cluster-devel.redhat.com archive mirror
 help / color / mirror / Atom feed
From: Andreas Gruenbacher <agruenba@redhat.com>
To: cluster-devel.redhat.com
Subject: [Cluster-devel] [PATCH 09/17] gfs2: Also reflect single-block allocations in rgd->rd_extfail_pt
Date: Fri, 30 Oct 2020 22:51:44 +0100	[thread overview]
Message-ID: <20201030215152.1567790-10-agruenba@redhat.com> (raw)
In-Reply-To: <20201030215152.1567790-1-agruenba@redhat.com>

Pass a non-NULL minext to gfs2_rbm_find even for single-block allocations.  In
gfs2_rbm_find, also set rgd->rd_extfail_pt when a single-block allocation
fails in a resource group: there is no reason for treating that case
differently.  In gfs2_reservation_check_and_update, only check how many free
blocks we have if more than one block is requested; we already know there's at
least one free block.

In addition, when allocating N blocks fails in gfs2_rbm_find, we need to set
rd_extfail_pt to N - 1 rather than N:  rd_extfail_pt defines the biggest
allocation that might still succeed.

Finally, reset rd_extfail_pt when updating the resource group statistics in
update_rgrp_lvb, as we already do in gfs2_rgrp_bh_get.

Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
---
 fs/gfs2/rgrp.c | 18 ++++++++++--------
 1 file changed, 10 insertions(+), 8 deletions(-)

diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c
index 5f72ae9677a7..8ed72a8dbb94 100644
--- a/fs/gfs2/rgrp.c
+++ b/fs/gfs2/rgrp.c
@@ -1274,6 +1274,8 @@ static int update_rgrp_lvb(struct gfs2_rgrpd *rgd)
 	rgd->rd_free = be32_to_cpu(rgd->rd_rgl->rl_free);
 	rgrp_set_bitmap_flags(rgd);
 	rgd->rd_free_clone = rgd->rd_free;
+	/* max out the rgrp allocation failure point */
+	rgd->rd_extfail_pt = rgd->rd_free;
 	rgd->rd_dinodes = be32_to_cpu(rgd->rd_rgl->rl_dinodes);
 	rgd->rd_igeneration = be64_to_cpu(rgd->rd_rgl->rl_igeneration);
 	return 0;
@@ -1669,7 +1671,7 @@ static int gfs2_reservation_check_and_update(struct gfs2_rbm *rbm,
 	 * If we have a minimum extent length, then skip over any extent
 	 * which is less than the min extent length in size.
 	 */
-	if (minext) {
+	if (minext > 1) {
 		extlen = gfs2_free_extlen(rbm, minext);
 		if (extlen <= maxext->len)
 			goto fail;
@@ -1704,7 +1706,7 @@ static int gfs2_reservation_check_and_update(struct gfs2_rbm *rbm,
  * gfs2_rbm_find - Look for blocks of a particular state
  * @rbm: Value/result starting position and final position
  * @state: The state which we want to find
- * @minext: Pointer to the requested extent length (NULL for a single block)
+ * @minext: Pointer to the requested extent length
  *          This is updated to be the actual reservation size.
  * @ip: If set, check for reservations
  * @nowrap: Stop looking at the end of the rgrp, rather than wrapping
@@ -1760,8 +1762,7 @@ static int gfs2_rbm_find(struct gfs2_rbm *rbm, u8 state, u32 *minext,
 		if (ip == NULL)
 			return 0;
 
-		ret = gfs2_reservation_check_and_update(rbm, ip,
-							minext ? *minext : 0,
+		ret = gfs2_reservation_check_and_update(rbm, ip, *minext,
 							&maxext);
 		if (ret == 0)
 			return 0;
@@ -1793,7 +1794,7 @@ static int gfs2_rbm_find(struct gfs2_rbm *rbm, u8 state, u32 *minext,
 			break;
 	}
 
-	if (minext == NULL || state != GFS2_BLKST_FREE)
+	if (state != GFS2_BLKST_FREE)
 		return -ENOSPC;
 
 	/* If the extent was too small, and it's smaller than the smallest
@@ -1801,7 +1802,7 @@ static int gfs2_rbm_find(struct gfs2_rbm *rbm, u8 state, u32 *minext,
 	   useless to search this rgrp again for this amount or more. */
 	if (wrapped && (scan_from_start || rbm->bii > last_bii) &&
 	    *minext < rbm->rgd->rd_extfail_pt)
-		rbm->rgd->rd_extfail_pt = *minext;
+		rbm->rgd->rd_extfail_pt = *minext - 1;
 
 	/* If the maximum extent we found is big enough to fulfill the
 	   minimum requirements, use it anyway. */
@@ -2375,14 +2376,15 @@ int gfs2_alloc_blocks(struct gfs2_inode *ip, u64 *bn, unsigned int *nblocks,
 	struct buffer_head *dibh;
 	struct gfs2_rbm rbm = { .rgd = ip->i_res.rs_rgd, };
 	u64 block; /* block, within the file system scope */
+	u32 minext = 1;
 	int error;
 
 	gfs2_set_alloc_start(&rbm, ip, dinode);
-	error = gfs2_rbm_find(&rbm, GFS2_BLKST_FREE, NULL, ip, false);
+	error = gfs2_rbm_find(&rbm, GFS2_BLKST_FREE, &minext, ip, false);
 
 	if (error == -ENOSPC) {
 		gfs2_set_alloc_start(&rbm, ip, dinode);
-		error = gfs2_rbm_find(&rbm, GFS2_BLKST_FREE, NULL, NULL, false);
+		error = gfs2_rbm_find(&rbm, GFS2_BLKST_FREE, &minext, NULL, false);
 	}
 
 	/* Since all blocks are reserved in advance, this shouldn't happen */
-- 
2.26.2



  parent reply	other threads:[~2020-10-30 21:51 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-10-30 21:51 [Cluster-devel] [PATCH 00/17] gfs2: resource group glock sharing Andreas Gruenbacher
2020-10-30 21:51 ` [Cluster-devel] [PATCH 01/17] gfs2: Turn gfs2_rbm_incr into gfs2_rbm_add Andreas Gruenbacher
2020-10-30 21:51 ` [Cluster-devel] [PATCH 02/17] gfs2: Only use struct gfs2_rbm for bitmap manipulations Andreas Gruenbacher
2020-10-30 21:51 ` [Cluster-devel] [PATCH 03/17] gfs2: Get rid of unnecessary variable in gfs2_alloc_blocks Andreas Gruenbacher
2020-10-30 21:51 ` [Cluster-devel] [PATCH 04/17] gfs2: Minor gfs2_inplace_reserve cleanup Andreas Gruenbacher
2020-10-30 21:51 ` [Cluster-devel] [PATCH 05/17] Revert "gfs2: Don't reject a supposedly full bitmap if we have blocks reserved" Andreas Gruenbacher
2020-10-30 21:51 ` [Cluster-devel] [PATCH 06/17] gfs2: Don't clear GBF_FULL flags in rs_deltree Andreas Gruenbacher
2020-10-30 21:51 ` [Cluster-devel] [PATCH 07/17] gfs2: Set GBF_FULL flags when reading resource group Andreas Gruenbacher
2020-10-30 21:51 ` [Cluster-devel] [PATCH 08/17] gfs2: When looking for blocks to allocate, don't clamp at free blocks Andreas Gruenbacher
2020-10-30 21:51 ` Andreas Gruenbacher [this message]
2020-10-30 21:51 ` [Cluster-devel] [PATCH 10/17] gfs2: Only pass reservation down to gfs2_rbm_find Andreas Gruenbacher
2020-10-30 21:51 ` [Cluster-devel] [PATCH 11/17] gfs: Don't search for unreserved space twice Andreas Gruenbacher
2020-10-30 21:51 ` [Cluster-devel] [PATCH 12/17] gfs2: Check for active reservation in gfs2_release Andreas Gruenbacher
2020-10-30 21:51 ` [Cluster-devel] [PATCH 13/17] gfs2: Rename rs_{free -> requested} and rd_{reserved -> requested} Andreas Gruenbacher
2020-10-30 21:51 ` [Cluster-devel] [PATCH 14/17] gfs2: Add per-reservation reserved block accounting Andreas Gruenbacher
2020-10-30 21:51 ` [Cluster-devel] [PATCH 15/17] gfs2: Add local resource group locking Andreas Gruenbacher
2020-10-30 21:51 ` [Cluster-devel] [PATCH 16/17] gfs2: Allow node-wide exclusive glock sharing Andreas Gruenbacher
2020-10-30 21:51 ` [Cluster-devel] [PATCH 17/17] gfs2: Use resource group " Andreas Gruenbacher

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=20201030215152.1567790-10-agruenba@redhat.com \
    --to=agruenba@redhat.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).