linux-kernel - [PATCH 4.20 041/183] qed: Fix qed_chain_set_prod() for PBL chains with non power of 2 page count

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]

Message-Id: <20190225195101.558437663@linuxfoundation.org>
Date:   Mon, 25 Feb 2019 22:10:14 +0100
From:   Greg Kroah-Hartman <gregkh@...uxfoundation.org>
To:     linux-kernel@...r.kernel.org
Cc:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        stable@...r.kernel.org, Denis Bolotin <dbolotin@...vell.com>,
        Ariel Elior <aelior@...vell.com>,
        "David S. Miller" <davem@...emloft.net>,
        Sasha Levin <sashal@...nel.org>
Subject: [PATCH 4.20 041/183] qed: Fix qed_chain_set_prod() for PBL chains with non power of 2 page count

4.20-stable review patch.  If anyone has any objections, please let me know.

------------------

[ Upstream commit 2d533a9287f2011632977e87ce2783f4c689c984 ]

In PBL chains with non power of 2 page count, the producer is not at the
beginning of the chain when index is 0 after a wrap. Therefore, after the
producer index wrap around, page index should be calculated more carefully.

Signed-off-by: Denis Bolotin <dbolotin@...vell.com>
Signed-off-by: Ariel Elior <aelior@...vell.com>
Signed-off-by: David S. Miller <davem@...emloft.net>
Signed-off-by: Sasha Levin <sashal@...nel.org>
---
 include/linux/qed/qed_chain.h | 31 +++++++++++++++++++++++++++++++
 1 file changed, 31 insertions(+)

diff --git a/include/linux/qed/qed_chain.h b/include/linux/qed/qed_chain.h
index 59ddf9af909e4..2dd0a9ed5b361 100644
--- a/include/linux/qed/qed_chain.h
+++ b/include/linux/qed/qed_chain.h
@@ -663,6 +663,37 @@ static inline void *qed_chain_get_last_elem(struct qed_chain *p_chain)
 static inline void qed_chain_set_prod(struct qed_chain *p_chain,
 				      u32 prod_idx, void *p_prod_elem)
 {
+	if (p_chain->mode == QED_CHAIN_MODE_PBL) {
+		u32 cur_prod, page_mask, page_cnt, page_diff;
+
+		cur_prod = is_chain_u16(p_chain) ? p_chain->u.chain16.prod_idx :
+			   p_chain->u.chain32.prod_idx;
+
+		/* Assume that number of elements in a page is power of 2 */
+		page_mask = ~p_chain->elem_per_page_mask;
+
+		/* Use "cur_prod - 1" and "prod_idx - 1" since producer index
+		 * reaches the first element of next page before the page index
+		 * is incremented. See qed_chain_produce().
+		 * Index wrap around is not a problem because the difference
+		 * between current and given producer indices is always
+		 * positive and lower than the chain's capacity.
+		 */
+		page_diff = (((cur_prod - 1) & page_mask) -
+			     ((prod_idx - 1) & page_mask)) /
+			    p_chain->elem_per_page;
+
+		page_cnt = qed_chain_get_page_cnt(p_chain);
+		if (is_chain_u16(p_chain))
+			p_chain->pbl.c.u16.prod_page_idx =
+				(p_chain->pbl.c.u16.prod_page_idx -
+				 page_diff + page_cnt) % page_cnt;
+		else
+			p_chain->pbl.c.u32.prod_page_idx =
+				(p_chain->pbl.c.u32.prod_page_idx -
+				 page_diff + page_cnt) % page_cnt;
+	}
+
 	if (is_chain_u16(p_chain))
 		p_chain->u.chain16.prod_idx = (u16) prod_idx;
 	else
-- 
2.19.1