linux-kernel - [PATCH 1/4] lightnvm: update closed list outside of intr context

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Sat, 20 Feb 2016 08:52:40 +0100
From:	Matias Bjørling <m@...rling.me>
To:	linux-block@...r.kernel.org, linux-kernel@...r.kernel.org,
	axboe@...com
Cc:	Javier González <javier@...xlabs.com>,
	Matias Bjørling <m@...rling.me>
Subject: [PATCH 1/4] lightnvm: update closed list outside of intr context

From: Javier González <javier@...xlabs.com>

When an I/O finishes, full blocks are moved from the open to the closed
list - a lock is taken to protect the list. This happens at the moment
in the interrupt context, which is not correct.

This patch moves this logic to the block workqueue instead, avoiding
holding a spinlock without interrupt save in an interrupt context.

Signed-off-by: Javier González <javier@...xlabs.com>
Fixes: ff0e498bfa18 ("lightnvm: manage open and closed blocks sepa...")
Signed-off-by: Matias Bjørling <m@...rling.me>
---
 drivers/lightnvm/rrpc.c | 23 ++++++++++-------------
 1 file changed, 10 insertions(+), 13 deletions(-)

diff --git a/drivers/lightnvm/rrpc.c b/drivers/lightnvm/rrpc.c
index 307db1e..b7ddfb3 100644
--- a/drivers/lightnvm/rrpc.c
+++ b/drivers/lightnvm/rrpc.c
@@ -499,12 +499,21 @@ static void rrpc_gc_queue(struct work_struct *work)
 	struct rrpc *rrpc = gcb->rrpc;
 	struct rrpc_block *rblk = gcb->rblk;
 	struct nvm_lun *lun = rblk->parent->lun;
+	struct nvm_block *blk = rblk->parent;
 	struct rrpc_lun *rlun = &rrpc->luns[lun->id - rrpc->lun_offset];
 
 	spin_lock(&rlun->lock);
 	list_add_tail(&rblk->prio, &rlun->prio_list);
 	spin_unlock(&rlun->lock);
 
+	spin_lock(&lun->lock);
+	lun->nr_open_blocks--;
+	lun->nr_closed_blocks++;
+	blk->state &= ~NVM_BLK_ST_OPEN;
+	blk->state |= NVM_BLK_ST_CLOSED;
+	list_move_tail(&rblk->list, &rlun->closed_list);
+	spin_unlock(&lun->lock);
+
 	mempool_free(gcb, rrpc->gcb_pool);
 	pr_debug("nvm: block '%lu' is full, allow GC (sched)\n",
 							rblk->parent->id);
@@ -668,20 +677,8 @@ static void rrpc_end_io_write(struct rrpc *rrpc, struct rrpc_rq *rrqd,
 		lun = rblk->parent->lun;
 
 		cmnt_size = atomic_inc_return(&rblk->data_cmnt_size);
-		if (unlikely(cmnt_size == rrpc->dev->pgs_per_blk)) {
-			struct nvm_block *blk = rblk->parent;
-			struct rrpc_lun *rlun = rblk->rlun;
-
-			spin_lock(&lun->lock);
-			lun->nr_open_blocks--;
-			lun->nr_closed_blocks++;
-			blk->state &= ~NVM_BLK_ST_OPEN;
-			blk->state |= NVM_BLK_ST_CLOSED;
-			list_move_tail(&rblk->list, &rlun->closed_list);
-			spin_unlock(&lun->lock);
-
+		if (unlikely(cmnt_size == rrpc->dev->pgs_per_blk))
 			rrpc_run_gc(rrpc, rblk);
-		}
 	}
 }
 
-- 
2.5.0