From: Jeff Liu <jeff.liu@oracle.com>
To: xfs@oss.sgi.com
Subject: [PATCH 05/15] xfs: Teach inode/space allocator aware of XFS_AG_STATE_ALLOC_DENY
Date: Fri, 16 Nov 2012 14:45:31 +0800 [thread overview]
Message-ID: <50A5E10B.5010109@oracle.com> (raw)
Teach inode/space allocators aware of the new AG state(XFS_AG_STATE_ALLOC_DENY).
- Don't chose an AG with this state.
- Don't mark pagi_inodeok upon an AG for xfs_set_inode32/64 mount options if the
given AG incore object has this state.
Signed-off-by: Jie Liu <jeff.liu@oracle.com>
---
fs/xfs/xfs_alloc.c | 15 ++++++++++++++-
fs/xfs/xfs_ialloc.c | 6 +++++-
fs/xfs/xfs_super.c | 10 ++++++++++
3 files changed, 29 insertions(+), 2 deletions(-)
diff --git a/fs/xfs/xfs_alloc.c b/fs/xfs/xfs_alloc.c
index bd9cc41..d2eb856 100644
--- a/fs/xfs/xfs_alloc.c
+++ b/fs/xfs/xfs_alloc.c
@@ -493,6 +493,7 @@ STATIC int /* error */
xfs_alloc_ag_vextent(
xfs_alloc_arg_t *args) /* argument structure for allocation */
{
+ xfs_perag_t *pag = args->pag;
int error=0;
ASSERT(args->minlen > 0);
@@ -500,6 +501,18 @@ xfs_alloc_ag_vextent(
ASSERT(args->minlen <= args->maxlen);
ASSERT(args->mod < args->prod);
ASSERT(args->alignment > 0);
+
+ /*
+ * Return NULLAGBLOCK if the a.g. state is set to offline.
+ * i.e. deny allocation.
+ * FIXME: how about replacing XFS_AG_STATE_ALLOC_DENY to
+ * XFS_AG_STATE_OFFLINE? maybe looks a bit more reasonable.
+ */
+ if (unlikely(pag->pag_state & XFS_AG_STATE_ALLOC_DENY)) {
+ args->agbno = NULLAGBLOCK;
+ return 0;
+ }
+
/*
* Branch to correct routine based on the type.
*/
@@ -1893,7 +1906,7 @@ xfs_alloc_fix_freelist(
/*
* Stop if we run out. Won't happen if callers are obeying
* the restrictions correctly. Can happen for free calls
- * on a completely full ag.
+ * on a completely full ag, or an ag is set to offline.
*/
if (targs.agbno == NULLAGBLOCK) {
if (flags & XFS_ALLOC_FLAG_FREEING)
diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c
index 445bf1a..04fbb49 100644
--- a/fs/xfs/xfs_ialloc.c
+++ b/fs/xfs/xfs_ialloc.c
@@ -488,7 +488,11 @@ xfs_ialloc_ag_select(
flags = XFS_ALLOC_FLAG_TRYLOCK;
for (;;) {
pag = xfs_perag_get(mp, agno);
- if (!pag->pagi_inodeok) {
+ /*
+ * Skip the current AG if it's state is set to offline.
+ */
+ if (!pag->pagi_inodeok ||
+ unlikely(pag->pag_state & XFS_AG_STATE_ALLOC_DENY)) {
xfs_ialloc_next_ag(mp);
goto nextag;
}
diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c
index 26a09bd..e01f64c 100644
--- a/fs/xfs/xfs_super.c
+++ b/fs/xfs/xfs_super.c
@@ -642,6 +642,11 @@ xfs_set_inode32(struct xfs_mount *mp)
}
pag = xfs_perag_get(mp, index);
+ if (unlikely(pag->pag_state & XFS_AG_STATE_ALLOC_DENY)) {
+ xfs_perag_put(pag);
+ continue;
+ }
+
pag->pagi_inodeok = 1;
maxagi++;
if (index < max_metadata)
@@ -663,6 +668,11 @@ xfs_set_inode64(struct xfs_mount *mp)
struct xfs_perag *pag;
pag = xfs_perag_get(mp, index);
+ if (unlikely(pag->pag_state & XFS_AG_STATE_ALLOC_DENY)) {
+ xfs_perag_put(pag);
+ continue;
+ }
+
pag->pagi_inodeok = 1;
pag->pagf_metadata = 0;
xfs_perag_put(pag);
--
1.7.4.1
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
reply other threads:[~2012-11-16 6:43 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
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=50A5E10B.5010109@oracle.com \
--to=jeff.liu@oracle.com \
--cc=xfs@oss.sgi.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.