From: Liviu.Dudau@arm.com
To: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Cc: Bjorn Helgaas <helgaas@kernel.org>,
Tomasz Nowicki <tn@semihalf.com>,
arnd@arndb.de, will.deacon@arm.com, catalin.marinas@arm.com,
rafael@kernel.org, hanjun.guo@linaro.org, okaya@codeaurora.org,
jiang.liu@linux.intel.com, jchandra@broadcom.com,
robert.richter@caviumnetworks.com, mw@semihalf.com,
ddaney@caviumnetworks.com, 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, jcm@redhat.com
Subject: Re: [PATCH V6 02/13] pci, acpi: Provide generic way to assign bus domain number.
Date: Thu, 28 Apr 2016 09:13:50 +0100 [thread overview]
Message-ID: <20160428081349.GU28464@e106497-lin.cambridge.arm.com> (raw)
In-Reply-To: <20160427173118.GA26653@red-moon>
On Wed, Apr 27, 2016 at 06:31:29PM +0100, Lorenzo Pieralisi wrote:
> On Wed, Apr 27, 2016 at 11:44:53AM -0500, Bjorn Helgaas wrote:
> > On Wed, Apr 27, 2016 at 12:17:58PM +0100, Lorenzo Pieralisi wrote:
> > > On Tue, Apr 26, 2016 at 09:26:49PM -0500, Bjorn Helgaas wrote:
> > > > On Fri, Apr 15, 2016 at 07:06:37PM +0200, Tomasz Nowicki wrote:
> > > > > As we now have valid PCI host bridge device reference we can
> > > > > introduce code that is going to find its bus domain number using
> > > > > ACPI _SEG method.
> > > > >
> > > > > Note that _SEG method is optional, therefore _SEG absence means
> > > > > that all PCI buses belong to domain 0.
> > > > >
> > > > > While at it, for the sake of code clarity we put ACPI and DT domain
> > > > > assign methods into the corresponding helpers.
> > > > >
> > > > > Signed-off-by: Tomasz Nowicki <tn@semihalf.com>
> > > > > Reviewed-by: Liviu Dudau <Liviu.Dudau@arm.com>
> > > > > Tested-by: Suravee Suthikulpanit <Suravee.Suthikulpanit@amd.com>
> > > > > Tested-by: Jeremy Linton <jeremy.linton@arm.com>
> > > > > Tested-by: Duc Dang <dhdang@apm.com>
> > > > > Tested-by: Dongdong Liu <liudongdong3@huawei.com>
> > > > > Tested-by: Hanjun Guo <hanjun.guo@linaro.org>
> > > > > Tested-by: Graeme Gregory <graeme.gregory@linaro.org>
> > > > > Tested-by: Sinan Kaya <okaya@codeaurora.org>
> > > > > ---
> > > > > drivers/acpi/pci_root.c | 18 ++++++++++++++++++
> > > > > drivers/pci/pci.c | 11 +++++++++--
> > > > > include/linux/pci-acpi.h | 2 ++
> > > > > 3 files changed, 29 insertions(+), 2 deletions(-)
> > > > >
> > > > > diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c
> > > > > index 4581e0e..d9a70c4 100644
> > > > > --- a/drivers/acpi/pci_root.c
> > > > > +++ b/drivers/acpi/pci_root.c
> > > > > @@ -419,6 +419,24 @@ out:
> > > > > }
> > > > > EXPORT_SYMBOL(acpi_pci_osc_control_set);
> > > > >
> > > > > +int acpi_pci_bus_domain_nr(struct device *parent)
> >
> > It looks like acpi_pci_bus_domain_nr() could be under #ifdef
> > CONFIG_PCI_DOMAINS_GENERIC, right?
>
> Yes it should.
>
> > > > > +{
> > > > > + struct acpi_device *acpi_dev = to_acpi_device(parent);
> > > > > + unsigned long long segment = 0;
> > > > > + acpi_status status;
> > > > > +
> > > > > + /*
> > > > > + * If _SEG method does not exist, following ACPI spec (6.5.6)
> > > > > + * all PCI buses belong to domain 0.
> > > > > + */
> > > > > + status = acpi_evaluate_integer(acpi_dev->handle, METHOD_NAME__SEG, NULL,
> > > > > + &segment);
> > > >
> > > > We already have code in acpi_pci_root_add() to evaluate _SEG. We
> > > > don't want to evaluate it *twice*, do we?
> > > >
> > > > I was sort of expecting that if you added it here, we'd remove the
> > > > existing call, but it looks like you're keeping both?
> > >
> > > We can't remove the existing call, since it is used on X86 and IA64
> > > to store the segment number that, in the process, is used in their
> > > pci_domain_nr() arch specific callback to retrieve the domain nr.
> > >
> > > On ARM64, that selects PCI_DOMAINS_GENERIC, we have to find a way
> > > to retrieve the domain number that is not arch dependent, since
> > > this is generic code, we can't rely on any bus->sysdata format (unless
> > > we do something like JC did below), therefore the only way is to call
> > > the _SEG method *again* here, which also forced Tomasz to go through
> > > the ACPI_COMPANION setting song and dance and pass the parent pointer
> > > to pci_create_root_bus() (see patch 1), which BTW is a source of
> > > trouble on its own as you noticed.
> > >
> > > JC solved it differently, via sysdata and pseudo-generic code:
> > >
> > > http://www.spinics.net/lists/arm-kernel/msg478167.html
> >
> > The thing I don't like about this is the special case of checking
> > parent and parent->of_node to figure out whether we should use the
> > segment from ACPI and the fragility of depending on the fact that the
> > companion hasn't been set yet.
> >
> > > http://www.spinics.net/lists/arm-kernel/msg478169.html
> > >
> > > I like neither, we need the lesser of two evils though.
> >
> > Today we call pci_bus_assign_domain_nr() from the PCI core (from
> > pci_create_root_bus()). This is only implemented for
> > PCI_DOMAINS_GENERIC, but even so, it fiddles around to figure out
> > whether to get the domain from DT or to assign a new one.
> >
> > That seems backwards to me. The host bridge drivers already know
> > where the domain should come from (ACPI _SEG, DT, etc.) and in the
> > long term, I think they should be responsible for looking up or
> > assigning a domain number *before* they call pci_create_root_bus().
>
> Yes, the question still is how pci_create_root_bus() can get that
> value (I am pretty certain this was heavily debated in the past, which
> does not mean we can't give it another try).
The main issue is that pci_create_root_bus() does a weird dance trying to
figure out if the root bus hasn't been already allocated. It allocates a
new bus, assigns a domain number and then it tries to find it in the list
of already allocated busses. Because pci_alloc_bus() does not pass any
additional information, pci_bus_assign_domain_nr() needs to try to guess
where the barely initialised bus should live and give you back a number.
Simplifying the creation of root busses to be the job of the host bridges
would greatly simplify the code as well.
Best regards,
Liviu
>
> > > > > + if (ACPI_FAILURE(status) && status != AE_NOT_FOUND)
> > > > > + dev_err(&acpi_dev->dev, "can't evaluate _SEG\n");
> > > > > +
> > > > > + return segment;
> > > > > +}
> > > > > +
> > > > > static void negotiate_os_control(struct acpi_pci_root *root, int *no_aspm)
> > > > > {
> > > > > u32 support, control, requested;
> > > > > diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
> > > > > index 25e0327..1a74e87 100644
> > > > > --- a/drivers/pci/pci.c
> > > > > +++ b/drivers/pci/pci.c
> > > > > @@ -19,6 +19,7 @@
> > > > > #include <linux/spinlock.h>
> > > > > #include <linux/string.h>
> > > > > #include <linux/log2.h>
> > > > > +#include <linux/pci-acpi.h>
> > > > > #include <linux/pci-aspm.h>
> > > > > #include <linux/pm_wakeup.h>
> > > > > #include <linux/interrupt.h>
> > > > > @@ -4779,7 +4780,7 @@ int pci_get_new_domain_nr(void)
> > > > > }
> > > > >
> > > > > #ifdef CONFIG_PCI_DOMAINS_GENERIC
> > > > > -void pci_bus_assign_domain_nr(struct pci_bus *bus, struct device *parent)
> > > > > +static int of_pci_bus_domain_nr(struct device *parent)
> > > > > {
> > > > > static int use_dt_domains = -1;
> > > > > int domain = -1;
> > > > > @@ -4823,7 +4824,13 @@ void pci_bus_assign_domain_nr(struct pci_bus *bus, struct device *parent)
> > > > > domain = -1;
> > > > > }
> > > > >
> > > > > - bus->domain_nr = domain;
> > > > > + return domain;
> > > > > +}
> > > > > +
> > > > > +void pci_bus_assign_domain_nr(struct pci_bus *bus, struct device *parent)
> > > > > +{
> > > > > + bus->domain_nr = acpi_disabled ? of_pci_bus_domain_nr(parent) :
> > > > > + acpi_pci_bus_domain_nr(parent);
> >
> > We have the pci_bus * here, so to_pci_host_bridge(bus->bridge) gives
> > us the struct pci_host_bridge. I can't remember why we put domain_nr
> > in the struct pci_bus instead of in the struct pci_host_bridge. It
> > seems like pci_host_bridge is the more logical place for it, because
> > every bus below the host bridge must have the same domain by
> > definition.
> >
> > Would it be feasible to either (a) move domain_nr to the
> > pci_host_bridge, or (b) change acpi_pci_bus_domain_nr() so it uses the
> > struct pci_bus * or the struct device * to find the struct
> > acpi_pci_root where segment has already been stored by
> > acpi_pci_root_add()?
>
> (b) is what JC implemented even though it works differently for
> different hosts since it all depends on what's in bus->sysdata.
>
> It can certainly be done in a generic way (that works on X86 and IA64
> too), let's give it more thought.
>
> > Another wrinkle is the quirk added by 1f09b09b4de0 ("x86/PCI: Ignore
> > _SEG on HP xw9300"). x86 doesn't use PCI_DOMAINS_GENERIC yet, so this
> > patch wouldn't break it, but I hope x86 can use PCI_DOMAINS_GENERIC in
> > the future, and then it will be a problem if we evaluate _SEG again.
>
> Yes, I share your concern here and I thought about that, if that's the
> end goal let's find a solution that works across arches (or we temporarily
> use JC's code and we then generalize it).
>
> Thanks,
> Lorenzo
>
> >
> > > > > }
> > > > > #endif
> > > > > #endif
> > > > > diff --git a/include/linux/pci-acpi.h b/include/linux/pci-acpi.h
> > > > > index 89ab057..a72e22d 100644
> > > > > --- a/include/linux/pci-acpi.h
> > > > > +++ b/include/linux/pci-acpi.h
> > > > > @@ -22,6 +22,7 @@ static inline acpi_status pci_acpi_remove_pm_notifier(struct acpi_device *dev)
> > > > > {
> > > > > return acpi_remove_pm_notifier(dev);
> > > > > }
> > > > > +extern int acpi_pci_bus_domain_nr(struct device *parent);
> > > > > extern phys_addr_t acpi_pci_root_get_mcfg_addr(acpi_handle handle);
> > > > >
> > > > > static inline acpi_handle acpi_find_root_bridge_handle(struct pci_dev *pdev)
> > > > > @@ -109,6 +110,7 @@ extern const u8 pci_acpi_dsm_uuid[];
> > > > > #else /* CONFIG_ACPI */
> > > > > static inline void acpi_pci_add_bus(struct pci_bus *bus) { }
> > > > > static inline void acpi_pci_remove_bus(struct pci_bus *bus) { }
> > > > > +static inline int acpi_pci_bus_domain_nr(struct device *parent) { return -1; }
> > > > > #endif /* CONFIG_ACPI */
> > > > >
> > > > > #ifdef CONFIG_ACPI_APEI
> > > > > --
> > > > > 1.9.1
> > > > >
> > > > > --
> > > > > To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
> > > > > the body of a message to majordomo@vger.kernel.org
> > > > > More majordomo info at http://vger.kernel.org/majordomo-info.html
> > > >
> >
>
--
====================
| I would like to |
| fix the world, |
| but they're not |
| giving me the |
\ source code! /
---------------
¯\_(ツ)_/¯
next prev parent reply other threads:[~2016-04-28 8:14 UTC|newest]
Thread overview: 106+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-04-15 17:06 [PATCH V6 00/13] Support for generic ACPI based PCI host controller Tomasz Nowicki
2016-04-15 17:06 ` [PATCH V6 01/13] pci, acpi, x86, ia64: Move ACPI host bridge device companion assignment to core code Tomasz Nowicki
2016-04-15 20:41 ` kbuild test robot
2016-04-26 22:36 ` Bjorn Helgaas
2016-04-27 10:12 ` Tomasz Nowicki
2016-04-27 2:45 ` Bjorn Helgaas
2016-05-04 8:10 ` Tomasz Nowicki
2016-05-09 22:18 ` Rafael J. Wysocki
2016-05-10 10:27 ` Lorenzo Pieralisi
2016-05-09 22:56 ` Rafael J. Wysocki
2016-05-10 1:53 ` Bjorn Helgaas
2016-05-10 10:07 ` Lorenzo Pieralisi
2016-04-15 17:06 ` [PATCH V6 02/13] pci, acpi: Provide generic way to assign bus domain number Tomasz Nowicki
2016-04-27 2:26 ` Bjorn Helgaas
2016-04-27 11:17 ` Lorenzo Pieralisi
2016-04-27 16:44 ` Bjorn Helgaas
2016-04-27 17:31 ` Lorenzo Pieralisi
2016-04-28 8:13 ` Liviu.Dudau [this message]
2016-04-28 15:12 ` Bjorn Helgaas
2016-04-28 15:34 ` Arnd Bergmann
2016-04-29 22:50 ` Arnd Bergmann
2016-05-02 12:43 ` Tomasz Nowicki
2016-05-02 13:26 ` Jayachandran C
2016-05-03 11:02 ` Lorenzo Pieralisi
2016-05-03 14:22 ` Jayachandran C
2016-05-03 14:55 ` Lorenzo Pieralisi
2016-04-27 11:59 ` Tomasz Nowicki
2016-04-15 17:06 ` [PATCH V6 03/13] x86, ia64: Include acpi_pci_{add|remove}_bus to the default pcibios_{add|remove}_bus implementation Tomasz Nowicki
2016-04-27 2:34 ` Bjorn Helgaas
2016-04-27 13:19 ` Tomasz Nowicki
2016-04-15 17:06 ` [PATCH V6 04/13] pci, of: Move the PCI I/O space management to PCI core code Tomasz Nowicki
2016-04-15 17:06 ` [PATCH V6 05/13] acpi, pci: Support IO resources when parsing PCI host bridge resources Tomasz Nowicki
2016-04-27 2:39 ` Bjorn Helgaas
2016-04-27 5:36 ` Jon Masters
2016-04-28 21:53 ` Jon Masters
2016-04-27 14:26 ` Lorenzo Pieralisi
2016-04-27 15:10 ` Liviu.Dudau
2016-04-27 16:09 ` Lorenzo Pieralisi
2016-04-28 15:45 ` Bjorn Helgaas
2016-04-15 17:06 ` [PATCH V6 06/13] arm64, pci, acpi: ACPI support for legacy IRQs parsing and consolidation with DT code Tomasz Nowicki
2016-04-27 2:44 ` Bjorn Helgaas
2016-04-27 11:46 ` Lorenzo Pieralisi
2016-04-15 17:06 ` [PATCH V6 07/13] PCI: Provide common functions for ECAM mapping Tomasz Nowicki
2016-04-15 18:41 ` Arnd Bergmann
2016-04-28 21:47 ` Bjorn Helgaas
2016-04-29 8:01 ` Jayachandran C
2016-05-05 9:24 ` Jayachandran C
2016-05-05 10:38 ` Tomasz Nowicki
2016-04-15 17:06 ` [PATCH V6 08/13] PCI: generic, thunder: update to use generic ECAM API Tomasz Nowicki
2016-04-15 18:39 ` Arnd Bergmann
2016-04-16 7:20 ` Jayachandran C
2016-04-16 7:31 ` Arnd Bergmann
2016-04-16 14:36 ` Jayachandran C
2016-04-18 13:03 ` Tomasz Nowicki
2016-04-18 14:44 ` Arnd Bergmann
2016-04-18 19:31 ` Tomasz Nowicki
2016-04-19 13:06 ` Arnd Bergmann
2016-04-21 9:28 ` Tomasz Nowicki
2016-04-21 9:36 ` Arnd Bergmann
2016-04-21 10:08 ` Tomasz Nowicki
2016-04-22 14:30 ` Jon Masters
2016-04-22 16:00 ` David Daney
2016-04-28 20:14 ` Bjorn Helgaas
2016-04-28 20:40 ` Arnd Bergmann
2016-04-28 21:18 ` Bjorn Helgaas
2016-04-28 21:47 ` Jon Masters
2016-04-29 9:41 ` Lorenzo Pieralisi
2016-04-19 21:40 ` Arnd Bergmann
2016-04-20 0:22 ` Jayachandran C
2016-04-15 17:06 ` [PATCH V6 09/13] pci, acpi: Support for ACPI based generic PCI host controller Tomasz Nowicki
2016-04-20 19:12 ` Jayachandran C
2016-04-21 9:06 ` Tomasz Nowicki
2016-04-22 12:49 ` Jayachandran C
2016-04-22 14:40 ` Jon Masters
2016-04-23 15:23 ` Jon Masters
2016-04-28 21:48 ` Bjorn Helgaas
2016-04-29 8:37 ` Lorenzo Pieralisi
2016-04-29 17:35 ` Jayachandran C
2016-05-02 11:31 ` Tomasz Nowicki
2016-05-03 8:46 ` Lorenzo Pieralisi
2016-05-02 11:03 ` Tomasz Nowicki
2016-04-15 17:06 ` [PATCH V6 10/13] arm64, pci, acpi: Start using ACPI based PCI host controller driver for ARM64 Tomasz Nowicki
2016-04-15 17:06 ` [PATCH V6 11/13] pci, acpi: Match PCI config space accessors against platfrom specific quirks Tomasz Nowicki
2016-04-18 11:37 ` liudongdong (C)
2016-04-18 12:21 ` Tomasz Nowicki
2016-04-15 17:06 ` [PATCH V6 12/13] pci, pci-thunder-ecam: Add ACPI support for ThunderX ECAM Tomasz Nowicki
2016-04-19 10:26 ` Tomasz Nowicki
2016-04-19 10:41 ` [Linaro-acpi] " G Gregory
2016-04-19 11:12 ` Graeme Gregory
2016-04-19 11:22 ` Tomasz Nowicki
2016-04-19 12:29 ` G Gregory
2016-04-15 17:06 ` [PATCH V6 13/13] pci, pci-thunder-pem: Add ACPI support for ThunderX PEM Tomasz Nowicki
2016-04-15 18:19 ` [PATCH V6 00/13] Support for generic ACPI based PCI host controller Jon Masters
2016-04-16 15:31 ` Jayachandran C
2016-04-18 13:33 ` Tomasz Nowicki
2016-04-18 14:38 ` Arnd Bergmann
2016-04-18 15:26 ` Tomasz Nowicki
2016-04-17 9:23 ` Martinez Kristofer
2016-04-16 18:30 ` Duc Dang
2016-04-17 4:18 ` Sinan Kaya
2016-04-22 16:08 ` Robert Richter
2016-04-22 20:46 ` Suravee Suthikulpanit
2016-04-25 17:23 ` Jeremy Linton
2016-04-26 9:07 ` liudongdong (C)
2016-04-28 21:27 ` [PATCH] acpi: pci: QDF2432 32 bit config space accessors Christopher Covington
2016-04-28 21:35 ` Rafael J. Wysocki
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=20160428081349.GU28464@e106497-lin.cambridge.arm.com \
--to=liviu.dudau@arm.com \
--cc=Suravee.Suthikulpanit@amd.com \
--cc=arnd@arndb.de \
--cc=catalin.marinas@arm.com \
--cc=ddaney@caviumnetworks.com \
--cc=hanjun.guo@linaro.org \
--cc=helgaas@kernel.org \
--cc=jchandra@broadcom.com \
--cc=jcm@redhat.com \
--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=lorenzo.pieralisi@arm.com \
--cc=msalter@redhat.com \
--cc=mw@semihalf.com \
--cc=okaya@codeaurora.org \
--cc=rafael@kernel.org \
--cc=robert.richter@caviumnetworks.com \
--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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).