xen-devel.lists.xenproject.org archive mirror
 help / color / mirror / Atom feed
From: Ian Campbell <ian.campbell@citrix.com>
To: Yechen Li <lccycc123@gmail.com>
Cc: xen-devel@lists.xen.org
Subject: Re: [PATCH v1][RFC] xen balloon driver numa support, libxl interface
Date: Tue, 13 Aug 2013 21:53:51 +0100	[thread overview]
Message-ID: <1376427231.9273.35.camel@hastur.hellion.org.uk> (raw)
In-Reply-To: <1376313528-16530-1-git-send-email-lccycc123@gmail.com>

On Mon, 2013-08-12 at 21:18 +0800, Yechen Li wrote:
> ---
> 
>   This small patch implements a numa support of memory operation for libxl
>   The command is: xl mem-set-numa [-e] vmid memorysize nodeid
>   To pass the parameters to balloon driver in kernel, I add a file of xen-store
> as /local/domain/(id)/memory/target_nid, hoping this is ok....

It might be OK if you document it in docs/misc/xenstore-paths.markdown.

>   It's my first time submitting a patch, please point out the problems so that
> I could work better in future, thanks very much!

Please see http://wiki.xen.org/wiki/Submitting_Xen_Patches, in
particular the bit about Signed-off-by.

> 
>  tools/libxl/libxl.c       | 14 ++++++++++++--
>  tools/libxl/libxl.h       |  1 +
>  tools/libxl/xl.h          |  1 +
>  tools/libxl/xl_cmdimpl.c  | 45 +++++++++++++++++++++++++++++++++++++++++++++
>  tools/libxl/xl_cmdtable.c |  7 +++++++
>  5 files changed, 66 insertions(+), 2 deletions(-)
> 
> diff --git a/tools/libxl/libxl.c b/tools/libxl/libxl.c
> index 81785df..f027d59 100644
> --- a/tools/libxl/libxl.c
> +++ b/tools/libxl/libxl.c
> @@ -3642,10 +3642,17 @@ retry:
>      }
>      return 0;
>  }
> -
>  int libxl_set_memory_target(libxl_ctx *ctx, uint32_t domid,
>          int32_t target_memkb, int relative, int enforce)
>  {
> +    return libxl_set_memory_target_numa(ctx, domid, target_memkb, relative,
> +                enforce, -1, 0);
> +}
> +
> +int libxl_set_memory_target_numa(libxl_ctx *ctx, uint32_t domid,
> +        int32_t target_memkb, int relative, int enforce,
> +        int node_specify, bool nodeexact)
> +{
>      GC_INIT(ctx);
>      int rc = 1, abort_transaction = 0;
>      uint32_t memorykb = 0, videoram = 0;
> @@ -3754,7 +3761,10 @@ retry_transaction:
>          abort_transaction = 1;
>          goto out;
>      }
> -
> +    //lcc:

Please don't leave debug dropping in place.

> +    LIBXL__LOG(ctx, LIBXL__LOG_DEBUG, "target_nid = %d focus= %d", node_specify, (int) nodeexact);
> +    libxl__xs_write(gc, t, libxl__sprintf(gc, "%s/memory/target_nid",
> +                dompath), "%"PRId32" %"PRId32, node_specify, (int)nodeexact);
>      libxl__xs_write(gc, t, libxl__sprintf(gc, "%s/memory/target",
>                  dompath), "%"PRIu32, new_target_memkb);
>      rc = xc_domain_getinfolist(ctx->xch, domid, 1, &info);
> diff --git a/tools/libxl/libxl.h b/tools/libxl/libxl.h
> index be19bf5..e21d8c3 100644
> --- a/tools/libxl/libxl.h
> +++ b/tools/libxl/libxl.h
> @@ -628,6 +628,7 @@ int libxl_domain_core_dump(libxl_ctx *ctx, uint32_t domid,
>  
>  int libxl_domain_setmaxmem(libxl_ctx *ctx, uint32_t domid, uint32_t target_memkb);
>  int libxl_set_memory_target(libxl_ctx *ctx, uint32_t domid, int32_t target_memkb, int relative, int enforce);
> +int libxl_set_memory_target_numa(libxl_ctx *ctx, uint32_t domid, int32_t target_memkb, int relative, int enforce, int node_specify, bool nodeexact);

This needs a LIBXL_HAVE style declaration. I'm unsure about adding
another function as opposed to extending the current ABI using the
LIBXL_API_VERSION compatibility provisions.
> diff --git a/tools/libxl/xl_cmdtable.c b/tools/libxl/xl_cmdtable.c
> index 326a660..ab918c0 100644
> --- a/tools/libxl/xl_cmdtable.c
> +++ b/tools/libxl/xl_cmdtable.c
> @@ -199,6 +199,13 @@ struct cmd_spec cmd_table[] = {
>        "Set the current memory usage for a domain",
>        "<Domain> <MemMB['b'[bytes]|'k'[KB]|'m'[MB]|'g'[GB]|'t'[TB]]>",
>      },
> +    { "mem-set-numa",

Perhaps instead of adding a new function the existing mem-set should
take a -n <node> parameter?

> +        &main_memset_numa, 0, 1,
> +        "Set the current memory usage for a domain, with numa node specified",
> +        "[-e] <Domain> <MemMB['b'[bytes]|'k'[KB]|'m'[MB]|'g'[GB]|'t'[TB]]> <nid>",
> +        "-e, --exact: operatrion will force on this node exactly"

"operation"

> +        "nid: the machine(physical) node id\n"
> +    },
>      { "button-press",
>        &main_button_press, 0, 1,
>        "Indicate an ACPI button press to the domain",

      parent reply	other threads:[~2013-08-13 20:53 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-08-12 13:18 [PATCH v1][RFC] xen balloon driver numa support, libxl interface Yechen Li
2013-08-12 14:00 ` Jan Beulich
2013-08-12 14:18   ` Li Yechen
2013-08-12 14:31     ` Jan Beulich
2013-08-12 14:57       ` Li Yechen
2013-08-12 15:15         ` Li Yechen
2013-08-12 15:29           ` Jan Beulich
2013-08-12 15:51         ` Dario Faggioli
2013-08-13 20:56           ` Ian Campbell
2013-08-13 23:32             ` Dario Faggioli
2013-08-12 16:16   ` Dario Faggioli
2013-08-12 16:58 ` Dario Faggioli
2013-08-12 17:07 ` Dario Faggioli
2013-08-13 20:53 ` Ian Campbell [this message]

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=1376427231.9273.35.camel@hastur.hellion.org.uk \
    --to=ian.campbell@citrix.com \
    --cc=lccycc123@gmail.com \
    --cc=xen-devel@lists.xen.org \
    /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).