From 083874549fdfefa629dfa752785e20427dde1511 Mon Sep 17 00:00:00 2001 From: Daniel Drake Date: Thu, 27 Sep 2018 15:47:33 -0500 Subject: PCI: Reprogram bridge prefetch registers on resume On 38+ Intel-based ASUS products, the NVIDIA GPU becomes unusable after S3 suspend/resume. The affected products include multiple generations of NVIDIA GPUs and Intel SoCs. After resume, nouveau logs many errors such as: fifo: fault 00 [READ] at 0000005555555000 engine 00 [GR] client 04 [HUB/FE] reason 4a [] on channel -1 [007fa91000 unknown] DRM: failed to idle channel 0 [DRM] Similarly, the NVIDIA proprietary driver also fails after resume (black screen, 100% CPU usage in Xorg process). We shipped a sample to NVIDIA for diagnosis, and their response indicated that it's a problem with the parent PCI bridge (on the Intel SoC), not the GPU. Runtime suspend/resume works fine, only S3 suspend is affected. We found a workaround: on resume, rewrite the Intel PCI bridge 'Prefetchable Base Upper 32 Bits' register (PCI_PREF_BASE_UPPER32). In the cases that I checked, this register has value 0 and we just have to rewrite that value. Linux already saves and restores PCI config space during suspend/resume, but this register was being skipped because upon resume, it already has value 0 (the correct, pre-suspend value). Intel appear to have previously acknowledged this behaviour and the requirement to rewrite this register: https://bugzilla.kernel.org/show_bug.cgi?id=116851#c23 Based on that, rewrite the prefetch register values even when that appears unnecessary. We have confirmed this solution on all the affected models we have in-hands (X542UQ, UX533FD, X530UN, V272UN). Additionally, this solves an issue where r8169 MSI-X interrupts were broken after S3 suspend/resume on ASUS X441UAR. This issue was recently worked around in commit 7bb05b85bc2d ("r8169: don't use MSI-X on RTL8106e"). It also fixes the same issue on RTL6186evl/8111evl on an Aimfor-tech laptop that we had not yet patched. I suspect it will also fix the issue that was worked around in commit 7c53a722459c ("r8169: don't use MSI-X on RTL8168g"). Thomas Martitz reports that this change also solves an issue where the AMD Radeon Polaris 10 GPU on the HP Zbook 14u G5 is unresponsive after S3 suspend/resume. Link: https://bugzilla.kernel.org/show_bug.cgi?id=201069 Signed-off-by: Daniel Drake Signed-off-by: Bjorn Helgaas Reviewed-by: Rafael J. Wysocki Reviewed-By: Peter Wu CC: stable@vger.kernel.org --- drivers/pci/pci.c | 27 +++++++++++++++++++-------- 1 file changed, 19 insertions(+), 8 deletions(-) (limited to 'drivers/pci') diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index 1835f3a7aa8d..51b6c81671c1 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -1289,12 +1289,12 @@ int pci_save_state(struct pci_dev *dev) EXPORT_SYMBOL(pci_save_state); static void pci_restore_config_dword(struct pci_dev *pdev, int offset, - u32 saved_val, int retry) + u32 saved_val, int retry, bool force) { u32 val; pci_read_config_dword(pdev, offset, &val); - if (val == saved_val) + if (!force && val == saved_val) return; for (;;) { @@ -1313,25 +1313,36 @@ static void pci_restore_config_dword(struct pci_dev *pdev, int offset, } static void pci_restore_config_space_range(struct pci_dev *pdev, - int start, int end, int retry) + int start, int end, int retry, + bool force) { int index; for (index = end; index >= start; index--) pci_restore_config_dword(pdev, 4 * index, pdev->saved_config_space[index], - retry); + retry, force); } static void pci_restore_config_space(struct pci_dev *pdev) { if (pdev->hdr_type == PCI_HEADER_TYPE_NORMAL) { - pci_restore_config_space_range(pdev, 10, 15, 0); + pci_restore_config_space_range(pdev, 10, 15, 0, false); /* Restore BARs before the command register. */ - pci_restore_config_space_range(pdev, 4, 9, 10); - pci_restore_config_space_range(pdev, 0, 3, 0); + pci_restore_config_space_range(pdev, 4, 9, 10, false); + pci_restore_config_space_range(pdev, 0, 3, 0, false); + } else if (pdev->hdr_type == PCI_HEADER_TYPE_BRIDGE) { + pci_restore_config_space_range(pdev, 12, 15, 0, false); + + /* + * Force rewriting of prefetch registers to avoid S3 resume + * issues on Intel PCI bridges that occur when these + * registers are not explicitly written. + */ + pci_restore_config_space_range(pdev, 9, 11, 0, true); + pci_restore_config_space_range(pdev, 0, 8, 0, false); } else { - pci_restore_config_space_range(pdev, 0, 15, 0); + pci_restore_config_space_range(pdev, 0, 15, 0, false); } } -- cgit v1.2.3 From 95375f2ab2960c135484d83ea9f8f357cb1be26a Mon Sep 17 00:00:00 2001 From: Thomas Petazzoni Date: Mon, 1 Oct 2018 14:49:34 +0200 Subject: PCI: mvebu: Fix PCI I/O mapping creation sequence MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Commit ee1604381a371 ("PCI: mvebu: Only remap I/O space if configured") had the side effect that the PCI I/O mapping was created much earlier than before, at a point where the probe() of the driver could still fail. This is for example a problem if one gets an -EPROBE_DEFER at some point during probe(), after pci_ioremap_io() has been called. Indeed, there is currently no function to undo what pci_ioremap_io() did, and switching to pci_remap_iospace() is not an option in pci-mvebu due to the need for special memory attributes on Armada 38x. Reverting ee1604381a371 ("PCI: mvebu: Only remap I/O space if configured") would be a possibility, but it would require also reverting 42342073e38b5 ("PCI: mvebu: Convert to use pci_host_bridge directly"). So instead, we use an open-coded version of pci_host_probe() that creates the PCI I/O mapping at a point where we are guaranteed not to fail anymore. Fixes: ee1604381a371 ("PCI: mvebu: Only remap I/O space if configured") Reported-by: Jan Kundrát Tested-by: Jan Kundrát Signed-off-by: Thomas Petazzoni Signed-off-by: Bjorn Helgaas Acked-by: Lorenzo Pieralisi --- drivers/pci/controller/pci-mvebu.c | 52 +++++++++++++++++++++++++++++++++++--- 1 file changed, 48 insertions(+), 4 deletions(-) (limited to 'drivers/pci') diff --git a/drivers/pci/controller/pci-mvebu.c b/drivers/pci/controller/pci-mvebu.c index 50eb0729385b..a41d79b8d46a 100644 --- a/drivers/pci/controller/pci-mvebu.c +++ b/drivers/pci/controller/pci-mvebu.c @@ -1145,7 +1145,6 @@ static int mvebu_pcie_parse_request_resources(struct mvebu_pcie *pcie) { struct device *dev = &pcie->pdev->dev; struct device_node *np = dev->of_node; - unsigned int i; int ret; INIT_LIST_HEAD(&pcie->resources); @@ -1179,13 +1178,58 @@ static int mvebu_pcie_parse_request_resources(struct mvebu_pcie *pcie) resource_size(&pcie->io) - 1); pcie->realio.name = "PCI I/O"; + pci_add_resource(&pcie->resources, &pcie->realio); + } + + return devm_request_pci_bus_resources(dev, &pcie->resources); +} + +/* + * This is a copy of pci_host_probe(), except that it does the I/O + * remap as the last step, once we are sure we won't fail. + * + * It should be removed once the I/O remap error handling issue has + * been sorted out. + */ +static int mvebu_pci_host_probe(struct pci_host_bridge *bridge) +{ + struct mvebu_pcie *pcie; + struct pci_bus *bus, *child; + int ret; + + ret = pci_scan_root_bus_bridge(bridge); + if (ret < 0) { + dev_err(bridge->dev.parent, "Scanning root bridge failed"); + return ret; + } + + pcie = pci_host_bridge_priv(bridge); + if (resource_size(&pcie->io) != 0) { + unsigned int i; + for (i = 0; i < resource_size(&pcie->realio); i += SZ_64K) pci_ioremap_io(i, pcie->io.start + i); + } - pci_add_resource(&pcie->resources, &pcie->realio); + bus = bridge->bus; + + /* + * We insert PCI resources into the iomem_resource and + * ioport_resource trees in either pci_bus_claim_resources() + * or pci_bus_assign_resources(). + */ + if (pci_has_flag(PCI_PROBE_ONLY)) { + pci_bus_claim_resources(bus); + } else { + pci_bus_size_bridges(bus); + pci_bus_assign_resources(bus); + + list_for_each_entry(child, &bus->children, node) + pcie_bus_configure_settings(child); } - return devm_request_pci_bus_resources(dev, &pcie->resources); + pci_bus_add_devices(bus); + return 0; } static int mvebu_pcie_probe(struct platform_device *pdev) @@ -1268,7 +1312,7 @@ static int mvebu_pcie_probe(struct platform_device *pdev) bridge->align_resource = mvebu_pcie_align_resource; bridge->msi = pcie->msi; - return pci_host_probe(bridge); + return mvebu_pci_host_probe(bridge); } static const struct of_device_id mvebu_pcie_of_match_table[] = { -- cgit v1.2.3