qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Jeff Cody <jcody@redhat.com>
To: qemu-block@nongnu.org
Cc: peter.maydell@linaro.org, jcody@redhat.com, qemu-devel@nongnu.org
Subject: [Qemu-devel] [PULL 03/10] block/nfs: add support for libnfs pagecache
Date: Tue, 28 Jun 2016 23:10:37 -0400	[thread overview]
Message-ID: <1467169844-18111-4-git-send-email-jcody@redhat.com> (raw)
In-Reply-To: <1467169844-18111-1-git-send-email-jcody@redhat.com>

From: Peter Lieven <pl@kamp.de>

upcoming libnfs will have support for a read cache that can
significantly help to speed up requests since libnfs by design
circumvents the kernel cache.

Example:
 qemu -cdrom nfs://127.0.0.1/iso/my.iso?pagecache=1024

The pagecache parameters takes the maximum amount of pages to
cache.  A page in libnfs is always the NFS_BLKSIZE which is
4KB.

Signed-off-by: Peter Lieven <pl@kamp.de>
Reviewed-by: Jeff Cody <jcody@redhat.com>
Message-id: 1463662083-20814-3-git-send-email-pl@kamp.de
Signed-off-by: Jeff Cody <jcody@redhat.com>
---
 block/nfs.c | 37 ++++++++++++++++++++++++++++++++++++-
 1 file changed, 36 insertions(+), 1 deletion(-)

diff --git a/block/nfs.c b/block/nfs.c
index 60be45e..15d6832 100644
--- a/block/nfs.c
+++ b/block/nfs.c
@@ -38,6 +38,7 @@
 #include <nfsc/libnfs.h>
 
 #define QEMU_NFS_MAX_READAHEAD_SIZE 1048576
+#define QEMU_NFS_MAX_PAGECACHE_SIZE (8388608 / NFS_BLKSIZE)
 #define QEMU_NFS_MAX_DEBUG_LEVEL 2
 
 typedef struct NFSClient {
@@ -342,6 +343,26 @@ static int64_t nfs_client_open(NFSClient *client, const char *filename,
                 val = QEMU_NFS_MAX_READAHEAD_SIZE;
             }
             nfs_set_readahead(client->context, val);
+#ifdef LIBNFS_FEATURE_PAGECACHE
+            nfs_set_pagecache_ttl(client->context, 0);
+#endif
+            client->cache_used = true;
+#endif
+#ifdef LIBNFS_FEATURE_PAGECACHE
+            nfs_set_pagecache_ttl(client->context, 0);
+        } else if (!strcmp(qp->p[i].name, "pagecache")) {
+            if (open_flags & BDRV_O_NOCACHE) {
+                error_setg(errp, "Cannot enable NFS pagecache "
+                                 "if cache.direct = on");
+                goto fail;
+            }
+            if (val > QEMU_NFS_MAX_PAGECACHE_SIZE) {
+                error_report("NFS Warning: Truncating NFS pagecache"
+                             " size to %d pages", QEMU_NFS_MAX_PAGECACHE_SIZE);
+                val = QEMU_NFS_MAX_PAGECACHE_SIZE;
+            }
+            nfs_set_pagecache(client->context, val);
+            nfs_set_pagecache_ttl(client->context, 0);
             client->cache_used = true;
 #endif
 #ifdef LIBNFS_FEATURE_DEBUG
@@ -524,7 +545,8 @@ static int nfs_reopen_prepare(BDRVReopenState *state,
     }
 
     if ((state->flags & BDRV_O_NOCACHE) && client->cache_used) {
-        error_setg(errp, "Cannot disable cache if libnfs readahead is enabled");
+        error_setg(errp, "Cannot disable cache if libnfs readahead or"
+                         " pagecache is enabled");
         return -EINVAL;
     }
 
@@ -542,6 +564,15 @@ static int nfs_reopen_prepare(BDRVReopenState *state,
     return 0;
 }
 
+#ifdef LIBNFS_FEATURE_PAGECACHE
+static void nfs_invalidate_cache(BlockDriverState *bs,
+                                 Error **errp)
+{
+    NFSClient *client = bs->opaque;
+    nfs_pagecache_invalidate(client->context, client->fh);
+}
+#endif
+
 static BlockDriver bdrv_nfs = {
     .format_name                    = "nfs",
     .protocol_name                  = "nfs",
@@ -565,6 +596,10 @@ static BlockDriver bdrv_nfs = {
 
     .bdrv_detach_aio_context        = nfs_detach_aio_context,
     .bdrv_attach_aio_context        = nfs_attach_aio_context,
+
+#ifdef LIBNFS_FEATURE_PAGECACHE
+    .bdrv_invalidate_cache          = nfs_invalidate_cache,
+#endif
 };
 
 static void nfs_block_init(void)
-- 
1.9.3

  parent reply	other threads:[~2016-06-29  3:11 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-06-29  3:10 [Qemu-devel] [PULL 00/10] Block patches Jeff Cody
2016-06-29  3:10 ` [Qemu-devel] [PULL 01/10] block/gluster: add support for SEEK_DATA/SEEK_HOLE Jeff Cody
2016-06-29  3:10 ` [Qemu-devel] [PULL 02/10] block/nfs: refuse readahead if cache.direct is on Jeff Cody
2016-06-29  3:10 ` Jeff Cody [this message]
2016-06-29  3:10 ` [Qemu-devel] [PULL 04/10] mirror: fix trace_mirror_yield_in_flight usage in mirror_iteration() Jeff Cody
2016-06-29  3:10 ` [Qemu-devel] [PULL 05/10] block/gluster: add support for selecting debug logging level Jeff Cody
2016-06-29  3:10 ` [Qemu-devel] [PULL 06/10] mirror: clarify mirror_do_read return code Jeff Cody
2016-06-29  3:10 ` [Qemu-devel] [PULL 07/10] mirror: limit niov to IOV_MAX elements, again Jeff Cody
2016-06-29  3:10 ` [Qemu-devel] [PULL 08/10] iotests: add small-granularity mirror test Jeff Cody
2016-06-29  3:10 ` [Qemu-devel] [PULL 09/10] blockjob: assert(cb) when create job Jeff Cody
2016-06-29  3:10 ` [Qemu-devel] [PULL 10/10] mirror: fix misleading comments Jeff Cody
2016-06-29 18:13 ` [Qemu-devel] [PULL 00/10] Block patches Peter Maydell

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=1467169844-18111-4-git-send-email-jcody@redhat.com \
    --to=jcody@redhat.com \
    --cc=peter.maydell@linaro.org \
    --cc=qemu-block@nongnu.org \
    --cc=qemu-devel@nongnu.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).