All of lore.kernel.org
 help / color / mirror / Atom feed
From: Ivo van Doorn <ivdoorn@gmail.com>
To: "John W. Linville" <linville@tuxdriver.com>
Cc: linux-wireless@vger.kernel.org, rt2400-devel@lists.sourceforge.net
Subject: [PATCH 12/24] rt2x00: Fix system freeze on device removal
Date: Sun, 16 Sep 2007 14:18:49 +0200	[thread overview]
Message-ID: <200709161418.51894.IvDoorn@gmail.com> (raw)
In-Reply-To: <200709161403.11332.IvDoorn@gmail.com>

When a PCI device is unplugged reading from the
register will return the value from the state of just
before the device is unplugged.

When a rt61pci device was removed while the txdone
handler was running it ended in a endless loop because
it would run the txdone handler from the same frame
(the one last indicated by STA_CSR4) over and over again.

Check in the loop if the freshly read register value is
the same as the previously one and bail out if this is the case.

Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
---
 drivers/net/wireless/rt2x00/rt61pci.c |   22 ++++++++++++++++++----
 1 files changed, 18 insertions(+), 4 deletions(-)

diff --git a/drivers/net/wireless/rt2x00/rt61pci.c b/drivers/net/wireless/rt2x00/rt61pci.c
index 6199606..0c53b61 100644
--- a/drivers/net/wireless/rt2x00/rt61pci.c
+++ b/drivers/net/wireless/rt2x00/rt61pci.c
@@ -1801,23 +1801,37 @@ static void rt61pci_txdone(struct rt2x00_dev *rt2x00dev)
 	struct data_desc *txd;
 	u32 word;
 	u32 reg;
+	u32 old_reg;
+	int type;
 	int index;
 	int tx_status;
 	int retry;
 
+	/*
+	 * During each loop we will compare the freshly read
+	 * STA_CSR4 register value with the value read from
+	 * the previous loop. If the 2 values are equal then
+	 * we should stop processing because the chance it
+	 * quite big that the device has been unplugged and
+	 * we risk going into an endless loop.
+	 */
+	old_reg = 0;
+
 	while (1) {
 		rt2x00pci_register_read(rt2x00dev, STA_CSR4, &reg);
 		if (!rt2x00_get_field32(reg, STA_CSR4_VALID))
 			break;
 
+		if (old_reg == reg)
+			break;
+		old_reg = reg;
+
 		/*
 		 * Skip this entry when it contains an invalid
 		 * ring identication number.
 		 */
-		ring =
-		    rt2x00lib_get_ring(rt2x00dev,
-				       rt2x00_get_field32(reg,
-							  STA_CSR4_PID_TYPE));
+		type = rt2x00_get_field32(reg, STA_CSR4_PID_TYPE);
+		ring = rt2x00lib_get_ring(rt2x00dev, type);
 		if (unlikely(!ring))
 			continue;
 
-- 
1.5.3

  parent reply	other threads:[~2007-09-16 12:08 UTC|newest]

Thread overview: 24+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <200709161403.11332.IvDoorn@gmail.com>
2007-09-16 12:17 ` [PATCH 1/24] rt2x00: Remove firmware not-NULL check Ivo van Doorn
2007-09-16 12:17 ` [PATCH 2/24] rt2x00: Don't check for IEEE80211_TXCTL_REQ_TX_STATUS Ivo van Doorn
2007-09-16 12:18 ` [PATCH 3/24] rt2x00: Cleanup rxdone Ivo van Doorn
2007-09-16 12:18 ` [PATCH 4/24] rt2x00: Don't allow configuration calls when uninitialized Ivo van Doorn
2007-09-16 12:18 ` [PATCH 5/24] rt2x00: Fix rt61pci and rt73usb beacon handling Ivo van Doorn
2007-09-16 12:18 ` [PATCH 6/24] rt2x00: Recalculate link quality Ivo van Doorn
2007-09-16 12:18 ` [PATCH 7/24] rt2x00: Cleanup entry->flags Ivo van Doorn
2007-09-16 12:18 ` [PATCH 8/24] rt2x00: Reduce LNA flags Ivo van Doorn
2007-09-16 12:18 ` [PATCH 9/24] rt2x00: Rework RT61 and RT73 Antenna handling Ivo van Doorn
2007-09-16 12:18 ` [PATCH 10/24] rt2x00: Rename DEVICE_SUPPORT_ATIM to REQUIRE_BEACON_RING Ivo van Doorn
2007-09-16 12:18 ` [PATCH 11/24] rt2x00: Remove rt2x00mac_reset() Ivo van Doorn
2007-09-16 12:18 ` Ivo van Doorn [this message]
2007-09-16 12:18 ` [PATCH 13/24] rt2x00: Reduce magic value writing to device Ivo van Doorn
2007-09-16 12:19 ` [PATCH 14/24] rt2x00: New USB ID's for rt73usb and rt2500usb Ivo van Doorn
2007-09-16 12:19 ` [PATCH 15/24] rt2x00: Beacon ring entries should have QID_MGMT Ivo van Doorn
2007-09-16 12:19 ` [PATCH 16/24] rt2x00: Fix DEV_RATEBIT_ definitions Ivo van Doorn
2007-09-16 12:19 ` [PATCH 17/24] rt2x00: Fix rfkill handling Ivo van Doorn
2007-09-16 12:19 ` [PATCH 18/24] rt2x00: Merge allocation/free register components Ivo van Doorn
2007-09-16 12:19 ` [PATCH 19/24] rt2x00: macro's shouldn't use hidden arguments Ivo van Doorn
2007-09-16 12:19 ` [PATCH 20/24] rt2x00: Fix channel initialization Ivo van Doorn
2007-09-16 12:19 ` [PATCH 21/24] rt2x00: Add better CONFIG_PM checks Ivo van Doorn
2007-09-16 12:19 ` [PATCH 22/24] rt2x00: Add start/stop handlers Ivo van Doorn
2007-09-16 12:19 ` [PATCH 23/24] rt2x00: Add additional bit to MAX_FRAME_UNIT Ivo van Doorn
2007-09-16 12:19 ` [PATCH 24/24] rt2x00: Release rt2x00 2.0.8 Ivo van Doorn

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=200709161418.51894.IvDoorn@gmail.com \
    --to=ivdoorn@gmail.com \
    --cc=linux-wireless@vger.kernel.org \
    --cc=linville@tuxdriver.com \
    --cc=rt2400-devel@lists.sourceforge.net \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.