public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Greg KH <gregkh@suse.de>
To: linux-kernel@vger.kernel.org, stable@kernel.org,
	openib-general@openib.org, Roland Dreier <rolandd@cisco.com>
Cc: Justin Forbes <jmforbes@linuxtx.org>,
	Zwane Mwaikambo <zwane@arm.linux.org.uk>,
	"Theodore Ts'o" <tytso@mit.edu>,
	Randy Dunlap <rdunlap@xenotime.net>,
	Dave Jones <davej@redhat.com>,
	Chuck Wolber <chuckw@quantumlinux.com>,
	Chris Wedgwood <reviews@ml.cw.f00f.org>,
	torvalds@osdl.org, akpm@osdl.org, alan@lxorguk.ukuu.org.uk,
	"Michael S. Tsirkin" <mst@mellanox.co.il>,
	Chris Wright <chrisw@sous-sol.org>,
	Greg Kroah-Hartman <gregkh@suse.de>
Subject: [patch 02/45] IB/mthca: restore missing PCI registers after reset
Date: Mon, 17 Jul 2006 09:25:31 -0700	[thread overview]
Message-ID: <20060717162531.GC4829@kroah.com> (raw)
In-Reply-To: <20060717162452.GA4829@kroah.com>

[-- Attachment #1: ib-mthca-restore-missing-pci-registers-after-reset.patch --]
[-- Type: text/plain, Size: 3584 bytes --]

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

------------------
mthca does not restore the following PCI-X/PCI Express registers after reset:
  PCI-X device: PCI-X command register
  PCI-X bridge: upstream and downstream split transaction registers
  PCI Express : PCI Express device control and link control registers

This causes instability and/or bad performance on systems where one of
these registers is set to a non-default value by BIOS.

Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Chris Wright <chrisw@sous-sol.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

---
 drivers/infiniband/hw/mthca/mthca_reset.c |   59 ++++++++++++++++++++++++++++++
 1 file changed, 59 insertions(+)

--- linux-2.6.17.2.orig/drivers/infiniband/hw/mthca/mthca_reset.c
+++ linux-2.6.17.2/drivers/infiniband/hw/mthca/mthca_reset.c
@@ -49,6 +49,12 @@ int mthca_reset(struct mthca_dev *mdev)
 	u32 *hca_header    = NULL;
 	u32 *bridge_header = NULL;
 	struct pci_dev *bridge = NULL;
+	int bridge_pcix_cap = 0;
+	int hca_pcie_cap = 0;
+	int hca_pcix_cap = 0;
+
+	u16 devctl;
+	u16 linkctl;
 
 #define MTHCA_RESET_OFFSET 0xf0010
 #define MTHCA_RESET_VALUE  swab32(1)
@@ -110,6 +116,9 @@ int mthca_reset(struct mthca_dev *mdev)
 		}
 	}
 
+	hca_pcix_cap = pci_find_capability(mdev->pdev, PCI_CAP_ID_PCIX);
+	hca_pcie_cap = pci_find_capability(mdev->pdev, PCI_CAP_ID_EXP);
+
 	if (bridge) {
 		bridge_header = kmalloc(256, GFP_KERNEL);
 		if (!bridge_header) {
@@ -129,6 +138,13 @@ int mthca_reset(struct mthca_dev *mdev)
 				goto out;
 			}
 		}
+		bridge_pcix_cap = pci_find_capability(bridge, PCI_CAP_ID_PCIX);
+		if (!bridge_pcix_cap) {
+				err = -ENODEV;
+				mthca_err(mdev, "Couldn't locate HCA bridge "
+					  "PCI-X capability, aborting.\n");
+				goto out;
+		}
 	}
 
 	/* actually hit reset */
@@ -178,6 +194,20 @@ int mthca_reset(struct mthca_dev *mdev)
 good:
 	/* Now restore the PCI headers */
 	if (bridge) {
+		if (pci_write_config_dword(bridge, bridge_pcix_cap + 0x8,
+				 bridge_header[(bridge_pcix_cap + 0x8) / 4])) {
+			err = -ENODEV;
+			mthca_err(mdev, "Couldn't restore HCA bridge Upstream "
+				  "split transaction control, aborting.\n");
+			goto out;
+		}
+		if (pci_write_config_dword(bridge, bridge_pcix_cap + 0xc,
+				 bridge_header[(bridge_pcix_cap + 0xc) / 4])) {
+			err = -ENODEV;
+			mthca_err(mdev, "Couldn't restore HCA bridge Downstream "
+				  "split transaction control, aborting.\n");
+			goto out;
+		}
 		/*
 		 * Bridge control register is at 0x3e, so we'll
 		 * naturally restore it last in this loop.
@@ -203,6 +233,35 @@ good:
 		}
 	}
 
+	if (hca_pcix_cap) {
+		if (pci_write_config_dword(mdev->pdev, hca_pcix_cap,
+				 hca_header[hca_pcix_cap / 4])) {
+			err = -ENODEV;
+			mthca_err(mdev, "Couldn't restore HCA PCI-X "
+				  "command register, aborting.\n");
+			goto out;
+		}
+	}
+
+	if (hca_pcie_cap) {
+		devctl = hca_header[(hca_pcie_cap + PCI_EXP_DEVCTL) / 4];
+		if (pci_write_config_word(mdev->pdev, hca_pcie_cap + PCI_EXP_DEVCTL,
+					   devctl)) {
+			err = -ENODEV;
+			mthca_err(mdev, "Couldn't restore HCA PCI Express "
+				  "Device Control register, aborting.\n");
+			goto out;
+		}
+		linkctl = hca_header[(hca_pcie_cap + PCI_EXP_LNKCTL) / 4];
+		if (pci_write_config_word(mdev->pdev, hca_pcie_cap + PCI_EXP_LNKCTL,
+					   linkctl)) {
+			err = -ENODEV;
+			mthca_err(mdev, "Couldn't restore HCA PCI Express "
+				  "Link control register, aborting.\n");
+			goto out;
+		}
+	}
+
 	for (i = 0; i < 16; ++i) {
 		if (i * 4 == PCI_COMMAND)
 			continue;

--

  parent reply	other threads:[~2006-07-17 16:34 UTC|newest]

Thread overview: 64+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <20060717160652.408007000@blue.kroah.org>
2006-07-17 16:24 ` [patch 00/45] 2.6.17.y -stable review Greg KH
2006-07-17 16:25   ` [patch 01/45] XFS: corruption fix Greg KH
2006-07-17 19:10     ` Chris Wedgwood
2006-07-18 13:27     ` Jan Engelhardt
2006-07-18 22:24       ` Nathan Scott
2006-07-17 16:25   ` Greg KH [this message]
2006-07-26 10:29     ` restore missing PCI registers after reset Michael S. Tsirkin
2006-07-26 16:20       ` Greg KH
2006-07-26 16:32         ` Michael S. Tsirkin
2006-07-26 16:42           ` Greg KH
2006-08-21 13:22             ` Michael S. Tsirkin
2006-07-17 16:25   ` [patch 03/45] x86_64: Fix modular pc speaker Greg KH
2006-07-17 18:02     ` Dmitry Torokhov
2006-07-17 16:25   ` [patch 04/45] BLOCK: Fix bounce limit address check Greg KH
2006-07-17 16:25   ` [patch 05/45] memory hotplug: solve config broken: undefined reference to `online_page Greg KH
2006-07-17 16:26   ` [patch 06/45] v4l/dvb: Fix budget-av frontend detection Greg KH
2006-07-17 16:26   ` [patch 07/45] v4l/dvb: Fix CI on old KNC1 DVBC cards Greg KH
2006-07-25 13:01     ` Edgar Hucek
2006-07-25 15:15       ` Michael Krufky
2006-07-25 15:22         ` Edgar Hucek
2006-07-25 17:10           ` Greg KH
2006-08-01 23:23             ` Michael Krufky
2006-07-17 16:26   ` [patch 08/45] v4l/dvb: Fix CI interface on PRO KNC1 cards Greg KH
2006-07-17 16:26   ` [patch 09/45] pnp: suppress request_irq() warning Greg KH
2006-07-17 16:26   ` [patch 10/45] Reduce ACPI verbosity on null handle condition Greg KH
2006-07-17 16:26   ` [patch 11/45] via-velocity: the link is not correctly detected when the device starts Greg KH
2006-07-17 16:26   ` [patch 12/45] 2 oopses in ethtool Greg KH
2006-07-17 20:51     ` Matthew Wilcox
2006-07-17 16:26   ` [patch 13/45] v4l/dvb: Kconfig: fix description and dependencies for saa7115 module Greg KH
2006-07-17 16:26   ` [patch 14/45] PKT_SCHED: Fix illegal memory dereferences when dumping actions Greg KH
2006-07-17 16:27   ` [patch 15/45] PKT_SCHED: Return ENOENT if action module is unavailable Greg KH
2006-07-17 16:27   ` [patch 16/45] PKT_SCHED: Fix error handling while dumping actions Greg KH
2006-07-17 16:27   ` [patch 17/45] v4l/dvb: Backport fix to artec USB DVB devices Greg KH
2006-07-17 16:27   ` [patch 18/45] v4l/dvb: Backport the DISEQC regression fix to 2.6.17.x Greg KH
2006-07-17 16:27   ` [patch 19/45] v4l/dvb: Backport the budget driver DISEQC instability fix Greg KH
2006-07-17 16:27   ` [patch 20/45] v4l/dvb: stradis: dont export MODULE_DEVICE_TABLE Greg KH
2006-07-17 16:27   ` [patch 21/45] dvb-bt8xx: fix frontend detection for DViCO FusionHDTV DVB-T Lite rev 1.2 Greg KH
2006-07-17 16:27   ` [patch 22/45] Make powernow-k7 work on SMP kernels Greg KH
2006-07-17 16:27   ` [patch 23/45] Fix powernow-k8 SMP kernel on UP hardware bug Greg KH
2006-07-17 16:27   ` [patch 24/45] cdrom: fix bad cgc.buflen assignment Greg KH
2006-07-17 16:27   ` [patch 25/45] splice: fix problems with sys_tee() Greg KH
2006-07-17 16:28   ` [patch 26/45] USB serial ftdi_sio: Prevent userspace DoS (CVE-2006-2936) Greg KH
2006-07-17 16:28   ` [patch 27/45] tpm: interrupt clear fix Greg KH
2006-07-17 16:56     ` Kylene Jo Hall
2006-07-17 17:52       ` [stable] " Greg KH
2006-07-17 16:28   ` [patch 28/45] pdflush: handle resume wakeups Greg KH
2006-07-17 16:28   ` [patch 29/45] ieee80211: TKIP requires CRC32 Greg KH
2006-07-17 16:28   ` [patch 30/45] : Fix IPv4/DECnet routing rule dumping Greg KH
2006-07-17 16:28   ` [patch 31/45] : Add missing UFO initialisations Greg KH
2006-07-17 16:28   ` [patch 32/45] ALSA: Suppress irq handler mismatch messages in ALSA ISA drivers Greg KH
2006-07-17 16:28   ` [patch 33/45] ALSA: RME HDSP - fixed proc interface (missing {}) Greg KH
2006-07-17 16:28   ` [patch 34/45] ALSA: hda-intel - Fix race in remove Greg KH
2006-07-17 16:28   ` [patch 35/45] ALSA: Fix workaround for AD1988A rev2 codec Greg KH
2006-07-17 16:28   ` [patch 36/45] ALSA: Fix undefined (missing) references in ISA MIRO sound driver Greg KH
2006-07-17 16:28   ` [patch 37/45] ALSA: fix the SND_FM801_TEA575X dependencies Greg KH
2006-07-17 16:29   ` [patch 38/45] ALSA: Fix mute switch on VAIO laptops with STAC7661 Greg KH
2006-07-17 16:29   ` [patch 39/45] ALSA: Fix model for HP dc7600 Greg KH
2006-07-17 16:29   ` [patch 40/45] ALSA: Fix missing array terminators in AD1988 codec support Greg KH
2006-07-17 16:29   ` [patch 41/45] ALSA: Fix a deadlock in snd-rtctimer Greg KH
2006-07-17 16:29   ` [patch 42/45] ALSA: au88x0 - Fix 64bit address of MPU401 MMIO port Greg KH
2006-07-17 16:29   ` [patch 43/45] struct file leakage Greg KH
2006-07-17 16:29   ` [patch 44/45] serial 8250: sysrq deadlock fix Greg KH
2006-07-17 16:29   ` [patch 45/45] fix fdset leakage Greg KH
2006-07-18 13:24   ` [patch 00/45] 2.6.17.y -stable review akrout70

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=20060717162531.GC4829@kroah.com \
    --to=gregkh@suse.de \
    --cc=akpm@osdl.org \
    --cc=alan@lxorguk.ukuu.org.uk \
    --cc=chrisw@sous-sol.org \
    --cc=chuckw@quantumlinux.com \
    --cc=davej@redhat.com \
    --cc=jmforbes@linuxtx.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mst@mellanox.co.il \
    --cc=openib-general@openib.org \
    --cc=rdunlap@xenotime.net \
    --cc=reviews@ml.cw.f00f.org \
    --cc=rolandd@cisco.com \
    --cc=stable@kernel.org \
    --cc=torvalds@osdl.org \
    --cc=tytso@mit.edu \
    --cc=zwane@arm.linux.org.uk \
    /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