From patchwork Mon Nov 19 22:26:02 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Stephen Warren X-Patchwork-Id: 10689595 X-Patchwork-Delegate: bhelgaas@google.com Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id A159E13BF for ; Mon, 19 Nov 2018 22:26:52 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 8E21229D9A for ; Mon, 19 Nov 2018 22:26:52 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 7D3B42A4DB; Mon, 19 Nov 2018 22:26:52 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-7.9 required=2.0 tests=BAYES_00,MAILING_LIST_MULTI, RCVD_IN_DNSWL_HI autolearn=ham version=3.3.1 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 5C6D229D9A for ; Mon, 19 Nov 2018 22:26:51 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1730244AbeKTIwi (ORCPT ); Tue, 20 Nov 2018 03:52:38 -0500 Received: from avon.wwwdotorg.org ([104.237.132.123]:37954 "EHLO avon.wwwdotorg.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726431AbeKTIwi (ORCPT ); Tue, 20 Nov 2018 03:52:38 -0500 Received: from swarren-lx1.nvidia.com (unknown [216.228.112.24]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-SHA256 (128/128 bits)) (No client certificate requested) by avon.wwwdotorg.org (Postfix) with ESMTPSA id 68E061C00E3; Mon, 19 Nov 2018 15:26:47 -0700 (MST) X-Virus-Status: Clean X-Virus-Scanned: clamav-milter 0.100.2 at avon.wwwdotorg.org From: Stephen Warren To: Jingoo Han , Gustavo Pimentel , Lorenzo Pieralisi , Bjorn Helgaas Cc: linux-pci@vger.kernel.org, Vidya Sagar , Manikanta Maddireddy , Stephen Warren Subject: [PATCH] PCI: dwc ep: cache config until DBI regs available Date: Mon, 19 Nov 2018 15:26:02 -0700 Message-Id: <20181119222602.28001-1-swarren@wwwdotorg.org> X-Mailer: git-send-email 2.19.1 MIME-Version: 1.0 X-NVConfidentiality: public Sender: linux-pci-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-pci@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP From: Stephen Warren Some implementations of the DWC PCIe endpoint controller do not allow access to DBI registers until the attached host has started REFCLK, released PERST, and the endpoint driver has initialized clocking of the DBI registers based on that. One such system is NVIDIA's T194 SoC. The PCIe endpoint subsystem and DWC driver currently don't work on such hardware, since they assume that all endpoint configuration can happen at any arbitrary time. Enhance the DWC endpoint driver to support such systems by caching all endpoint configuration in software, and only writing the configuration to hardware once it's been initialized. This is implemented by splitting all endpoint controller ops into two functions; the first which simply records/caches the desired configuration whenever called by the associated function driver and optionally calls the second, and the second which actually programs the configuration into hardware, which may be called either by the first function, or later when it's known that the DBI registers are available. Signed-off-by: Stephen Warren --- Note: I have only compiled-tested this patch in the context of the mainline kernel since NVIDIA Tegra PCIe EP support is not yet present. However, I've built and tested an equivalent patch in our downstream kernel. I did have to manually port it to mainline due to conflicts though. This patch was developed on top of "PCI: designware: don't hard-code DBI/ATU offset" which I sent last week, so may rely on that for patch context. --- .../pci/controller/dwc/pcie-designware-ep.c | 199 ++++++++++++++---- drivers/pci/controller/dwc/pcie-designware.h | 26 ++- 2 files changed, 181 insertions(+), 44 deletions(-) diff --git a/drivers/pci/controller/dwc/pcie-designware-ep.c b/drivers/pci/controller/dwc/pcie-designware-ep.c index 880210366e71..6910b64bfdcb 100644 --- a/drivers/pci/controller/dwc/pcie-designware-ep.c +++ b/drivers/pci/controller/dwc/pcie-designware-ep.c @@ -73,11 +73,10 @@ static u8 dw_pcie_ep_find_capability(struct dw_pcie *pci, u8 cap) return __dw_pcie_ep_find_next_cap(pci, next_cap_ptr, cap); } -static int dw_pcie_ep_write_header(struct pci_epc *epc, u8 func_no, - struct pci_epf_header *hdr) +static void dw_pcie_ep_write_header_regs(struct dw_pcie_ep *ep) { - struct dw_pcie_ep *ep = epc_get_drvdata(epc); struct dw_pcie *pci = to_dw_pcie_from_ep(ep); + struct pci_epf_header *hdr = &(ep->cached_hdr); dw_pcie_dbi_ro_wr_en(pci); dw_pcie_writew_dbi(pci, PCI_VENDOR_ID, hdr->vendorid); @@ -94,6 +93,19 @@ static int dw_pcie_ep_write_header(struct pci_epc *epc, u8 func_no, dw_pcie_writeb_dbi(pci, PCI_INTERRUPT_PIN, hdr->interrupt_pin); dw_pcie_dbi_ro_wr_dis(pci); +} + +static int dw_pcie_ep_write_header(struct pci_epc *epc, u8 func_no, + struct pci_epf_header *hdr) +{ + struct dw_pcie_ep *ep = epc_get_drvdata(epc); + + ep->cached_hdr = *hdr; + + if (!ep->hw_regs_available) + return 0; + + dw_pcie_ep_write_header_regs(ep); return 0; } @@ -112,6 +124,15 @@ static int dw_pcie_ep_inbound_atu(struct dw_pcie_ep *ep, enum pci_barno bar, return -EINVAL; } + ep->cached_inbound_atus[free_win].bar = bar; + ep->cached_inbound_atus[free_win].cpu_addr = cpu_addr; + ep->cached_inbound_atus[free_win].as_type = as_type; + ep->cached_bars[bar].atu_index = free_win; + set_bit(free_win, ep->ib_window_map); + + if (!ep->hw_regs_available) + return 0; + ret = dw_pcie_prog_inbound_atu(pci, free_win, bar, cpu_addr, as_type); if (ret < 0) { @@ -119,9 +140,6 @@ static int dw_pcie_ep_inbound_atu(struct dw_pcie_ep *ep, enum pci_barno bar, return ret; } - ep->bar_to_atu[bar] = free_win; - set_bit(free_win, ep->ib_window_map); - return 0; } @@ -137,27 +155,66 @@ static int dw_pcie_ep_outbound_atu(struct dw_pcie_ep *ep, phys_addr_t phys_addr, return -EINVAL; } + ep->cached_outbound_atus[free_win].addr = phys_addr; + ep->cached_outbound_atus[free_win].pci_addr = pci_addr; + ep->cached_outbound_atus[free_win].size = size; + set_bit(free_win, ep->ob_window_map); + + if (!ep->hw_regs_available) + return 0; + dw_pcie_prog_outbound_atu(pci, free_win, PCIE_ATU_TYPE_MEM, phys_addr, pci_addr, size); - set_bit(free_win, ep->ob_window_map); - ep->outbound_addr[free_win] = phys_addr; - return 0; } +static void dw_pcie_ep_clear_bar_regs(struct dw_pcie_ep *ep, u8 func_no, + enum pci_barno bar) +{ + struct dw_pcie *pci = to_dw_pcie_from_ep(ep); + int flags = ep->cached_bars[bar].flags; + u32 atu_index = ep->cached_bars[bar].atu_index; + + __dw_pcie_ep_reset_bar(pci, bar, flags); + + dw_pcie_disable_atu(pci, atu_index, DW_PCIE_REGION_INBOUND); +} + static void dw_pcie_ep_clear_bar(struct pci_epc *epc, u8 func_no, struct pci_epf_bar *epf_bar) { struct dw_pcie_ep *ep = epc_get_drvdata(epc); - struct dw_pcie *pci = to_dw_pcie_from_ep(ep); enum pci_barno bar = epf_bar->barno; - u32 atu_index = ep->bar_to_atu[bar]; + u32 atu_index = ep->cached_bars[bar].atu_index; - __dw_pcie_ep_reset_bar(pci, bar, epf_bar->flags); - - dw_pcie_disable_atu(pci, atu_index, DW_PCIE_REGION_INBOUND); clear_bit(atu_index, ep->ib_window_map); + + if (!ep->hw_regs_available) + return; + + dw_pcie_ep_clear_bar_regs(ep, func_no, bar); +} + +static void dw_pcie_ep_set_bar_regs(struct dw_pcie_ep *ep, u8 func_no, + enum pci_barno bar) +{ + struct dw_pcie *pci = to_dw_pcie_from_ep(ep); + size_t size = ep->cached_bars[bar].size; + int flags = ep->cached_bars[bar].flags; + u32 reg = PCI_BASE_ADDRESS_0 + (4 * bar); + + dw_pcie_dbi_ro_wr_en(pci); + + dw_pcie_writel_dbi2(pci, reg, lower_32_bits(size - 1)); + dw_pcie_writel_dbi(pci, reg, flags); + + if (flags & PCI_BASE_ADDRESS_MEM_TYPE_64) { + dw_pcie_writel_dbi2(pci, reg + 4, upper_32_bits(size - 1)); + dw_pcie_writel_dbi(pci, reg + 4, 0); + } + + dw_pcie_dbi_ro_wr_dis(pci); } static int dw_pcie_ep_set_bar(struct pci_epc *epc, u8 func_no, @@ -165,12 +222,10 @@ static int dw_pcie_ep_set_bar(struct pci_epc *epc, u8 func_no, { int ret; struct dw_pcie_ep *ep = epc_get_drvdata(epc); - struct dw_pcie *pci = to_dw_pcie_from_ep(ep); enum pci_barno bar = epf_bar->barno; size_t size = epf_bar->size; int flags = epf_bar->flags; enum dw_pcie_as_type as_type; - u32 reg = PCI_BASE_ADDRESS_0 + (4 * bar); if (!(flags & PCI_BASE_ADDRESS_SPACE)) as_type = DW_PCIE_AS_MEM; @@ -181,17 +236,13 @@ static int dw_pcie_ep_set_bar(struct pci_epc *epc, u8 func_no, if (ret) return ret; - dw_pcie_dbi_ro_wr_en(pci); - - dw_pcie_writel_dbi2(pci, reg, lower_32_bits(size - 1)); - dw_pcie_writel_dbi(pci, reg, flags); + ep->cached_bars[bar].size = size; + ep->cached_bars[bar].flags = flags; - if (flags & PCI_BASE_ADDRESS_MEM_TYPE_64) { - dw_pcie_writel_dbi2(pci, reg + 4, upper_32_bits(size - 1)); - dw_pcie_writel_dbi(pci, reg + 4, 0); - } + if (!ep->hw_regs_available) + return 0; - dw_pcie_dbi_ro_wr_dis(pci); + dw_pcie_ep_set_bar_regs(ep, func_no, bar); return 0; } @@ -202,7 +253,7 @@ static int dw_pcie_find_index(struct dw_pcie_ep *ep, phys_addr_t addr, u32 index; for (index = 0; index < ep->num_ob_windows; index++) { - if (ep->outbound_addr[index] != addr) + if (ep->cached_outbound_atus[index].addr != addr) continue; *atu_index = index; return 0; @@ -223,8 +274,12 @@ static void dw_pcie_ep_unmap_addr(struct pci_epc *epc, u8 func_no, if (ret < 0) return; - dw_pcie_disable_atu(pci, atu_index, DW_PCIE_REGION_OUTBOUND); clear_bit(atu_index, ep->ob_window_map); + + if (!ep->hw_regs_available) + return; + + dw_pcie_disable_atu(pci, atu_index, DW_PCIE_REGION_OUTBOUND); } static int dw_pcie_ep_map_addr(struct pci_epc *epc, u8 func_no, @@ -254,7 +309,10 @@ static int dw_pcie_ep_get_msi(struct pci_epc *epc, u8 func_no) return -EINVAL; reg = ep->msi_cap + PCI_MSI_FLAGS; - val = dw_pcie_readw_dbi(pci, reg); + if (ep->hw_regs_available) + val = dw_pcie_readw_dbi(pci, reg); + else + val = ep->cached_msi_flags; if (!(val & PCI_MSI_FLAGS_ENABLE)) return -EINVAL; @@ -273,12 +331,19 @@ static int dw_pcie_ep_set_msi(struct pci_epc *epc, u8 func_no, u8 interrupts) return -EINVAL; reg = ep->msi_cap + PCI_MSI_FLAGS; - val = dw_pcie_readw_dbi(pci, reg); + if (ep->hw_regs_available) + val = dw_pcie_readw_dbi(pci, reg); + else + val = ep->cached_msi_flags; val &= ~PCI_MSI_FLAGS_QMASK; val |= (interrupts << 1) & PCI_MSI_FLAGS_QMASK; - dw_pcie_dbi_ro_wr_en(pci); - dw_pcie_writew_dbi(pci, reg, val); - dw_pcie_dbi_ro_wr_dis(pci); + if (ep->hw_regs_available) { + dw_pcie_dbi_ro_wr_en(pci); + dw_pcie_writew_dbi(pci, reg, val); + dw_pcie_dbi_ro_wr_dis(pci); + } else { + ep->cached_msi_flags = val; + } return 0; } @@ -293,7 +358,10 @@ static int dw_pcie_ep_get_msix(struct pci_epc *epc, u8 func_no) return -EINVAL; reg = ep->msix_cap + PCI_MSIX_FLAGS; - val = dw_pcie_readw_dbi(pci, reg); + if (ep->hw_regs_available) + val = dw_pcie_readw_dbi(pci, reg); + else + val = ep->cached_msix_flags; if (!(val & PCI_MSIX_FLAGS_ENABLE)) return -EINVAL; @@ -312,16 +380,49 @@ static int dw_pcie_ep_set_msix(struct pci_epc *epc, u8 func_no, u16 interrupts) return -EINVAL; reg = ep->msix_cap + PCI_MSIX_FLAGS; - val = dw_pcie_readw_dbi(pci, reg); + if (ep->hw_regs_available) + val = dw_pcie_readw_dbi(pci, reg); + else + val = ep->cached_msix_flags; val &= ~PCI_MSIX_FLAGS_QSIZE; val |= interrupts; - dw_pcie_dbi_ro_wr_en(pci); - dw_pcie_writew_dbi(pci, reg, val); - dw_pcie_dbi_ro_wr_dis(pci); + if (ep->hw_regs_available) { + dw_pcie_dbi_ro_wr_en(pci); + dw_pcie_writew_dbi(pci, reg, val); + dw_pcie_dbi_ro_wr_dis(pci); + } else { + ep->cached_msix_flags = val; + } return 0; } +void dw_pcie_set_regs_available(struct dw_pcie *pci) +{ + struct dw_pcie_ep *ep = &(pci->ep); + int i; + + ep->hw_regs_available = true; + + dw_pcie_ep_write_header_regs(ep); + for_each_set_bit(i, ep->ib_window_map, ep->num_ib_windows) { + dw_pcie_prog_inbound_atu(pci, i, + ep->cached_inbound_atus[i].bar, + ep->cached_inbound_atus[i].cpu_addr, + ep->cached_inbound_atus[i].as_type); + dw_pcie_ep_set_bar_regs(ep, 0, ep->cached_inbound_atus[i].bar); + } + for_each_set_bit(i, ep->ob_window_map, ep->num_ob_windows) + dw_pcie_prog_outbound_atu(pci, i, PCIE_ATU_TYPE_MEM, + ep->cached_outbound_atus[i].addr, + ep->cached_outbound_atus[i].pci_addr, + ep->cached_outbound_atus[i].size); + dw_pcie_dbi_ro_wr_en(pci); + dw_pcie_writew_dbi(pci, PCI_MSI_FLAGS, ep->cached_msi_flags); + dw_pcie_writew_dbi(pci, PCI_MSIX_FLAGS, ep->cached_msix_flags); + dw_pcie_dbi_ro_wr_dis(pci); +} + static int dw_pcie_ep_raise_irq(struct pci_epc *epc, u8 func_no, enum pci_epc_irq_type type, u16 interrupt_num) { @@ -330,6 +431,9 @@ static int dw_pcie_ep_raise_irq(struct pci_epc *epc, u8 func_no, if (!ep->ops->raise_irq) return -EINVAL; + if (!ep->hw_regs_available) + return -EAGAIN; + return ep->ops->raise_irq(ep, func_no, type, interrupt_num); } @@ -391,6 +495,9 @@ int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no, bool has_upper; int ret; + if (!ep->hw_regs_available) + return -EAGAIN; + if (!ep->msi_cap) return -EINVAL; @@ -436,6 +543,9 @@ int dw_pcie_ep_raise_msix_irq(struct dw_pcie_ep *ep, u8 func_no, void __iomem *msix_tbl; int ret; + if (!ep->hw_regs_available) + return -EAGAIN; + reg = ep->msix_cap + PCI_MSIX_TABLE; tbl_offset = dw_pcie_readl_dbi(pci, reg); bir = (tbl_offset & PCI_MSIX_TABLE_BIR); @@ -494,7 +604,6 @@ void dw_pcie_ep_exit(struct dw_pcie_ep *ep) int dw_pcie_ep_init(struct dw_pcie_ep *ep) { int ret; - void *addr; struct pci_epc *epc; struct dw_pcie *pci = to_dw_pcie_from_ep(ep); struct device *dev = pci->dev; @@ -543,11 +652,17 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep) if (!ep->ob_window_map) return -ENOMEM; - addr = devm_kcalloc(dev, ep->num_ob_windows, sizeof(phys_addr_t), - GFP_KERNEL); - if (!addr) + ep->cached_inbound_atus = devm_kzalloc(dev, + sizeof(ep->cached_inbound_atus[0]) * ep->num_ib_windows, + GFP_KERNEL); + if (!ep->cached_inbound_atus) + return -ENOMEM; + + ep->cached_outbound_atus = devm_kzalloc(dev, + sizeof(ep->cached_outbound_atus[0]) * ep->num_ob_windows, + GFP_KERNEL); + if (!ep->cached_outbound_atus) return -ENOMEM; - ep->outbound_addr = addr; epc = devm_pci_epc_create(dev, &epc_ops); if (IS_ERR(epc)) { diff --git a/drivers/pci/controller/dwc/pcie-designware.h b/drivers/pci/controller/dwc/pcie-designware.h index 02fb532c5db9..c116b289aa51 100644 --- a/drivers/pci/controller/dwc/pcie-designware.h +++ b/drivers/pci/controller/dwc/pcie-designware.h @@ -192,8 +192,6 @@ struct dw_pcie_ep { phys_addr_t phys_base; size_t addr_size; size_t page_size; - u8 bar_to_atu[6]; - phys_addr_t *outbound_addr; unsigned long *ib_window_map; unsigned long *ob_window_map; u32 num_ib_windows; @@ -202,6 +200,25 @@ struct dw_pcie_ep { phys_addr_t msi_mem_phys; u8 msi_cap; /* MSI capability offset */ u8 msix_cap; /* MSI-X capability offset */ + bool hw_regs_available; + struct pci_epf_header cached_hdr; + struct { + size_t size; + int flags; + u8 atu_index; + } cached_bars[6]; + struct { + enum pci_barno bar; + dma_addr_t cpu_addr; + enum dw_pcie_as_type as_type; + } *cached_inbound_atus; + struct { + phys_addr_t addr; + u64 pci_addr; + size_t size; + } *cached_outbound_atus; + u32 cached_msi_flags; + u32 cached_msix_flags; }; struct dw_pcie_ops { @@ -363,6 +380,7 @@ static inline int dw_pcie_allocate_domains(struct pcie_port *pp) void dw_pcie_ep_linkup(struct dw_pcie_ep *ep); int dw_pcie_ep_init(struct dw_pcie_ep *ep); void dw_pcie_ep_exit(struct dw_pcie_ep *ep); +void dw_pcie_ep_set_regs_available(struct dw_pcie *pci); int dw_pcie_ep_raise_legacy_irq(struct dw_pcie_ep *ep, u8 func_no); int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no, u8 interrupt_num); @@ -383,6 +401,10 @@ static inline void dw_pcie_ep_exit(struct dw_pcie_ep *ep) { } +static inline void dw_pcie_ep_set_regs_available(struct dw_pcie *pci) +{ +} + static inline int dw_pcie_ep_raise_legacy_irq(struct dw_pcie_ep *ep, u8 func_no) { return 0;