netdev - Experimental fix for MSI-X issue on r8169

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-next>] [day] [month] [year] [list]

Message-ID: <9f605d51-c1b3-2f00-9d44-0f53dcb38848@gmail.com>
Date:   Sun, 19 Aug 2018 22:34:59 +0200
From:   Heiner Kallweit <hkallweit1@...il.com>
To:     Steve Dodd <steved424@...il.com>, Lou Reed <gogen@...root.org>,
        Jian-Hong Pan <jian-hong@...lessm.com>
Cc:     "netdev@...r.kernel.org" <netdev@...r.kernel.org>
Subject: Experimental fix for MSI-X issue on r8169

The three of you reported an MSI-X-related error when the system
resumes from suspend. This has been fixed for now by disabling MSI-X
on certain chip versions. However more versions may be affected.

I checked with Realtek and they confirmed that on certain chip
versions a MSIX-related value in PCI config space is reset when
resuming from S3.

I would appreciate if you could test the following experimental patch
and whether warning "MSIX address lost, re-configuring" appears in
your dmesg output after resume from suspend.

Thanks a lot for your efforts.

Heiner

---
 drivers/net/ethernet/realtek/r8169.c | 27 ++++++++++++++++++++++-----
 1 file changed, 22 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/realtek/r8169.c b/drivers/net/ethernet/realtek/r8169.c
index 0d9c38318..56b4bdff9 100644
--- a/drivers/net/ethernet/realtek/r8169.c
+++ b/drivers/net/ethernet/realtek/r8169.c
@@ -690,6 +690,8 @@ struct rtl8169_private {
 	struct rtl8169_counters *counters;
 	struct rtl8169_tc_offsets tc_offset;
 	u32 saved_wolopts;
+	u32 saved_msix_addr_lo;
+	u32 saved_msix_addr_hi;
 
 	struct rtl_fw {
 		const struct firmware *fw;
@@ -6876,6 +6878,19 @@ static int rtl8169_resume(struct device *device)
 {
 	struct pci_dev *pdev = to_pci_dev(device);
 	struct net_device *dev = pci_get_drvdata(pdev);
+	struct rtl8169_private *tp = netdev_priv(dev);
+	u32 val;
+
+	/* Some chip versions loose these values when resuming */
+	if (pdev->msix_enabled) {
+		pci_read_config_dword(pdev, PCI_BASE_ADDRESS_4, &val);
+		if (!val)
+			dev_warn(device, "MSIX address lost, re-configuring\n");
+		pci_write_config_dword(pdev, PCI_BASE_ADDRESS_4,
+				       tp->saved_msix_addr_lo);
+		pci_write_config_dword(pdev, PCI_BASE_ADDRESS_5,
+				       tp->saved_msix_addr_hi);
+	}
 
 	if (netif_running(dev))
 		__rtl8169_resume(dev);
@@ -7076,11 +7091,6 @@ static int rtl_alloc_irq(struct rtl8169_private *tp)
 		RTL_W8(tp, Config2, RTL_R8(tp, Config2) & ~MSIEnable);
 		RTL_W8(tp, Cfg9346, Cfg9346_Lock);
 		flags = PCI_IRQ_LEGACY;
-	} else if (tp->mac_version == RTL_GIGA_MAC_VER_40) {
-		/* This version was reported to have issues with resume
-		 * from suspend when using MSI-X
-		 */
-		flags = PCI_IRQ_LEGACY | PCI_IRQ_MSI;
 	} else {
 		flags = PCI_IRQ_ALL_TYPES;
 	}
@@ -7355,6 +7365,13 @@ static int rtl_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 		return rc;
 	}
 
+	if (pdev->msix_enabled) {
+		pci_read_config_dword(pdev, PCI_BASE_ADDRESS_4,
+				      &tp->saved_msix_addr_lo);
+		pci_read_config_dword(pdev, PCI_BASE_ADDRESS_5,
+				      &tp->saved_msix_addr_hi);
+	}
+
 	tp->saved_wolopts = __rtl8169_get_wol(tp);
 
 	mutex_init(&tp->wk.mutex);
-- 
2.18.0