| Message ID | 20090826232056.20840.86116.stgit@bob.kio (mailing list archive) |
|---|---|
| State | New, archived |
| Delegated to: | Bjorn Helgaas |
| Headers | show |
Bjorn Helgaas wrote: > This patch adds a new pci_configure_slot() function that programs the > PCI bus characteristics for a newly-added device. This is based on > code in pciehp_pci.c, but this should be generic enough to be used by > pciehp, shpchp, and acpiphp. > diff --git a/include/linux/pci_hotplug.h b/include/linux/pci_hotplug.h > index 0f3e1db..8cff1d5 100644 > --- a/include/linux/pci_hotplug.h > +++ b/include/linux/pci_hotplug.h > @@ -230,6 +230,10 @@ int acpi_get_hp_params(struct pci_dev *dev, struct > hotplug_params *hpp); int acpi_get_hp_hw_control_from_firmware(struct > pci_dev *dev, u32 flags); int acpi_pci_check_ejectable(struct pci_bus > *pbus, acpi_handle handle); int acpi_pci_detect_ejectable(struct pci_bus > *pbus); > +#else > +#define acpi_get_hp_params(dev, hpp) (-ENODEV) Shouldn't that be "static inline acpi_get_hp_params(...) { return -ENODEV; }" so it behaves exactly like the original function (e.g. function pointer can be taken)? > #endif > + > +void pci_configure_slot(struct pci_dev *dev); > #endif Eike
On Wednesday 26 August 2009 11:00:28 pm Rolf Eike Beer wrote: > Bjorn Helgaas wrote: > > This patch adds a new pci_configure_slot() function that programs the > > PCI bus characteristics for a newly-added device. This is based on > > code in pciehp_pci.c, but this should be generic enough to be used by > > pciehp, shpchp, and acpiphp. > > > diff --git a/include/linux/pci_hotplug.h b/include/linux/pci_hotplug.h > > index 0f3e1db..8cff1d5 100644 > > --- a/include/linux/pci_hotplug.h > > +++ b/include/linux/pci_hotplug.h > > @@ -230,6 +230,10 @@ int acpi_get_hp_params(struct pci_dev *dev, struct > > hotplug_params *hpp); int acpi_get_hp_hw_control_from_firmware(struct > > pci_dev *dev, u32 flags); int acpi_pci_check_ejectable(struct pci_bus > > *pbus, acpi_handle handle); int acpi_pci_detect_ejectable(struct pci_bus > > *pbus); > > +#else > > +#define acpi_get_hp_params(dev, hpp) (-ENODEV) > > Shouldn't that be "static inline acpi_get_hp_params(...) { return -ENODEV; }" > so it behaves exactly like the original function (e.g. function pointer can be > taken)? Thanks, Eike, that's a good point. I copied the style from the original pciehp_get_hp_params_from_firmware() definitions (which my next patch removes), but it would be an improvement to use an inline function. It also occurred to me last night that pci_configure_slot() will probably need to be exported, and that there's no need to check "dev->is_pcie" before calling program_hpp_type2(), because we bail out of program_hpp_type2() anyway if there's no PCI_CAP_ID_EXP. I'll tidy up these bits and address any other issues, then post a new series next week. Bjorn > > #endif > > + > > +void pci_configure_slot(struct pci_dev *dev); > > #endif > > Eike > -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Thu, Aug 27, 2009 at 09:00:38AM -0600, Bjorn Helgaas wrote: > On Wednesday 26 August 2009 11:00:28 pm Rolf Eike Beer wrote: > > Bjorn Helgaas wrote: > > > This patch adds a new pci_configure_slot() function that programs the > > > PCI bus characteristics for a newly-added device. This is based on > > > code in pciehp_pci.c, but this should be generic enough to be used by > > > pciehp, shpchp, and acpiphp. > > > > > diff --git a/include/linux/pci_hotplug.h b/include/linux/pci_hotplug.h > > > index 0f3e1db..8cff1d5 100644 > > > --- a/include/linux/pci_hotplug.h > > > +++ b/include/linux/pci_hotplug.h > > > @@ -230,6 +230,10 @@ int acpi_get_hp_params(struct pci_dev *dev, struct > > > hotplug_params *hpp); int acpi_get_hp_hw_control_from_firmware(struct > > > pci_dev *dev, u32 flags); int acpi_pci_check_ejectable(struct pci_bus > > > *pbus, acpi_handle handle); int acpi_pci_detect_ejectable(struct pci_bus > > > *pbus); > > > +#else > > > +#define acpi_get_hp_params(dev, hpp) (-ENODEV) > > > > Shouldn't that be "static inline acpi_get_hp_params(...) { return -ENODEV; }" > > so it behaves exactly like the original function (e.g. function pointer can be > > taken)? > > Thanks, Eike, that's a good point. I copied the style from the > original pciehp_get_hp_params_from_firmware() definitions (which > my next patch removes), but it would be an improvement to use an > inline function. > > It also occurred to me last night that pci_configure_slot() will > probably need to be exported, and that there's no need to check > "dev->is_pcie" before calling program_hpp_type2(), because we > bail out of program_hpp_type2() anyway if there's no PCI_CAP_ID_EXP. Bjorn, I didn't see any 'acpiphp' functional issues when I just tried 2.6.31-rc7 with your patches (plus the definitely needed pci_configure_slot export) on one of our systems (IBM x3850). Gary
On Thursday 27 August 2009 02:08:22 pm Gary Hade wrote: > On Thu, Aug 27, 2009 at 09:00:38AM -0600, Bjorn Helgaas wrote: > > On Wednesday 26 August 2009 11:00:28 pm Rolf Eike Beer wrote: > > > Bjorn Helgaas wrote: > > > > This patch adds a new pci_configure_slot() function that programs the > > > > PCI bus characteristics for a newly-added device. This is based on > > > > code in pciehp_pci.c, but this should be generic enough to be used by > > > > pciehp, shpchp, and acpiphp. > > > > > > > diff --git a/include/linux/pci_hotplug.h b/include/linux/pci_hotplug.h > > > > index 0f3e1db..8cff1d5 100644 > > > > --- a/include/linux/pci_hotplug.h > > > > +++ b/include/linux/pci_hotplug.h > > > > @@ -230,6 +230,10 @@ int acpi_get_hp_params(struct pci_dev *dev, struct > > > > hotplug_params *hpp); int acpi_get_hp_hw_control_from_firmware(struct > > > > pci_dev *dev, u32 flags); int acpi_pci_check_ejectable(struct pci_bus > > > > *pbus, acpi_handle handle); int acpi_pci_detect_ejectable(struct pci_bus > > > > *pbus); > > > > +#else > > > > +#define acpi_get_hp_params(dev, hpp) (-ENODEV) > > > > > > Shouldn't that be "static inline acpi_get_hp_params(...) { return -ENODEV; }" > > > so it behaves exactly like the original function (e.g. function pointer can be > > > taken)? > > > > Thanks, Eike, that's a good point. I copied the style from the > > original pciehp_get_hp_params_from_firmware() definitions (which > > my next patch removes), but it would be an improvement to use an > > inline function. > > > > It also occurred to me last night that pci_configure_slot() will > > probably need to be exported, and that there's no need to check > > "dev->is_pcie" before calling program_hpp_type2(), because we > > bail out of program_hpp_type2() anyway if there's no PCI_CAP_ID_EXP. > > Bjorn, I didn't see any 'acpiphp' functional issues when I just > tried 2.6.31-rc7 with your patches (plus the definitely needed > pci_configure_slot export) on one of our systems (IBM x3850). Oh, great! Thanks a lot for actually trying it out! I've tried acpiphp on an HP rx3600, but haven't been able to try out pciehp or shpchp yet. Bjorn -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Bjorn Helgaas wrote: > On Thursday 27 August 2009 02:08:22 pm Gary Hade wrote: >> On Thu, Aug 27, 2009 at 09:00:38AM -0600, Bjorn Helgaas wrote: >>> On Wednesday 26 August 2009 11:00:28 pm Rolf Eike Beer wrote: >>>> Bjorn Helgaas wrote: >>>>> This patch adds a new pci_configure_slot() function that programs the >>>>> PCI bus characteristics for a newly-added device. This is based on >>>>> code in pciehp_pci.c, but this should be generic enough to be used by >>>>> pciehp, shpchp, and acpiphp. >>>>> diff --git a/include/linux/pci_hotplug.h b/include/linux/pci_hotplug.h >>>>> index 0f3e1db..8cff1d5 100644 >>>>> --- a/include/linux/pci_hotplug.h >>>>> +++ b/include/linux/pci_hotplug.h >>>>> @@ -230,6 +230,10 @@ int acpi_get_hp_params(struct pci_dev *dev, struct >>>>> hotplug_params *hpp); int acpi_get_hp_hw_control_from_firmware(struct >>>>> pci_dev *dev, u32 flags); int acpi_pci_check_ejectable(struct pci_bus >>>>> *pbus, acpi_handle handle); int acpi_pci_detect_ejectable(struct pci_bus >>>>> *pbus); >>>>> +#else >>>>> +#define acpi_get_hp_params(dev, hpp) (-ENODEV) >>>> Shouldn't that be "static inline acpi_get_hp_params(...) { return -ENODEV; }" >>>> so it behaves exactly like the original function (e.g. function pointer can be >>>> taken)? >>> Thanks, Eike, that's a good point. I copied the style from the >>> original pciehp_get_hp_params_from_firmware() definitions (which >>> my next patch removes), but it would be an improvement to use an >>> inline function. >>> >>> It also occurred to me last night that pci_configure_slot() will >>> probably need to be exported, and that there's no need to check >>> "dev->is_pcie" before calling program_hpp_type2(), because we >>> bail out of program_hpp_type2() anyway if there's no PCI_CAP_ID_EXP. >> Bjorn, I didn't see any 'acpiphp' functional issues when I just >> tried 2.6.31-rc7 with your patches (plus the definitely needed >> pci_configure_slot export) on one of our systems (IBM x3850). > > Oh, great! Thanks a lot for actually trying it out! I've tried > acpiphp on an HP rx3600, but haven't been able to try out pciehp > or shpchp yet. I'm going to try pciehp and shpchp. But please wait a moment. Now I'm waiting for the machine time. I'll be able to try them in early in the next week. Thanks, Kenji Kaneshige -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Bjorn Helgaas wrote: > This patch adds a new pci_configure_slot() function that programs the > PCI bus characteristics for a newly-added device. This is based on > code in pciehp_pci.c, but this should be generic enough to be used by > pciehp, shpchp, and acpiphp. > > Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com> > Reviewed-by: Alex Chiang <achiang@hp.com> > --- > drivers/pci/hotplug/Makefile | 2 > drivers/pci/hotplug/pcihp_slot.c | 181 ++++++++++++++++++++++++++++++++++++++ > include/linux/pci_hotplug.h | 4 + > 3 files changed, 186 insertions(+), 1 deletions(-) > create mode 100644 drivers/pci/hotplug/pcihp_slot.c > > diff --git a/drivers/pci/hotplug/Makefile b/drivers/pci/hotplug/Makefile > index 2aa117c..3625b09 100644 > --- a/drivers/pci/hotplug/Makefile > +++ b/drivers/pci/hotplug/Makefile > @@ -22,7 +22,7 @@ obj-$(CONFIG_HOTPLUG_PCI_SGI) += sgi_hotplug.o > # Link this last so it doesn't claim devices that have a real hotplug driver > obj-$(CONFIG_HOTPLUG_PCI_FAKE) += fakephp.o > > -pci_hotplug-objs := pci_hotplug_core.o > +pci_hotplug-objs := pci_hotplug_core.o pcihp_slot.o > > ifdef CONFIG_HOTPLUG_PCI_CPCI > pci_hotplug-objs += cpci_hotplug_core.o \ > diff --git a/drivers/pci/hotplug/pcihp_slot.c b/drivers/pci/hotplug/pcihp_slot.c > new file mode 100644 > index 0000000..82d4698 > --- /dev/null > +++ b/drivers/pci/hotplug/pcihp_slot.c > @@ -0,0 +1,181 @@ > +/* > + * Copyright (C) 1995,2001 Compaq Computer Corporation > + * Copyright (C) 2001 Greg Kroah-Hartman (greg@kroah.com) > + * Copyright (C) 2001 IBM Corp. > + * Copyright (C) 2003-2004 Intel Corporation > + * (c) Copyright 2009 Hewlett-Packard Development Company, L.P. > + * > + * All rights reserved. > + * > + * This program is free software; you can redistribute it and/or modify > + * it under the terms of the GNU General Public License as published by > + * the Free Software Foundation; either version 2 of the License, or (at > + * your option) any later version. > + * > + * This program is distributed in the hope that it will be useful, but > + * WITHOUT ANY WARRANTY; without even the implied warranty of > + * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or > + * NON INFRINGEMENT. See the GNU General Public License for more > + * details. > + * > + * You should have received a copy of the GNU General Public License > + * along with this program; if not, write to the Free Software > + * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. > + */ > + > +#include <linux/pci.h> > +#include <linux/pci_hotplug.h> > + > +static struct hpp_type0 pci_default_type0 = { > + .revision = 1, > + .cache_line_size = 8, > + .latency_timer = 0x40, > + .enable_serr = 0, > + .enable_perr = 0, > +}; > + > +static void program_hpp_type0(struct pci_dev *dev, struct hpp_type0 *hpp) > +{ > + u16 pci_cmd, pci_bctl; > + > + if (hpp->revision > 1) { > + dev_warn(&dev->dev, > + "PCI setting record rev %d not supported\n", > + hpp->revision); > + return; > + } > + > + pci_write_config_byte(dev, PCI_CACHE_LINE_SIZE, hpp->cache_line_size); > + pci_write_config_byte(dev, PCI_LATENCY_TIMER, hpp->latency_timer); > + pci_read_config_word(dev, PCI_COMMAND, &pci_cmd); > + if (hpp->enable_serr) > + pci_cmd |= PCI_COMMAND_SERR; > + else > + pci_cmd &= ~PCI_COMMAND_SERR; > + if (hpp->enable_perr) > + pci_cmd |= PCI_COMMAND_PARITY; > + else > + pci_cmd &= ~PCI_COMMAND_PARITY; > + pci_write_config_word(dev, PCI_COMMAND, pci_cmd); > + > + /* Program bridge control value */ > + if ((dev->class >> 8) == PCI_CLASS_BRIDGE_PCI) { > + pci_write_config_byte(dev, PCI_SEC_LATENCY_TIMER, > + hpp->latency_timer); > + pci_read_config_word(dev, PCI_BRIDGE_CONTROL, &pci_bctl); > + if (hpp->enable_serr) > + pci_bctl |= PCI_BRIDGE_CTL_SERR; > + else > + pci_bctl &= ~PCI_BRIDGE_CTL_SERR; > + if (hpp->enable_perr) > + pci_bctl |= PCI_BRIDGE_CTL_PARITY; > + else > + pci_bctl &= ~PCI_BRIDGE_CTL_PARITY; > + pci_write_config_word(dev, PCI_BRIDGE_CONTROL, pci_bctl); > + } > +} > + > +static void program_hpp_type2(struct pci_dev *dev, struct hpp_type2 *hpp) > +{ > + int pos; > + u16 reg16; > + u32 reg32; > + > + if (hpp->revision > 1) { > + dev_warn(&dev->dev, > + "PCI Express setting record rev %d not supported\n", > + hpp->revision); > + return; > + } > + > + /* Find PCI Express capability */ > + pos = pci_find_capability(dev, PCI_CAP_ID_EXP); > + if (!pos) > + return; > + > + /* Initialize Device Control Register */ > + pci_read_config_word(dev, pos + PCI_EXP_DEVCTL, ®16); > + reg16 = (reg16 & hpp->pci_exp_devctl_and) | hpp->pci_exp_devctl_or; > + pci_write_config_word(dev, pos + PCI_EXP_DEVCTL, reg16); > + > + /* Initialize Link Control Register */ > + if (dev->subordinate) { > + pci_read_config_word(dev, pos + PCI_EXP_LNKCTL, ®16); > + reg16 = (reg16 & hpp->pci_exp_lnkctl_and) > + | hpp->pci_exp_lnkctl_or; > + pci_write_config_word(dev, pos + PCI_EXP_LNKCTL, reg16); > + } > + > + /* Find Advanced Error Reporting Enhanced Capability */ > + pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ERR); > + if (!pos) > + return; > + > + /* Initialize Uncorrectable Error Mask Register */ > + pci_read_config_dword(dev, pos + PCI_ERR_UNCOR_MASK, ®32); > + reg32 = (reg32 & hpp->unc_err_mask_and) | hpp->unc_err_mask_or; > + pci_write_config_dword(dev, pos + PCI_ERR_UNCOR_MASK, reg32); > + > + /* Initialize Uncorrectable Error Severity Register */ > + pci_read_config_dword(dev, pos + PCI_ERR_UNCOR_SEVER, ®32); > + reg32 = (reg32 & hpp->unc_err_sever_and) | hpp->unc_err_sever_or; > + pci_write_config_dword(dev, pos + PCI_ERR_UNCOR_SEVER, reg32); > + > + /* Initialize Correctable Error Mask Register */ > + pci_read_config_dword(dev, pos + PCI_ERR_COR_MASK, ®32); > + reg32 = (reg32 & hpp->cor_err_mask_and) | hpp->cor_err_mask_or; > + pci_write_config_dword(dev, pos + PCI_ERR_COR_MASK, reg32); > + > + /* Initialize Advanced Error Capabilities and Control Register */ > + pci_read_config_dword(dev, pos + PCI_ERR_CAP, ®32); > + reg32 = (reg32 & hpp->adv_err_cap_and) | hpp->adv_err_cap_or; > + pci_write_config_dword(dev, pos + PCI_ERR_CAP, reg32); > + > + /* > + * FIXME: The following two registers are not supported yet. > + * > + * o Secondary Uncorrectable Error Severity Register > + * o Secondary Uncorrectable Error Mask Register > + */ > +} > + Hi, All the struct hpp_* data structures and program_hpp_type?() functions are ACPI specific. So I think we had better place them to drivers/pci/hotplug/acpi_pcihp.c. Thanks, Kenji Kaneshige -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Thursday 27 August 2009 06:33:32 pm Kenji Kaneshige wrote: > All the struct hpp_* data structures and program_hpp_type?() > functions are ACPI specific. So I think we had better place > them to drivers/pci/hotplug/acpi_pcihp.c. Kenji-san, Thanks for the suggestion; I think you're absolutely right. I'll post some updated patches soon. Bjorn -- To unsubscribe from this list: send the line "unsubscribe linux-pci" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/drivers/pci/hotplug/Makefile b/drivers/pci/hotplug/Makefile index 2aa117c..3625b09 100644 --- a/drivers/pci/hotplug/Makefile +++ b/drivers/pci/hotplug/Makefile @@ -22,7 +22,7 @@ obj-$(CONFIG_HOTPLUG_PCI_SGI) += sgi_hotplug.o # Link this last so it doesn't claim devices that have a real hotplug driver obj-$(CONFIG_HOTPLUG_PCI_FAKE) += fakephp.o -pci_hotplug-objs := pci_hotplug_core.o +pci_hotplug-objs := pci_hotplug_core.o pcihp_slot.o ifdef CONFIG_HOTPLUG_PCI_CPCI pci_hotplug-objs += cpci_hotplug_core.o \ diff --git a/drivers/pci/hotplug/pcihp_slot.c b/drivers/pci/hotplug/pcihp_slot.c new file mode 100644 index 0000000..82d4698 --- /dev/null +++ b/drivers/pci/hotplug/pcihp_slot.c @@ -0,0 +1,181 @@ +/* + * Copyright (C) 1995,2001 Compaq Computer Corporation + * Copyright (C) 2001 Greg Kroah-Hartman (greg@kroah.com) + * Copyright (C) 2001 IBM Corp. + * Copyright (C) 2003-2004 Intel Corporation + * (c) Copyright 2009 Hewlett-Packard Development Company, L.P. + * + * All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or (at + * your option) any later version. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or + * NON INFRINGEMENT. See the GNU General Public License for more + * details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. + */ + +#include <linux/pci.h> +#include <linux/pci_hotplug.h> + +static struct hpp_type0 pci_default_type0 = { + .revision = 1, + .cache_line_size = 8, + .latency_timer = 0x40, + .enable_serr = 0, + .enable_perr = 0, +}; + +static void program_hpp_type0(struct pci_dev *dev, struct hpp_type0 *hpp) +{ + u16 pci_cmd, pci_bctl; + + if (hpp->revision > 1) { + dev_warn(&dev->dev, + "PCI setting record rev %d not supported\n", + hpp->revision); + return; + } + + pci_write_config_byte(dev, PCI_CACHE_LINE_SIZE, hpp->cache_line_size); + pci_write_config_byte(dev, PCI_LATENCY_TIMER, hpp->latency_timer); + pci_read_config_word(dev, PCI_COMMAND, &pci_cmd); + if (hpp->enable_serr) + pci_cmd |= PCI_COMMAND_SERR; + else + pci_cmd &= ~PCI_COMMAND_SERR; + if (hpp->enable_perr) + pci_cmd |= PCI_COMMAND_PARITY; + else + pci_cmd &= ~PCI_COMMAND_PARITY; + pci_write_config_word(dev, PCI_COMMAND, pci_cmd); + + /* Program bridge control value */ + if ((dev->class >> 8) == PCI_CLASS_BRIDGE_PCI) { + pci_write_config_byte(dev, PCI_SEC_LATENCY_TIMER, + hpp->latency_timer); + pci_read_config_word(dev, PCI_BRIDGE_CONTROL, &pci_bctl); + if (hpp->enable_serr) + pci_bctl |= PCI_BRIDGE_CTL_SERR; + else + pci_bctl &= ~PCI_BRIDGE_CTL_SERR; + if (hpp->enable_perr) + pci_bctl |= PCI_BRIDGE_CTL_PARITY; + else + pci_bctl &= ~PCI_BRIDGE_CTL_PARITY; + pci_write_config_word(dev, PCI_BRIDGE_CONTROL, pci_bctl); + } +} + +static void program_hpp_type2(struct pci_dev *dev, struct hpp_type2 *hpp) +{ + int pos; + u16 reg16; + u32 reg32; + + if (hpp->revision > 1) { + dev_warn(&dev->dev, + "PCI Express setting record rev %d not supported\n", + hpp->revision); + return; + } + + /* Find PCI Express capability */ + pos = pci_find_capability(dev, PCI_CAP_ID_EXP); + if (!pos) + return; + + /* Initialize Device Control Register */ + pci_read_config_word(dev, pos + PCI_EXP_DEVCTL, ®16); + reg16 = (reg16 & hpp->pci_exp_devctl_and) | hpp->pci_exp_devctl_or; + pci_write_config_word(dev, pos + PCI_EXP_DEVCTL, reg16); + + /* Initialize Link Control Register */ + if (dev->subordinate) { + pci_read_config_word(dev, pos + PCI_EXP_LNKCTL, ®16); + reg16 = (reg16 & hpp->pci_exp_lnkctl_and) + | hpp->pci_exp_lnkctl_or; + pci_write_config_word(dev, pos + PCI_EXP_LNKCTL, reg16); + } + + /* Find Advanced Error Reporting Enhanced Capability */ + pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ERR); + if (!pos) + return; + + /* Initialize Uncorrectable Error Mask Register */ + pci_read_config_dword(dev, pos + PCI_ERR_UNCOR_MASK, ®32); + reg32 = (reg32 & hpp->unc_err_mask_and) | hpp->unc_err_mask_or; + pci_write_config_dword(dev, pos + PCI_ERR_UNCOR_MASK, reg32); + + /* Initialize Uncorrectable Error Severity Register */ + pci_read_config_dword(dev, pos + PCI_ERR_UNCOR_SEVER, ®32); + reg32 = (reg32 & hpp->unc_err_sever_and) | hpp->unc_err_sever_or; + pci_write_config_dword(dev, pos + PCI_ERR_UNCOR_SEVER, reg32); + + /* Initialize Correctable Error Mask Register */ + pci_read_config_dword(dev, pos + PCI_ERR_COR_MASK, ®32); + reg32 = (reg32 & hpp->cor_err_mask_and) | hpp->cor_err_mask_or; + pci_write_config_dword(dev, pos + PCI_ERR_COR_MASK, reg32); + + /* Initialize Advanced Error Capabilities and Control Register */ + pci_read_config_dword(dev, pos + PCI_ERR_CAP, ®32); + reg32 = (reg32 & hpp->adv_err_cap_and) | hpp->adv_err_cap_or; + pci_write_config_dword(dev, pos + PCI_ERR_CAP, reg32); + + /* + * FIXME: The following two registers are not supported yet. + * + * o Secondary Uncorrectable Error Severity Register + * o Secondary Uncorrectable Error Mask Register + */ +} + +void pci_configure_slot(struct pci_dev *dev) +{ + struct pci_dev *cdev; + struct hotplug_params hpp; + struct hpp_type0 *type0 = NULL; + int ret; + + if (!(dev->hdr_type == PCI_HEADER_TYPE_NORMAL || + (dev->hdr_type == PCI_HEADER_TYPE_BRIDGE && + (dev->class >> 8) == PCI_CLASS_BRIDGE_PCI))) + return; + + memset(&hpp, 0, sizeof(hpp)); + ret = acpi_get_hp_params(dev, &hpp); + if (ret == 0 && hpp.t0) + type0 = hpp.t0; + else { + if (!dev->is_pcie) { + type0 = &pci_default_type0; + dev_warn(&dev->dev, + "no PCI setting record; using defaults\n"); + } + } + + if (dev->is_pcie && hpp.t2) + program_hpp_type2(dev, hpp.t2); + + if (hpp.t1) + dev_info(&dev->dev, + "PCI-X setting record not supported\n"); + + if (type0) + program_hpp_type0(dev, type0); + + if (dev->subordinate) { + list_for_each_entry(cdev, &dev->subordinate->devices, + bus_list) + pci_configure_slot(cdev); + } +} diff --git a/include/linux/pci_hotplug.h b/include/linux/pci_hotplug.h index 0f3e1db..8cff1d5 100644 --- a/include/linux/pci_hotplug.h +++ b/include/linux/pci_hotplug.h @@ -230,6 +230,10 @@ int acpi_get_hp_params(struct pci_dev *dev, struct hotplug_params *hpp); int acpi_get_hp_hw_control_from_firmware(struct pci_dev *dev, u32 flags); int acpi_pci_check_ejectable(struct pci_bus *pbus, acpi_handle handle); int acpi_pci_detect_ejectable(struct pci_bus *pbus); +#else +#define acpi_get_hp_params(dev, hpp) (-ENODEV) #endif + +void pci_configure_slot(struct pci_dev *dev); #endif