[RFC v1 0/2] x86: kconfig: simplify APIC entries

[RFC v1 0/2] x86: kconfig: simplify APIC entries

[Luis R. Rodriguez]

From: "Luis R. Rodriguez" <mcgrof@suse.com>

There's quite a bit of Kconfig entries related to X86_LOCAL_APIC
and X86_IO_APIC, and all for 32-bit purposes. Since 32-bit isn't in
a magical state of flux we now *should* know when to enable or disable
APIC stuff for these systems. Through history we've observed for
instance PCI_MSI requires these under specific scenarios, just enable
it for those learned scenarios and expect future requirements to also
be explicit about the requirements on 32-bit. This lets us remove
clutter (3 Kconfig entries) from x86 Kconfig simply for 32-bit build
purposes.

Luis R. Rodriguez (2):
  x86: kconfig: simplify enabling [IO]APIC for PCI_MSI
  x86: kconfig: remove UP [IO]APIC options

 Documentation/PCI/MSI-HOWTO.txt |  5 +----
 arch/x86/Kconfig                | 33 ++-------------------------------
 drivers/pci/Kconfig             |  2 ++
 3 files changed, 5 insertions(+), 35 deletions(-)

-- 
2.3.2.209.gd67f9d5.dirty

[RFC v1 1/2] x86: kconfig: simplify enabling [IO]APIC for PCI_MSI

[Luis R. Rodriguez]

From: "Luis R. Rodriguez" <mcgrof@suse.com>

Commit 38a1dfda by Bryan O'Donoghue fixed a build issue
when PCI_MSI support is enabled on non-SMP X86_32 systems,
this was done by introducing a new Kconfig entry,
X86_UP_APIC_MSI, with the only purpose of selecting
X86_UP_APIC under certain conditions. The real goal
in the end though is to enable X86_LOCAL_APIC and
X86_IO_APIC. We already have enough APIC related Kconfig
entries and adding more only confuses things, we can
remove X86_UP_APIC_MSI by just passing on to PCI_MSI
what it needs.

Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: Bryan O'Donoghue <pure.logic@nexus-software.ie>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Andy Shevchenko <andy.shevchenko@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Borislav Petkov <bp@suse.de>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Jan Beulich <JBeulich@suse.com>
Cc: Juergen Gross <jgross@suse.com>
Cc: linux-pci@vger.kernel.org
Cc: x86@kernel.org
Signed-off-by: Luis R. Rodriguez <mcgrof@suse.com>
---
 Documentation/PCI/MSI-HOWTO.txt | 5 +----
 arch/x86/Kconfig                | 6 +-----
 drivers/pci/Kconfig             | 2 ++
 3 files changed, 4 insertions(+), 9 deletions(-)

diff --git a/Documentation/PCI/MSI-HOWTO.txt b/Documentation/PCI/MSI-HOWTO.txt
index 0d920d5..7306d2c 100644
--- a/Documentation/PCI/MSI-HOWTO.txt
+++ b/Documentation/PCI/MSI-HOWTO.txt
@@ -71,10 +71,7 @@ will simply fail and the device will continue to use pin-based interrupts.
 4.1 Include kernel support for MSIs
 
 To support MSI or MSI-X, the kernel must be built with the CONFIG_PCI_MSI
-option enabled.  This option is only available on some architectures,
-and it may depend on some other options also being set.  For example,
-on x86, you must also enable X86_UP_APIC or SMP in order to see the
-CONFIG_PCI_MSI option.
+option enabled.  This option is only available on some architectures.
 
 4.2 Using MSI
 
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index d038475..0ecdeb1 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -902,10 +902,6 @@ config X86_UP_APIC
 	  performance counters), and the NMI watchdog which detects hard
 	  lockups.
 
-config X86_UP_APIC_MSI
-	def_bool y
-	select X86_UP_APIC if X86_32 && !SMP && !X86_32_NON_STANDARD && PCI_MSI
-
 config X86_UP_IOAPIC
 	bool "IO-APIC support on uniprocessors"
 	depends on X86_UP_APIC
@@ -924,7 +920,7 @@ config X86_LOCAL_APIC
 	select GENERIC_IRQ_LEGACY_ALLOC_HWIRQ
 
 config X86_IO_APIC
-	def_bool X86_64 || SMP || X86_32_NON_STANDARD || X86_UP_IOAPIC
+	def_bool X86_64 || SMP || X86_32_NON_STANDARD || X86_UP_IOAPIC || PCI_MSI
 	depends on X86_LOCAL_APIC
 	select IRQ_DOMAIN
 
diff --git a/drivers/pci/Kconfig b/drivers/pci/Kconfig
index 7a8f1c5..01370c0 100644
--- a/drivers/pci/Kconfig
+++ b/drivers/pci/Kconfig
@@ -5,6 +5,8 @@ config PCI_MSI
 	bool "Message Signaled Interrupts (MSI and MSI-X)"
 	depends on PCI
 	select GENERIC_MSI_IRQ
+	select X86_LOCAL_APIC if X86_32 && !SMP && !X86_32_NON_STANDARD
+	select X86_IO_APIC if X86_32 && !SMP && !X86_32_NON_STANDARD
 	help
 	   This allows device drivers to enable MSI (Message Signaled
 	   Interrupts).  Message Signaled Interrupts enable a device to
-- 
2.3.2.209.gd67f9d5.dirty

Re: [RFC v1 1/2] x86: kconfig: simplify enabling [IO]APIC for PCI_MSI

[David Rientjes]

On Tue, 10 Mar 2015, Luis R. Rodriguez wrote:

> diff --git a/Documentation/PCI/MSI-HOWTO.txt b/Documentation/PCI/MSI-HOWTO.txt
> index 0d920d5..7306d2c 100644
> --- a/Documentation/PCI/MSI-HOWTO.txt
> +++ b/Documentation/PCI/MSI-HOWTO.txt
> @@ -71,10 +71,7 @@ will simply fail and the device will continue to use pin-based interrupts.
>  4.1 Include kernel support for MSIs
>  
>  To support MSI or MSI-X, the kernel must be built with the CONFIG_PCI_MSI
> -option enabled.  This option is only available on some architectures,
> -and it may depend on some other options also being set.  For example,
> -on x86, you must also enable X86_UP_APIC or SMP in order to see the
> -CONFIG_PCI_MSI option.
> +option enabled.  This option is only available on some architectures.
>  
>  4.2 Using MSI
>  
> diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
> index d038475..0ecdeb1 100644
> --- a/arch/x86/Kconfig
> +++ b/arch/x86/Kconfig
> @@ -902,10 +902,6 @@ config X86_UP_APIC
>  	  performance counters), and the NMI watchdog which detects hard
>  	  lockups.
>  
> -config X86_UP_APIC_MSI
> -	def_bool y
> -	select X86_UP_APIC if X86_32 && !SMP && !X86_32_NON_STANDARD && PCI_MSI
> -
>  config X86_UP_IOAPIC
>  	bool "IO-APIC support on uniprocessors"
>  	depends on X86_UP_APIC
> @@ -924,7 +920,7 @@ config X86_LOCAL_APIC
>  	select GENERIC_IRQ_LEGACY_ALLOC_HWIRQ
>  
>  config X86_IO_APIC
> -	def_bool X86_64 || SMP || X86_32_NON_STANDARD || X86_UP_IOAPIC
> +	def_bool X86_64 || SMP || X86_32_NON_STANDARD || X86_UP_IOAPIC || PCI_MSI
>  	depends on X86_LOCAL_APIC
>  	select IRQ_DOMAIN
>  
> diff --git a/drivers/pci/Kconfig b/drivers/pci/Kconfig
> index 7a8f1c5..01370c0 100644
> --- a/drivers/pci/Kconfig
> +++ b/drivers/pci/Kconfig
> @@ -5,6 +5,8 @@ config PCI_MSI
>  	bool "Message Signaled Interrupts (MSI and MSI-X)"
>  	depends on PCI
>  	select GENERIC_MSI_IRQ
> +	select X86_LOCAL_APIC if X86_32 && !SMP && !X86_32_NON_STANDARD
> +	select X86_IO_APIC if X86_32 && !SMP && !X86_32_NON_STANDARD
>  	help
>  	   This allows device drivers to enable MSI (Message Signaled
>  	   Interrupts).  Message Signaled Interrupts enable a device to

This conflicts with commit 50849eefea3b ("x86/Kconfig: Simplify 
X86_UP_APIC handling"), which has already been merged for the 4.0 release 
cycle, and I think I prefer the simplification done there over the 
additional "select"s for PCI_MSI.

Re: [RFC v1 1/2] x86: kconfig: simplify enabling [IO]APIC for PCI_MSI

[Luis R. Rodriguez]

On Tue, Mar 10, 2015 at 05:49:10PM -0700, David Rientjes wrote:
> On Tue, 10 Mar 2015, Luis R. Rodriguez wrote:
> 
> > diff --git a/Documentation/PCI/MSI-HOWTO.txt b/Documentation/PCI/MSI-HOWTO.txt
> > index 0d920d5..7306d2c 100644
> > --- a/Documentation/PCI/MSI-HOWTO.txt
> > +++ b/Documentation/PCI/MSI-HOWTO.txt
> > @@ -71,10 +71,7 @@ will simply fail and the device will continue to use pin-based interrupts.
> >  4.1 Include kernel support for MSIs
> >  
> >  To support MSI or MSI-X, the kernel must be built with the CONFIG_PCI_MSI
> > -option enabled.  This option is only available on some architectures,
> > -and it may depend on some other options also being set.  For example,
> > -on x86, you must also enable X86_UP_APIC or SMP in order to see the
> > -CONFIG_PCI_MSI option.
> > +option enabled.  This option is only available on some architectures.
> >  
> >  4.2 Using MSI
> >  
> > diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
> > index d038475..0ecdeb1 100644
> > --- a/arch/x86/Kconfig
> > +++ b/arch/x86/Kconfig
> > @@ -902,10 +902,6 @@ config X86_UP_APIC
> >  	  performance counters), and the NMI watchdog which detects hard
> >  	  lockups.
> >  
> > -config X86_UP_APIC_MSI
> > -	def_bool y
> > -	select X86_UP_APIC if X86_32 && !SMP && !X86_32_NON_STANDARD && PCI_MSI
> > -
> >  config X86_UP_IOAPIC
> >  	bool "IO-APIC support on uniprocessors"
> >  	depends on X86_UP_APIC
> > @@ -924,7 +920,7 @@ config X86_LOCAL_APIC
> >  	select GENERIC_IRQ_LEGACY_ALLOC_HWIRQ
> >  
> >  config X86_IO_APIC
> > -	def_bool X86_64 || SMP || X86_32_NON_STANDARD || X86_UP_IOAPIC
> > +	def_bool X86_64 || SMP || X86_32_NON_STANDARD || X86_UP_IOAPIC || PCI_MSI
> >  	depends on X86_LOCAL_APIC
> >  	select IRQ_DOMAIN
> >  
> > diff --git a/drivers/pci/Kconfig b/drivers/pci/Kconfig
> > index 7a8f1c5..01370c0 100644
> > --- a/drivers/pci/Kconfig
> > +++ b/drivers/pci/Kconfig
> > @@ -5,6 +5,8 @@ config PCI_MSI
> >  	bool "Message Signaled Interrupts (MSI and MSI-X)"
> >  	depends on PCI
> >  	select GENERIC_MSI_IRQ
> > +	select X86_LOCAL_APIC if X86_32 && !SMP && !X86_32_NON_STANDARD
> > +	select X86_IO_APIC if X86_32 && !SMP && !X86_32_NON_STANDARD
> >  	help
> >  	   This allows device drivers to enable MSI (Message Signaled
> >  	   Interrupts).  Message Signaled Interrupts enable a device to
> 
> and I think I prefer the simplification done there over the 
> additional "select"s for PCI_MSI.

OK found a way to address that which makes it even easier to read.

 Luis

[RFC v1 2/2] x86: kconfig: remove UP [IO]APIC options

[Luis R. Rodriguez]

From: "Luis R. Rodriguez" <mcgrof@suse.com>

These were here mostly to help as wrappers for PCI_MSI
under certain UP 32-bit build scenarios, now that PCI_MSI
properly selects what it needs and when it needs it these
Kconfig options are now only enabling users reduce the size
of 32-bit kernels optionally. Since 32-bit kernels are not
in a huge state of flux and since there is no penalty for
adding APIC support other than a bit of size moving forward
we simply expect folks to enable APIC options on 32-bit systems
when a system needs it, as with PCI_MSI. This lets us remove
clutter from the x86 Kconfig file.

Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: Bryan O'Donoghue <pure.logic@nexus-software.ie>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Andy Shevchenko <andy.shevchenko@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Borislav Petkov <bp@suse.de>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Jan Beulich <JBeulich@suse.com>
Cc: Juergen Gross <jgross@suse.com>
Cc: linux-pci@vger.kernel.org
Cc: x86@kernel.org
Signed-off-by: Luis R. Rodriguez <mcgrof@suse.com>
---
 arch/x86/Kconfig | 29 ++---------------------------
 1 file changed, 2 insertions(+), 27 deletions(-)

diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 0ecdeb1..bfd4d66 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -889,38 +889,13 @@ config UP_LATE_INIT
        def_bool y
        depends on !SMP && X86_LOCAL_APIC
 
-config X86_UP_APIC
-	bool "Local APIC support on uniprocessors"
-	depends on X86_32 && !SMP && !X86_32_NON_STANDARD
-	---help---
-	  A local APIC (Advanced Programmable Interrupt Controller) is an
-	  integrated interrupt controller in the CPU. If you have a single-CPU
-	  system which has a processor with a local APIC, you can say Y here to
-	  enable and use it. If you say Y here even though your machine doesn't
-	  have a local APIC, then the kernel will still run with no slowdown at
-	  all. The local APIC supports CPU-generated self-interrupts (timer,
-	  performance counters), and the NMI watchdog which detects hard
-	  lockups.
-
-config X86_UP_IOAPIC
-	bool "IO-APIC support on uniprocessors"
-	depends on X86_UP_APIC
-	---help---
-	  An IO-APIC (I/O Advanced Programmable Interrupt Controller) is an
-	  SMP-capable replacement for PC-style interrupt controllers. Most
-	  SMP systems and many recent uniprocessor systems have one.
-
-	  If you have a single-CPU system with an IO-APIC, you can say Y here
-	  to use it. If you say Y here even though your machine doesn't have
-	  an IO-APIC, then the kernel will still run with no slowdown at all.
-
 config X86_LOCAL_APIC
 	def_bool y
-	depends on X86_64 || SMP || X86_32_NON_STANDARD || X86_UP_APIC || PCI_MSI
+	depends on X86_64 || SMP || X86_32_NON_STANDARD || PCI_MSI
 	select GENERIC_IRQ_LEGACY_ALLOC_HWIRQ
 
 config X86_IO_APIC
-	def_bool X86_64 || SMP || X86_32_NON_STANDARD || X86_UP_IOAPIC || PCI_MSI
+	def_bool X86_64 || SMP || X86_32_NON_STANDARD || PCI_MSI
 	depends on X86_LOCAL_APIC
 	select IRQ_DOMAIN
 
-- 
2.3.2.209.gd67f9d5.dirty

Re: [RFC v1 2/2] x86: kconfig: remove UP [IO]APIC options

[David Rientjes]

On Tue, 10 Mar 2015, Luis R. Rodriguez wrote:

> From: "Luis R. Rodriguez" <mcgrof@suse.com>
> 
> These were here mostly to help as wrappers for PCI_MSI
> under certain UP 32-bit build scenarios, now that PCI_MSI
> properly selects what it needs and when it needs it these
> Kconfig options are now only enabling users reduce the size
> of 32-bit kernels optionally. Since 32-bit kernels are not
> in a huge state of flux and since there is no penalty for
> adding APIC support other than a bit of size moving forward
> we simply expect folks to enable APIC options on 32-bit systems
> when a system needs it, as with PCI_MSI. This lets us remove
> clutter from the x86 Kconfig file.
> 

Check out Linus's latest tree at 
git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git and see 
how it looks after commit 50849eefea3b ("x86/Kconfig: Simplify X86_UP_APIC 
handling") and commit b1da1e715d4f ("x86/Kconfig: Simplify X86_IO_APIC 
dependencies").

Re: [RFC v1 2/2] x86: kconfig: remove UP [IO]APIC options

[Luis R. Rodriguez]

On Tue, Mar 10, 2015 at 05:58:27PM -0700, David Rientjes wrote:
> On Tue, 10 Mar 2015, Luis R. Rodriguez wrote:
> 
> > From: "Luis R. Rodriguez" <mcgrof@suse.com>
> > 
> > These were here mostly to help as wrappers for PCI_MSI
> > under certain UP 32-bit build scenarios, now that PCI_MSI
> > properly selects what it needs and when it needs it these
> > Kconfig options are now only enabling users reduce the size
> > of 32-bit kernels optionally. Since 32-bit kernels are not
> > in a huge state of flux and since there is no penalty for
> > adding APIC support other than a bit of size moving forward
> > we simply expect folks to enable APIC options on 32-bit systems
> > when a system needs it, as with PCI_MSI. This lets us remove
> > clutter from the x86 Kconfig file.
> > 
> 
> Check out Linus's latest tree at 
> git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git and see 
> how it looks after commit 50849eefea3b ("x86/Kconfig: Simplify X86_UP_APIC 
> handling") and commit b1da1e715d4f ("x86/Kconfig: Simplify X86_IO_APIC 
> dependencies").

OK thanks yeah I see that now, yeah this can be simplified further, will
send a follow up based on top of these two.

 Luis