ocfs2-devel.oss.oracle.com archive mirror
 help / color / mirror / Atom feed
From: Tao Ma <tao.ma@oracle.com>
To: ocfs2-devel@oss.oracle.com
Subject: [Ocfs2-devel] [PATCH v2] ocfs2: make ocfs2_adjust_resv_from_alloc simple.
Date: Thu,  8 Apr 2010 16:33:02 +0800	[thread overview]
Message-ID: <1270715582-13702-1-git-send-email-tao.ma@oracle.com> (raw)
In-Reply-To: <20100407212423.GP11402@wotan.suse.de>

When we allocate some bits from the reservation, we always
allocate from the r_start(see ocfs2_resmap_resv_bits).
So there should be no sense for checking between r_start
and start. And I don't think we will change this behaviour
somehow later by allocating from some bits after r_start.
Why not make ocfs2_adjust_resv_from_alloc simple now?

So the only chance we have to adjust the reservation is that
we haven't reached the end. With this patch, the function is
more readable.

Note:
btw, this patch also fixes an original bug in the function
which I haven't found before.
	if (end < ocfs2_resv_end(resv))
		rhs = end - ocfs2_resv_end(resv);
This code is of course buggy. ;)

Cc: Mark Fasheh <mfasheh@suse.com>
Signed-off-by: Tao Ma <tao.ma@oracle.com>
---
 fs/ocfs2/reservations.c |   32 ++++++++++++--------------------
 fs/ocfs2/reservations.h |    3 ++-
 2 files changed, 14 insertions(+), 21 deletions(-)

diff --git a/fs/ocfs2/reservations.c b/fs/ocfs2/reservations.c
index 32bad4a..4065002 100644
--- a/fs/ocfs2/reservations.c
+++ b/fs/ocfs2/reservations.c
@@ -406,7 +406,7 @@ ocfs2_find_resv_lhs(struct ocfs2_reservation_map *resmap, unsigned int goal)
  * The start value of *rstart is insignificant.
  *
  * This function searches the bitmap range starting at search_start
- * with length csearch_len for a set of contiguous free bits. We try
+ * with length search_len for a set of contiguous free bits. We try
  * to find up to 'wanted' bits, but can sometimes return less.
  *
  * Returns the length of allocation, 0 if no free bits are found.
@@ -776,38 +776,28 @@ static void
 				     struct ocfs2_alloc_reservation *resv,
 				     unsigned int start, unsigned int end)
 {
-	unsigned int lhs = 0, rhs = 0;
+	unsigned int rhs = 0;
+	unsigned int old_end = ocfs2_resv_end(resv);
 
-	BUG_ON(start < resv->r_start);
+	BUG_ON(start != resv->r_start || old_end < end);
 
 	/*
 	 * Completely used? We can remove it then.
 	 */
-	if (ocfs2_resv_end(resv) <= end && resv->r_start >= start) {
+	if (old_end == end) {
 		__ocfs2_resv_discard(resmap, resv);
 		return;
 	}
 
-	if (end < ocfs2_resv_end(resv))
-		rhs = end - ocfs2_resv_end(resv);
-
-	if (start > resv->r_start)
-		lhs = start - resv->r_start;
+	rhs = old_end - end;
 
 	/*
-	 * This should have been trapped above. At the very least, rhs
-	 * should be non zero.
+	 * This should have been trapped above.
 	 */
-	BUG_ON(rhs == 0 && lhs == 0);
-
-	if (rhs >= lhs) {
-		unsigned int old_end = ocfs2_resv_end(resv);
+	BUG_ON(rhs == 0);
 
-		resv->r_start = end + 1;
-		resv->r_len = old_end - resv->r_start + 1;
-	} else {
-		resv->r_len = start - resv->r_start;
-	}
+	resv->r_start = end + 1;
+	resv->r_len = old_end - resv->r_start + 1;
 }
 
 void ocfs2_resmap_claimed_bits(struct ocfs2_reservation_map *resmap,
@@ -822,6 +812,8 @@ void ocfs2_resmap_claimed_bits(struct ocfs2_reservation_map *resmap,
 	if (resv == NULL)
 		return;
 
+	BUG_ON(cstart != resv->r_start);
+
 	spin_lock(&resv_lock);
 
 	mlog(0, "claim bits: cstart: %u cend: %u clen: %u r_start: %u "
diff --git a/fs/ocfs2/reservations.h b/fs/ocfs2/reservations.h
index 25b0c0e..1e49cc2 100644
--- a/fs/ocfs2/reservations.h
+++ b/fs/ocfs2/reservations.h
@@ -149,7 +149,8 @@ int ocfs2_resmap_resv_bits(struct ocfs2_reservation_map *resmap,
  * reservation. But we must always call this function when bits are claimed.
  * Internally, the reservations code will use this information to mark the
  * reservations bitmap. If resv is passed, it's next allocation window will be
- * calculated.
+ * calculated. It also expects that 'cstart' is the same as we passed back
+ * from ocfs2_resmap_resv_bits().
  */
 void ocfs2_resmap_claimed_bits(struct ocfs2_reservation_map *resmap,
 			       struct ocfs2_alloc_reservation *resv,
-- 
1.5.5

  parent reply	other threads:[~2010-04-08  8:33 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-04-06  7:40 [Ocfs2-devel] [PATCH] ocfs2: make ocfs2_adjust_resv_from_alloc simple Tao Ma
2010-04-07 21:24 ` Mark Fasheh
2010-04-07 23:47   ` Tao Ma
2010-04-08  8:33   ` Tao Ma [this message]
2010-04-23 21:44     ` [Ocfs2-devel] [PATCH v2] " Mark Fasheh
2010-04-23 22:03     ` Joel Becker
2010-04-26  6:45       ` Tao Ma

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=1270715582-13702-1-git-send-email-tao.ma@oracle.com \
    --to=tao.ma@oracle.com \
    --cc=ocfs2-devel@oss.oracle.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).