          void *host, int len)
>  {
> -    /* To be done */
> +    int idx, thread_count;
> +
> +    thread_count = migrate_decompress_threads();
> +    while (true) {
> +        for (idx = 0; idx < thread_count; idx++) {
> +            if (decomp_param[idx].state == DONE) {
> +                memcpy(decomp_param[idx].compbuf, compbuf, len);
> +                decomp_param[idx].des = host;
> +                decomp_param[idx].len = len;
> +                start_decompression(&decomp_param[idx]);
> +                break;
> +            }
> +        }
> +        if (idx < thread_count) {
> +            break;
> +        }
> +    }
>  }
>  
>  static int ram_load(QEMUFile *f, void *opaque, int version_id)
> -- 
> 1.8.3.1
> 
--
Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK

  reply	other threads:[~2015-01-23 13:42 UTC|newest]

Thread overview: 44+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-12-12  1:28 [Qemu-devel] [PATCH v3 0/13] migration: Add a new feature to do live migration Liang Li
2014-12-12  1:28 ` [Qemu-devel] [v3 01/13] docs: Add a doc about multiple thread compression Liang Li
2015-01-23 13:17   ` Dr. David Alan Gilbert
2015-01-23 15:24   ` Eric Blake
2014-12-12  1:28 ` [Qemu-devel] [v3 02/13] migration: Add the framework of multi-thread compression Liang Li
2015-01-23 13:23   ` Dr. David Alan Gilbert
2015-01-23 16:09   ` Eric Blake
2014-12-12  1:28 ` [Qemu-devel] [v3 03/13] migration: Add the framework of muti-thread decompression Liang Li
2015-01-23 13:26   ` Dr. David Alan Gilbert
2015-01-23 16:22   ` Eric Blake
2014-12-12  1:28 ` [Qemu-devel] [v3 04/13] qemu-file: Add tow function will be used in migration Liang Li
2015-01-23 13:31   ` Dr. David Alan Gilbert
2015-01-24 13:42     ` Li, Liang Z
2014-12-12  1:28 ` [Qemu-devel] [v3 05/13] arch_init: alloc and free data struct in multi-thread compression Liang Li
2015-01-23 13:35   ` Dr. David Alan Gilbert
2015-01-24 13:46     ` Li, Liang Z
2014-12-12  1:28 ` [Qemu-devel] [v3 06/13] arch_init: Add data struct used by decompression Liang Li
2014-12-12  1:29 ` [Qemu-devel] [v3 07/13] migraion: Rewrite the function ram_save_page() Liang Li
2015-01-23 13:38   ` Dr. David Alan Gilbert
2014-12-12  1:29 ` [Qemu-devel] [v3 08/13] migration: Add the core code of multi-thread compresion Liang Li
2015-01-23 13:39   ` Dr. David Alan Gilbert
2015-01-24 13:51     ` Li, Liang Z
2014-12-12  1:29 ` [Qemu-devel] [v3 09/13] migration: Make compression co-work with xbzrle Liang Li
2015-01-23 13:40   ` Dr. David Alan Gilbert
2014-12-12  1:29 ` [Qemu-devel] [v3 10/13] migration: Add the core code of multi-thread decompression Liang Li
2015-01-23 13:42   ` Dr. David Alan Gilbert [this message]
2014-12-12  1:29 ` [Qemu-devel] [v3 11/13] migration: Add interface to control compression Liang Li
2015-01-23 13:44   ` Dr. David Alan Gilbert
2015-01-23 15:26   ` Eric Blake
2014-12-12  1:29 ` [Qemu-devel] [v3 12/13] migration: Add command to set migration parameter Liang Li
2015-01-23 13:48   ` Dr. David Alan Gilbert
2015-01-23 15:42     ` Eric Blake
2015-01-23 15:59       ` Dr. David Alan Gilbert
2015-01-23 16:06         ` Eric Blake
2015-01-24 14:14     ` Li, Liang Z
2015-01-26  9:22       ` Dr. David Alan Gilbert
2015-01-23 15:39   ` Eric Blake
2014-12-12  1:29 ` [Qemu-devel] [v3 13/13] migration: Add command to query " Liang Li
2015-01-23 13:49   ` Dr. David Alan Gilbert
2015-01-23 15:47   ` Eric Blake
2014-12-24  5:08 ` [Qemu-devel] [PATCH v3 0/13] migration: Add a new feature to do live migration Li, Liang Z
2015-01-07  3:12 ` Li, Liang Z
2015-01-23 13:10 ` Dr. David Alan Gilbert
2015-01-24 13:25   ` Li, Liang Z

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=20150123134239.GJ2370@work-vm \
    --to=dgilbert@redhat.com \
    --cc=armbru@redhat.com \
    --cc=lcapitulino@redhat.com \
    --cc=liang.z.li@intel.com \
    --cc=qemu-devel@nongnu.org \
    --cc=quintela@redhat.com \
    --cc=yang.z.zhang@intel.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).