All of lore.kernel.org
 help / color / mirror / Atom feed
From: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
To: Tomasz Nowicki <tn@semihalf.com>
Cc: bhelgaas@google.com, arnd@arndb.de, will.deacon@arm.com,
	catalin.marinas@arm.com, rjw@rjwysocki.net,
	hanjun.guo@linaro.org, jiang.liu@linux.intel.com,
	robert.richter@caviumnetworks.com,
	Narinder.Dhillon@caviumnetworks.com, ddaney@caviumnetworks.com,
	Liviu.Dudau@arm.com, tglx@linutronix.de, wangyijing@huawei.com,
	Suravee.Suthikulpanit@amd.com, msalter@redhat.com,
	linux-pci@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
	linux-acpi@vger.kernel.org, linux-kernel@vger.kernel.org,
	linaro-acpi@lists.linaro.org
Subject: Re: [PATCH V1 10/11] pci, acpi: Provide generic way to assign bus domain number.
Date: Tue, 3 Nov 2015 16:10:59 +0000	[thread overview]
Message-ID: <20151103161059.GE3574@red-moon> (raw)
In-Reply-To: <1445963922-22711-11-git-send-email-tn@semihalf.com>

On Tue, Oct 27, 2015 at 05:38:41PM +0100, Tomasz Nowicki wrote:
> Architectures which support PCI_DOMAINS_GENERIC (like ARM64)
> cannot call pci_bus_assign_domain_nr along ACPI PCI host bridge
> initialization since this function needs valid parent device reference
> to be able to retrieve domain number (aka segment).
> 
> We can omit that blocker and pass down host bridge device via
> pci_create_root_bus parameter and then be able to evaluate _SEG method
> being in pci_bus_assign_domain_nr.
> 
> Note that _SEG method is optional, therefore _SEG absence means
> that all PCI buses belong to domain 0.
> 
> Signed-off-by: Tomasz Nowicki <tn@semihalf.com>
> ---
>  drivers/acpi/pci_root.c |  2 +-
>  drivers/pci/pci.c       | 32 +++++++++++++++++++++++++++-----
>  2 files changed, 28 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c
> index 850d7bf..e682dc6 100644
> --- a/drivers/acpi/pci_root.c
> +++ b/drivers/acpi/pci_root.c
> @@ -839,7 +839,7 @@ struct pci_bus *acpi_pci_root_create(struct acpi_pci_root *root,
>  
>  	pci_acpi_root_add_resources(info);
>  	pci_add_resource(&info->resources, &root->secondary);
> -	bus = pci_create_root_bus(NULL, busnum, ops->pci_ops,
> +	bus = pci_create_root_bus(&device->dev, busnum, ops->pci_ops,
>  				  sysdata, &info->resources);

If I read x86 code correctly, they rely on the first argument to be
NULL, I think you would break x86 by doing that, see:

arch/x86/pci/acpi.c (pcibios_root_bridge_prepare())

By the way, can't we move the code setting up the ACPI_COMPANION
to core PCI code and stop relying on sysdata for that ?

Thanks,
Lorenzo

>  	if (!bus)
>  		goto out_release_info;
> diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
> index 6a9a111..17d1857 100644
> --- a/drivers/pci/pci.c
> +++ b/drivers/pci/pci.c
> @@ -25,6 +25,7 @@
>  #include <linux/device.h>
>  #include <linux/pm_runtime.h>
>  #include <linux/pci_hotplug.h>
> +#include <linux/acpi.h>
>  #include <asm-generic/pci-bridge.h>
>  #include <asm/setup.h>
>  #include "pci.h"
> @@ -4501,7 +4502,7 @@ int pci_get_new_domain_nr(void)
>  void pci_bus_assign_domain_nr(struct pci_bus *bus, struct device *parent)
>  {
>  	static int use_dt_domains = -1;
> -	int domain = of_get_pci_domain_nr(parent->of_node);
> +	int domain;
>  
>  	/*
>  	 * Check DT domain and use_dt_domains values.
> @@ -4523,14 +4524,35 @@ void pci_bus_assign_domain_nr(struct pci_bus *bus, struct device *parent)
>  	 * API and update the use_dt_domains value to keep track of method we
>  	 * are using to assign domain numbers (use_dt_domains = 0).
>  	 *
> +	 * IF ACPI, we expect non-DT method (use_dt_domains == -1)
> +	 * and call _SEG method for corresponding host bridge device.
> +	 * If _SEG method does not exist, following ACPI spec (6.5.6)
> +	 * all PCI buses belong to domain 0.
> +	 *
>  	 * All other combinations imply we have a platform that is trying
> -	 * to mix domain numbers obtained from DT and pci_get_new_domain_nr(),
> -	 * which is a recipe for domain mishandling and it is prevented by
> -	 * invalidating the domain value (domain = -1) and printing a
> -	 * corresponding error.
> +	 * to mix domain numbers obtained from DT, ACPI and
> +	 * pci_get_new_domain_nr(), which is a recipe for domain mishandling and
> +	 * it is prevented by invalidating the domain value (domain = -1) and
> +	 * printing a corresponding error.
>  	 */
> +
> +	domain = of_get_pci_domain_nr(parent->of_node);
>  	if (domain >= 0 && use_dt_domains) {
>  		use_dt_domains = 1;
> +#ifdef CONFIG_ACPI
> +	} else if (!acpi_disabled && use_dt_domains == -1) {
> +		struct acpi_device *acpi_dev = to_acpi_device(parent);
> +		unsigned long long segment = 0;
> +		acpi_status status;
> +
> +		status = acpi_evaluate_integer(acpi_dev->handle,
> +					       METHOD_NAME__SEG, NULL,
> +					       &segment);
> +		if (ACPI_FAILURE(status) && status != AE_NOT_FOUND)
> +			dev_err(&acpi_dev->dev,  "can't evaluate _SEG\n");
> +
> +		domain = segment;
> +#endif
>  	} else if (domain < 0 && use_dt_domains != 1) {
>  		use_dt_domains = 0;
>  		domain = pci_get_new_domain_nr();
> -- 
> 1.9.1
> 

WARNING: multiple messages have this Message-ID (diff)
From: lorenzo.pieralisi@arm.com (Lorenzo Pieralisi)
To: linux-arm-kernel@lists.infradead.org
Subject: [PATCH V1 10/11] pci, acpi: Provide generic way to assign bus domain number.
Date: Tue, 3 Nov 2015 16:10:59 +0000	[thread overview]
Message-ID: <20151103161059.GE3574@red-moon> (raw)
In-Reply-To: <1445963922-22711-11-git-send-email-tn@semihalf.com>

On Tue, Oct 27, 2015 at 05:38:41PM +0100, Tomasz Nowicki wrote:
> Architectures which support PCI_DOMAINS_GENERIC (like ARM64)
> cannot call pci_bus_assign_domain_nr along ACPI PCI host bridge
> initialization since this function needs valid parent device reference
> to be able to retrieve domain number (aka segment).
> 
> We can omit that blocker and pass down host bridge device via
> pci_create_root_bus parameter and then be able to evaluate _SEG method
> being in pci_bus_assign_domain_nr.
> 
> Note that _SEG method is optional, therefore _SEG absence means
> that all PCI buses belong to domain 0.
> 
> Signed-off-by: Tomasz Nowicki <tn@semihalf.com>
> ---
>  drivers/acpi/pci_root.c |  2 +-
>  drivers/pci/pci.c       | 32 +++++++++++++++++++++++++++-----
>  2 files changed, 28 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c
> index 850d7bf..e682dc6 100644
> --- a/drivers/acpi/pci_root.c
> +++ b/drivers/acpi/pci_root.c
> @@ -839,7 +839,7 @@ struct pci_bus *acpi_pci_root_create(struct acpi_pci_root *root,
>  
>  	pci_acpi_root_add_resources(info);
>  	pci_add_resource(&info->resources, &root->secondary);
> -	bus = pci_create_root_bus(NULL, busnum, ops->pci_ops,
> +	bus = pci_create_root_bus(&device->dev, busnum, ops->pci_ops,
>  				  sysdata, &info->resources);

If I read x86 code correctly, they rely on the first argument to be
NULL, I think you would break x86 by doing that, see:

arch/x86/pci/acpi.c (pcibios_root_bridge_prepare())

By the way, can't we move the code setting up the ACPI_COMPANION
to core PCI code and stop relying on sysdata for that ?

Thanks,
Lorenzo

>  	if (!bus)
>  		goto out_release_info;
> diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
> index 6a9a111..17d1857 100644
> --- a/drivers/pci/pci.c
> +++ b/drivers/pci/pci.c
> @@ -25,6 +25,7 @@
>  #include <linux/device.h>
>  #include <linux/pm_runtime.h>
>  #include <linux/pci_hotplug.h>
> +#include <linux/acpi.h>
>  #include <asm-generic/pci-bridge.h>
>  #include <asm/setup.h>
>  #include "pci.h"
> @@ -4501,7 +4502,7 @@ int pci_get_new_domain_nr(void)
>  void pci_bus_assign_domain_nr(struct pci_bus *bus, struct device *parent)
>  {
>  	static int use_dt_domains = -1;
> -	int domain = of_get_pci_domain_nr(parent->of_node);
> +	int domain;
>  
>  	/*
>  	 * Check DT domain and use_dt_domains values.
> @@ -4523,14 +4524,35 @@ void pci_bus_assign_domain_nr(struct pci_bus *bus, struct device *parent)
>  	 * API and update the use_dt_domains value to keep track of method we
>  	 * are using to assign domain numbers (use_dt_domains = 0).
>  	 *
> +	 * IF ACPI, we expect non-DT method (use_dt_domains == -1)
> +	 * and call _SEG method for corresponding host bridge device.
> +	 * If _SEG method does not exist, following ACPI spec (6.5.6)
> +	 * all PCI buses belong to domain 0.
> +	 *
>  	 * All other combinations imply we have a platform that is trying
> -	 * to mix domain numbers obtained from DT and pci_get_new_domain_nr(),
> -	 * which is a recipe for domain mishandling and it is prevented by
> -	 * invalidating the domain value (domain = -1) and printing a
> -	 * corresponding error.
> +	 * to mix domain numbers obtained from DT, ACPI and
> +	 * pci_get_new_domain_nr(), which is a recipe for domain mishandling and
> +	 * it is prevented by invalidating the domain value (domain = -1) and
> +	 * printing a corresponding error.
>  	 */
> +
> +	domain = of_get_pci_domain_nr(parent->of_node);
>  	if (domain >= 0 && use_dt_domains) {
>  		use_dt_domains = 1;
> +#ifdef CONFIG_ACPI
> +	} else if (!acpi_disabled && use_dt_domains == -1) {
> +		struct acpi_device *acpi_dev = to_acpi_device(parent);
> +		unsigned long long segment = 0;
> +		acpi_status status;
> +
> +		status = acpi_evaluate_integer(acpi_dev->handle,
> +					       METHOD_NAME__SEG, NULL,
> +					       &segment);
> +		if (ACPI_FAILURE(status) && status != AE_NOT_FOUND)
> +			dev_err(&acpi_dev->dev,  "can't evaluate _SEG\n");
> +
> +		domain = segment;
> +#endif
>  	} else if (domain < 0 && use_dt_domains != 1) {
>  		use_dt_domains = 0;
>  		domain = pci_get_new_domain_nr();
> -- 
> 1.9.1
> 

  parent reply	other threads:[~2015-11-03 16:10 UTC|newest]

Thread overview: 131+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-10-27 16:38 [PATCH V1 00/11] MMCONFIG refactoring and ARM64 PCI hostbridge init based on ACPI Tomasz Nowicki
2015-10-27 16:38 ` Tomasz Nowicki
2015-10-27 16:38 ` [PATCH V1 01/11] x86, pci: Reorder logic of pci_mmconfig_insert() function Tomasz Nowicki
2015-10-27 16:38   ` Tomasz Nowicki
2015-10-27 16:38   ` Tomasz Nowicki
2015-10-27 16:38 ` [PATCH V1 02/11] x86, pci, acpi: Move arch-agnostic MMCONFIG (aka ECAM) and ACPI code out of arch/x86/ directory Tomasz Nowicki
2015-10-27 16:38   ` Tomasz Nowicki
2015-10-27 16:38   ` Tomasz Nowicki
2015-10-27 16:38 ` [PATCH V1 03/11] pci, acpi, mcfg: Provide generic implementation of MCFG code initialization Tomasz Nowicki
2015-10-27 16:38   ` Tomasz Nowicki
2015-10-27 16:38 ` [PATCH V1 04/11] x86, pci: mmconfig_{32,64}.c code refactoring - remove code duplication Tomasz Nowicki
2015-10-27 16:38   ` [PATCH V1 04/11] x86, pci: mmconfig_{32, 64}.c " Tomasz Nowicki
2015-10-27 16:38 ` [PATCH V1 05/11] x86, pci, ecam: mmconfig_64.c becomes default implementation for ECAM driver Tomasz Nowicki
2015-10-27 16:38   ` Tomasz Nowicki
2015-10-27 16:38   ` Tomasz Nowicki
2015-10-27 16:38 ` [PATCH V1 06/11] pci, acpi, mcfg: Provide default RAW ACPI PCI config space accessors Tomasz Nowicki
2015-10-27 16:38   ` Tomasz Nowicki
2015-10-27 16:38   ` Tomasz Nowicki
2015-10-27 16:38 ` [PATCH V1 07/11] XEN / PCI: Remove the dependence on arch x86 when PCI_MMCONFIG=y Tomasz Nowicki
2015-10-27 16:38   ` Tomasz Nowicki
2015-10-27 16:47   ` [Linaro-acpi] " Tomasz Nowicki
2015-10-27 16:47     ` Tomasz Nowicki
2015-10-27 17:25     ` Boris Ostrovsky
2015-10-27 17:25       ` Boris Ostrovsky
2015-10-28 10:49       ` Stefano Stabellini
2015-10-28 10:49         ` Stefano Stabellini
2015-10-28 10:49         ` Stefano Stabellini
2015-10-28 10:56         ` Tomasz Nowicki
2015-10-28 10:56           ` Tomasz Nowicki
2015-10-28 13:45           ` Hanjun Guo
2015-10-28 13:45             ` Hanjun Guo
2015-10-28 14:07             ` Boris Ostrovsky
2015-10-28 14:07               ` Boris Ostrovsky
2015-10-27 16:38 ` [PATCH V1 08/11] pci, acpi, ecam: Add flag to indicate whether ECAM region was hot added or not Tomasz Nowicki
2015-10-27 16:38   ` Tomasz Nowicki
2015-10-27 16:38   ` Tomasz Nowicki
2015-10-27 16:38 ` [PATCH V1 09/11] x86, pci: Use previously added ECAM hot_added flag to remove ECAM regions Tomasz Nowicki
2015-10-27 16:38   ` Tomasz Nowicki
2015-10-27 16:38   ` Tomasz Nowicki
2015-10-27 16:38 ` [PATCH V1 10/11] pci, acpi: Provide generic way to assign bus domain number Tomasz Nowicki
2015-10-27 16:38   ` Tomasz Nowicki
2015-10-28 11:38   ` Liviu.Dudau
2015-10-28 11:38     ` Liviu.Dudau at arm.com
2015-10-28 12:47     ` [Linaro-acpi] " Tomasz Nowicki
2015-10-28 12:47       ` Tomasz Nowicki
2015-11-03 16:10   ` Lorenzo Pieralisi [this message]
2015-11-03 16:10     ` Lorenzo Pieralisi
2015-11-04 10:04     ` [Linaro-acpi] " Tomasz Nowicki
2015-11-04 10:04       ` Tomasz Nowicki
2015-10-27 16:38 ` [PATCH V1 11/11] arm64, pci, acpi: Support for ACPI based PCI hostbridge init Tomasz Nowicki
2015-10-27 16:38   ` Tomasz Nowicki
2015-10-28 11:49   ` Liviu.Dudau
2015-10-28 11:49     ` Liviu.Dudau at arm.com
2015-10-28 11:49     ` Liviu.Dudau
2015-10-28 13:42     ` Tomasz Nowicki
2015-10-28 13:42       ` Tomasz Nowicki
2015-10-28 13:51       ` Liviu.Dudau
2015-10-28 13:51         ` Liviu.Dudau at arm.com
2015-11-03 14:32     ` Lorenzo Pieralisi
2015-11-03 14:32       ` Lorenzo Pieralisi
2015-11-03 16:28       ` Liviu.Dudau
2015-11-03 16:28         ` Liviu.Dudau at arm.com
2015-10-28 18:46   ` Sinan Kaya
2015-10-28 18:46     ` Sinan Kaya
2015-10-28 20:36     ` Sinan Kaya
2015-10-28 20:36       ` Sinan Kaya
2015-10-29 11:38       ` Tomasz Nowicki
2015-10-29 11:38         ` Tomasz Nowicki
2015-10-29 15:01         ` Sinan Kaya
2015-10-29 15:01           ` Sinan Kaya
2015-10-29 15:53           ` Tomasz Nowicki
2015-10-29 15:53             ` Tomasz Nowicki
2015-10-29 16:20             ` Sinan Kaya
2015-10-29 16:20               ` Sinan Kaya
2015-10-29 14:57       ` Sinan Kaya
2015-10-29 14:57         ` Sinan Kaya
2015-10-29 16:27         ` Tomasz Nowicki
2015-10-29 16:27           ` Tomasz Nowicki
2015-11-03 14:15     ` Lorenzo Pieralisi
2015-11-03 14:15       ` Lorenzo Pieralisi
2015-11-03 14:39       ` Tomasz Nowicki
2015-11-03 14:39         ` Tomasz Nowicki
2015-11-03 15:10         ` Sinan Kaya
2015-11-03 15:10           ` Sinan Kaya
2015-11-03 15:59           ` Arnd Bergmann
2015-11-03 15:59             ` Arnd Bergmann
2015-11-03 16:33             ` Sinan Kaya
2015-11-03 16:33               ` Sinan Kaya
2015-11-03 16:55               ` Arnd Bergmann
2015-11-03 16:55                 ` Arnd Bergmann
2015-11-03 17:43                 ` Sinan Kaya
2015-11-03 17:43                   ` Sinan Kaya
2015-11-05 14:48                   ` [Linaro-acpi] " Sinan Kaya
2015-11-05 14:48                     ` Sinan Kaya
2015-11-03 15:19       ` Hanjun Guo
2015-11-03 15:19         ` Hanjun Guo
2015-11-03 15:19         ` Hanjun Guo
2015-11-03 17:39         ` David Daney
2015-11-03 17:39           ` David Daney
2015-11-03 17:39           ` David Daney
2015-11-03 18:00           ` Gabriele Paoloni
2015-11-03 18:00             ` Gabriele Paoloni
2015-11-03 18:00             ` Gabriele Paoloni
2015-11-03 16:55   ` Lorenzo Pieralisi
2015-11-03 16:55     ` Lorenzo Pieralisi
2015-11-04  9:59     ` Tomasz Nowicki
2015-11-04  9:59       ` Tomasz Nowicki
2015-11-04 10:11     ` Tomasz Nowicki
2015-11-04 10:11       ` Tomasz Nowicki
2015-10-30  4:07 ` [PATCH V1 00/11] MMCONFIG refactoring and ARM64 PCI hostbridge init based on ACPI Jon Masters
2015-10-30  4:07   ` Jon Masters
2015-10-30  4:50   ` Hanjun Guo
2015-10-30  4:50     ` Hanjun Guo
2015-10-30  4:50     ` Hanjun Guo
2015-10-30  8:26   ` Tomasz Nowicki
2015-10-30  8:26     ` Tomasz Nowicki
2015-10-30 16:38 ` Suravee Suthikulpanit
2015-10-30 16:38   ` Suravee Suthikulpanit
2015-10-30 16:38   ` Suravee Suthikulpanit
2015-12-07 20:31 ` Bjorn Helgaas
2015-12-07 20:31   ` Bjorn Helgaas
2015-12-09 10:01   ` Jayachandran C.
2015-12-09 10:01     ` Jayachandran C.
2015-12-09 15:55     ` Lorenzo Pieralisi
2015-12-09 15:55       ` Lorenzo Pieralisi
2015-12-16 12:51       ` Jayachandran C.
2015-12-16 12:51         ` Jayachandran C.
2015-12-16 14:05         ` Lorenzo Pieralisi
2015-12-16 14:05           ` Lorenzo Pieralisi
2015-12-08 17:43 ` Jeremy Linton
2015-12-08 17:43   ` Jeremy Linton

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=20151103161059.GE3574@red-moon \
    --to=lorenzo.pieralisi@arm.com \
    --cc=Liviu.Dudau@arm.com \
    --cc=Narinder.Dhillon@caviumnetworks.com \
    --cc=Suravee.Suthikulpanit@amd.com \
    --cc=arnd@arndb.de \
    --cc=bhelgaas@google.com \
    --cc=catalin.marinas@arm.com \
    --cc=ddaney@caviumnetworks.com \
    --cc=hanjun.guo@linaro.org \
    --cc=jiang.liu@linux.intel.com \
    --cc=linaro-acpi@lists.linaro.org \
    --cc=linux-acpi@vger.kernel.org \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-pci@vger.kernel.org \
    --cc=msalter@redhat.com \
    --cc=rjw@rjwysocki.net \
    --cc=robert.richter@caviumnetworks.com \
    --cc=tglx@linutronix.de \
    --cc=tn@semihalf.com \
    --cc=wangyijing@huawei.com \
    --cc=will.deacon@arm.com \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.