linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	stable@vger.kernel.org, Alex Vesker <valex@mellanox.com>,
	Saeed Mahameed <saeedm@mellanox.com>
Subject: [PATCH 3.18 25/29] net/mlx5: Fix command interface race in polling mode
Date: Fri, 20 Jul 2018 14:11:20 +0200	[thread overview]
Message-ID: <20180720115159.880064240@linuxfoundation.org> (raw)
In-Reply-To: <20180720115158.974693829@linuxfoundation.org>

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

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

From: Alex Vesker <valex@mellanox.com>

[ Upstream commit d412c31dae053bf30a1bc15582a9990df297a660 ]

The command interface can work in two modes: Events and Polling.
In the general case, each time we invoke a command, a work is
queued to handle it.

When working in events, the interrupt handler completes the
command execution. On the other hand, when working in polling
mode, the work itself completes it.

Due to a bug in the work handler, a command could have been
completed by the interrupt handler, while the work handler
hasn't finished yet, causing the it to complete once again
if the command interface mode was changed from Events to
polling after the interrupt handler was called.

mlx5_unload_one()
        mlx5_stop_eqs()
                // Destroy the EQ before cmd EQ
                ...cmd_work_handler()
                        write_doorbell()
                        --> EVENT_TYPE_CMD
                                mlx5_cmd_comp_handler() // First free
                                        free_ent(cmd, ent->idx)
                                        complete(&ent->done)

        <-- mlx5_stop_eqs //cmd was complete
                // move to polling before destroying the last cmd EQ
                mlx5_cmd_use_polling()
                        cmd->mode = POLL;

                --> cmd_work_handler (continues)
                        if (cmd->mode == POLL)
                                mlx5_cmd_comp_handler() // Double free

The solution is to store the cmd->mode before writing the doorbell.

Fixes: e126ba97dba9 ("mlx5: Add driver for Mellanox Connect-IB adapters")
Signed-off-by: Alex Vesker <valex@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 drivers/net/ethernet/mellanox/mlx5/core/cmd.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

--- a/drivers/net/ethernet/mellanox/mlx5/core/cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/cmd.c
@@ -480,6 +480,7 @@ static void cmd_work_handler(struct work
 	struct mlx5_core_dev *dev = container_of(cmd, struct mlx5_core_dev, cmd);
 	struct mlx5_cmd_layout *lay;
 	struct semaphore *sem;
+	int cmd_mode;
 
 	sem = ent->page_queue ? &cmd->pages_sem : &cmd->sem;
 	down(sem);
@@ -513,13 +514,14 @@ static void cmd_work_handler(struct work
 	set_signature(ent, !cmd->checksum_disabled);
 	dump_command(dev, ent, 1);
 	ent->ts1 = ktime_get_ns();
+	cmd_mode = cmd->mode;
 
 	/* ring doorbell after the descriptor is valid */
 	wmb();
 	iowrite32be(1 << ent->idx, &dev->iseg->cmd_dbell);
 	mlx5_core_dbg(dev, "write 0x%x to command doorbell\n", 1 << ent->idx);
 	mmiowb();
-	if (cmd->mode == CMD_MODE_POLLING) {
+	if (cmd_mode == CMD_MODE_POLLING) {
 		poll_timeout(ent);
 		/* make sure we read the descriptor after ownership is SW */
 		rmb();



  parent reply	other threads:[~2018-07-20 13:00 UTC|newest]

Thread overview: 32+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-07-20 12:10 [PATCH 3.18 00/29] 3.18.116-stable review Greg Kroah-Hartman
2018-07-20 12:10 ` [PATCH 3.18 01/29] ibmasm: dont write out of bounds in read handler Greg Kroah-Hartman
2018-07-20 12:10 ` [PATCH 3.18 02/29] USB: serial: keyspan_pda: fix modem-status error handling Greg Kroah-Hartman
2018-07-20 12:10 ` [PATCH 3.18 03/29] USB: yurex: fix out-of-bounds uaccess in read handler Greg Kroah-Hartman
2018-07-20 12:10 ` [PATCH 3.18 04/29] USB: serial: mos7840: fix status-register error handling Greg Kroah-Hartman
2018-07-20 12:11 ` [PATCH 3.18 05/29] usb: quirks: add delay quirks for Corsair Strafe Greg Kroah-Hartman
2018-07-20 12:11 ` [PATCH 3.18 06/29] xhci: xhci-mem: off by one in xhci_stream_id_to_ring() Greg Kroah-Hartman
2018-07-20 12:11 ` [PATCH 3.18 07/29] Fix up non-directory creation in SGID directories Greg Kroah-Hartman
2018-07-20 12:11 ` [PATCH 3.18 08/29] netfilter: x_tables: initialise match/target check parameter struct Greg Kroah-Hartman
2018-07-20 12:11 ` [PATCH 3.18 09/29] loop: add recursion validation to LOOP_CHANGE_FD Greg Kroah-Hartman
2018-07-20 12:11 ` [PATCH 3.18 10/29] PM / hibernate: Fix oops at snapshot_write() Greg Kroah-Hartman
2018-07-20 12:11 ` [PATCH 3.18 11/29] RDMA/ucm: Mark UCM interface as BROKEN Greg Kroah-Hartman
2018-07-20 12:11 ` [PATCH 3.18 12/29] loop: remember whether sysfs_create_group() was done Greg Kroah-Hartman
2018-07-20 12:11 ` [PATCH 3.18 13/29] bcm63xx_enet: correct clock usage Greg Kroah-Hartman
2018-07-20 12:11 ` [PATCH 3.18 14/29] bcm63xx_enet: do not write to random DMA channel on BCM6345 Greg Kroah-Hartman
2018-07-20 12:11 ` [PATCH 3.18 15/29] crypto: crypto4xx - remove bad list_del Greg Kroah-Hartman
2018-07-20 12:11 ` [PATCH 3.18 16/29] crypto: crypto4xx - fix crypto4xx_build_pdr, crypto4xx_build_sdr leak Greg Kroah-Hartman
2018-07-20 12:11 ` [PATCH 3.18 17/29] net: dccp: avoid crash in ccid3_hc_rx_send_feedback() Greg Kroah-Hartman
2018-07-20 12:11 ` [PATCH 3.18 18/29] net: dccp: switch rx_tstamp_last_feedback to monotonic clock Greg Kroah-Hartman
2018-07-20 12:11 ` [PATCH 3.18 19/29] net/mlx5: Fix incorrect raw command length parsing Greg Kroah-Hartman
2018-07-20 12:11 ` [PATCH 3.18 20/29] net: sungem: fix rx checksum support Greg Kroah-Hartman
2018-07-20 12:11 ` [PATCH 3.18 21/29] tcp: fix Fast Open key endianness Greg Kroah-Hartman
2018-07-20 12:11 ` [PATCH 3.18 23/29] vhost_net: validate sock before trying to put its fd Greg Kroah-Hartman
2018-07-20 12:11 ` [PATCH 3.18 24/29] net_sched: blackhole: tell upper qdisc about dropped packets Greg Kroah-Hartman
2018-07-20 12:11 ` Greg Kroah-Hartman [this message]
2018-07-20 12:11 ` [PATCH 3.18 26/29] netfilter: ebtables: reject non-bridge targets Greg Kroah-Hartman
2018-07-20 12:11 ` [PATCH 3.18 27/29] KEYS: DNS: fix parsing multiple options Greg Kroah-Hartman
2018-07-20 12:11 ` [PATCH 3.18 28/29] rds: avoid unenecessary cong_update in loop transport Greg Kroah-Hartman
2018-07-20 12:11 ` [PATCH 3.18 29/29] net/nfc: Avoid stalls when nfc_alloc_send_skb() returned NULL Greg Kroah-Hartman
2018-07-20 13:33 ` [PATCH 3.18 00/29] 3.18.116-stable review Nathan Chancellor
2018-07-21 13:38 ` Guenter Roeck
2018-07-22 11:42   ` Greg Kroah-Hartman

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=20180720115159.880064240@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=saeedm@mellanox.com \
    --cc=stable@vger.kernel.org \
    --cc=valex@mellanox.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).