Skip to contentSkip to conversation

hexagon: Handle HAS_IOPORT dependencies

PATCH Superseded Niklas Schnelle 3 participants 10 messages Apr 3, 2024 – Oct 2, 2024 Qualcomm hexagon architecture

v1 of 9 Superseded Latest is v9 →

lore
Share & tools
lore
#0 lore raw

cover hexagon: Handle HAS_IOPORT dependencies

Hi Brian,

This is a follow up in my ongoing effort of making inb()/outb() and similar I/O port accessors compile-time optional. Previously I sent this as a treewide series titled "treewide: Remove I/O port accessors for HAS_IOPORT=n" with the latest being its 5th version[0]. With a significant subset of patches merged I've changed over to per-subsystem series. These series are stand alone and should be merged via the relevant tree such that with all subsystems complete we can follow this up with the final patch that will make the I/O port accessors compile-time optional.

The current state of the full series with changes to the remaining subsystems and the aforementioned final patch can be found for your convenience on my git.kernel.org tree in the has_ioport_v6 branch[1] with signed tags. As for compile-time vs runtime see Linus' reply to my first attempt[2].

Thanks,
Niklas

[0] https://lore.kernel.org/all/[email protected]/
[1] https://git.kernel.org/pub/scm/linux/kernel/git/niks/linux.git/log/?h=has_ioport_v6
[2] https://lore.kernel.org/lkml/CAHk-=wg80je=K7madF4e7WrRNp37e3qh6y10Svhdc7O8SZ_-8g@mail.gmail.com/

Niklas Schnelle (1):
hexagon: Don't select GENERIC_IOMAP without HAS_IOPORT support

 arch/hexagon/Kconfig | 1 -
 1 file changed, 1 deletion(-)
Signature · 1 line
2.40.1
#1 lore raw

PATCH 1/1 hexagon: Don't select GENERIC_IOMAP without HAS_IOPORT support

In a future patch HAS_IOPORT=n will disable inb()/outb() and friends at compile time. As hexagon does not support I/O port access it also the GENERIC_IOMAP mechanism of dynamically choosing between I/O port and MMIO access doesn't work so don't select it.

---
 arch/hexagon/Kconfig | 1 -
 1 file changed, 1 deletion(-)
arch/hexagon/Kconfig +0 -1
index e233b5efa276..5ea1bf4b7d4f 100644
@@ -31,7 +31,6 @@ config HEXAGON
 	select HAVE_ARCH_TRACEHOOK
 	select NEED_SG_DMA_LENGTH
 	select NO_IOPORT_MAP
-	select GENERIC_IOMAP
 	select GENERIC_IOREMAP
 	select GENERIC_SMP_IDLE_THREAD
 	select STACKTRACE_SUPPORT
Signature · 1 line
2.40.1
Reviewed-by on v1 1/1
#2 lore raw

RE: [PATCH 1/1] hexagon: Don't select GENERIC_IOMAP without HAS_IOPORT support

Quoted 33 lines

-----Original Message-----
From: Niklas Schnelle <[email protected]>
Sent: Wednesday, April 3, 2024 8:25 AM
To: Brian Cain <[email protected]>
Cc: [email protected]; Arnd Bergmann <[email protected]>; Heiko Carstens <[email protected]>; [email protected]; Niklas Schnelle <[email protected]>
Subject: [PATCH 1/1] hexagon: Don't select GENERIC_IOMAP without HAS_IOPORT support

WARNING: This email originated from outside of Qualcomm. Please be wary of any links or attachments, and do not enable macros.

In a future patch HAS_IOPORT=n will disable inb()/outb() and friends at compile time. As hexagon does not support I/O port access it also the GENERIC_IOMAP mechanism of dynamically choosing between I/O port and
MMIO access doesn't work so don't select it.

---
 arch/hexagon/Kconfig | 1 -
 1 file changed, 1 deletion(-)
arch/hexagon/Kconfig +0 -0
index e233b5efa276..5ea1bf4b7d4f 100644

arch/hexagon/Kconfig config HEXAGON · line 36

Quoted 7 lines
arch/hexagon/Kconfig +0 -1
@@ -31,7 +31,6 @@ config HEXAGON
        select HAVE_ARCH_TRACEHOOK
        select NEED_SG_DMA_LENGTH
        select NO_IOPORT_MAP
-       select GENERIC_IOMAP
        select GENERIC_IOREMAP
        select GENERIC_SMP_IDLE_THREAD
        select STACKTRACE_SUPPORT
Quoted 2 lines

--
2.40.1

#3 lore raw

Re: [PATCH 1/1] hexagon: Don't select GENERIC_IOMAP without HAS_IOPORT support

Quoted 27 lines · Brian Cain
Quoted 27 lines

-----Original Message-----
From: Niklas Schnelle <[email protected]>
Sent: Wednesday, April 3, 2024 8:25 AM
To: Brian Cain <[email protected]>
Cc: [email protected]; Arnd Bergmann <[email protected]>; Heiko Carstens <[email protected]>; [email protected]; Niklas Schnelle <[email protected]>
Subject: [PATCH 1/1] hexagon: Don't select GENERIC_IOMAP without HAS_IOPORT support

WARNING: This email originated from outside of Qualcomm. Please be wary of any links or attachments, and do not enable macros.

In a future patch HAS_IOPORT=n will disable inb()/outb() and friends at compile time. As hexagon does not support I/O port access it also the GENERIC_IOMAP mechanism of dynamically choosing between I/O port and
MMIO access doesn't work so don't select it.

---
 arch/hexagon/Kconfig | 1 -
 1 file changed, 1 deletion(-)
Quoted 15 lines

arch/hexagon/Kconfig config HEXAGON · line 36

Quoted 12 lines
arch/hexagon/Kconfig +0 -1
index e233b5efa276..5ea1bf4b7d4f 100644
@@ -31,7 +31,6 @@ config HEXAGON
        select HAVE_ARCH_TRACEHOOK
        select NEED_SG_DMA_LENGTH
        select NO_IOPORT_MAP
-       select GENERIC_IOMAP
        select GENERIC_IOREMAP
        select GENERIC_SMP_IDLE_THREAD
        select STACKTRACE_SUPPORT

Thanks for the R-b, which tree should take this? If you have other hexagon changes for v6.10 I think it would be easiest to take this together with those. As mentioned in the cover letter the per-subsystem changes are all independent until HAS_IOPORT=n actually disables inb()/outb() which Arnd will take once all the subsystems are covered.

Thanks,
Niklas

#4 lore raw

RE: [PATCH 1/1] hexagon: Don't select GENERIC_IOMAP without HAS_IOPORT support

Quoted 61 lines

-----Original Message-----
From: Niklas Schnelle <[email protected]>
Sent: Wednesday, April 3, 2024 10:22 AM
To: Brian Cain <[email protected]>
Cc: [email protected]; Arnd Bergmann <[email protected]>; Heiko Carstens <[email protected]>; [email protected]
Subject: Re: [PATCH 1/1] hexagon: Don't select GENERIC_IOMAP without HAS_IOPORT support

WARNING: This email originated from outside of Qualcomm. Please be wary of any links or attachments, and do not enable macros.

Quoted 7 lines · Brian Cain
Quoted 7 lines

-----Original Message-----
From: Niklas Schnelle <[email protected]>
Sent: Wednesday, April 3, 2024 8:25 AM
To: Brian Cain <[email protected]>
Cc: [email protected]; Arnd Bergmann <[email protected]>; Heiko Carstens <[email protected]>; [email protected];

Niklas

Quoted 5 lines
Quoted 5 lines

Schnelle <[email protected]>
Subject: [PATCH 1/1] hexagon: Don't select GENERIC_IOMAP without HAS_IOPORT support

WARNING: This email originated from outside of Qualcomm. Please be

wary

Quoted 5 lines
Quoted 5 lines

of any links or attachments, and do not enable macros.

In a future patch HAS_IOPORT=n will disable inb()/outb() and friends at compile time. As hexagon does not support I/O port access it also the GENERIC_IOMAP mechanism of dynamically choosing between I/O

port

Quoted 10 lines
Quoted 10 lines

and
MMIO access doesn't work so don't select it.

---
 arch/hexagon/Kconfig | 1 -
 1 file changed, 1 deletion(-)
Quoted 15 lines

arch/hexagon/Kconfig config HEXAGON · line 36

Quoted 12 lines
arch/hexagon/Kconfig +0 -1
index e233b5efa276..5ea1bf4b7d4f 100644
@@ -31,7 +31,6 @@ config HEXAGON
        select HAVE_ARCH_TRACEHOOK
        select NEED_SG_DMA_LENGTH
        select NO_IOPORT_MAP
-       select GENERIC_IOMAP
        select GENERIC_IOREMAP
        select GENERIC_SMP_IDLE_THREAD
        select STACKTRACE_SUPPORT
Quoted 3 lines

Thanks for the R-b, which tree should take this? If you have other hexagon changes for v6.10 I think it would be easiest to take this together with those. As mentioned in the cover letter the per-subsystem

I don't have other changes planned for v6.10 but I'm happy to take the change in my tree.

Quoted 5 lines

changes are all independent until HAS_IOPORT=n actually disables inb()/outb() which Arnd will take once all the subsystems are covered.

Thanks,
Niklas

#5 lore raw

Re: [PATCH 1/1] hexagon: Don't select GENERIC_IOMAP without HAS_IOPORT support

Quoted 60 lines · Brian Cain
Quoted 60 lines

-----Original Message-----
From: Niklas Schnelle <[email protected]>
Sent: Wednesday, April 3, 2024 10:22 AM
To: Brian Cain <[email protected]>
Cc: [email protected]; Arnd Bergmann <[email protected]>; Heiko Carstens <[email protected]>; [email protected]
Subject: Re: [PATCH 1/1] hexagon: Don't select GENERIC_IOMAP without HAS_IOPORT support

WARNING: This email originated from outside of Qualcomm. Please be wary of any links or attachments, and do not enable macros.

Quoted 7 lines · Brian Cain
Quoted 7 lines

-----Original Message-----
From: Niklas Schnelle <[email protected]>
Sent: Wednesday, April 3, 2024 8:25 AM
To: Brian Cain <[email protected]>
Cc: [email protected]; Arnd Bergmann <[email protected]>; Heiko Carstens <[email protected]>; [email protected];

Niklas

Quoted 5 lines
Quoted 5 lines

Schnelle <[email protected]>
Subject: [PATCH 1/1] hexagon: Don't select GENERIC_IOMAP without HAS_IOPORT support

WARNING: This email originated from outside of Qualcomm. Please be

wary

Quoted 5 lines
Quoted 5 lines

of any links or attachments, and do not enable macros.

In a future patch HAS_IOPORT=n will disable inb()/outb() and friends at compile time. As hexagon does not support I/O port access it also the GENERIC_IOMAP mechanism of dynamically choosing between I/O

port

Quoted 25 lines
Quoted 15 lines

and
MMIO access doesn't work so don't select it.

---
 arch/hexagon/Kconfig | 1 -
 1 file changed, 1 deletion(-)
arch/hexagon/Kconfig +0 -0
index e233b5efa276..5ea1bf4b7d4f 100644

arch/hexagon/Kconfig config HEXAGON · line 36

Quoted 7 lines
arch/hexagon/Kconfig +0 -1
@@ -31,7 +31,6 @@ config HEXAGON
        select HAVE_ARCH_TRACEHOOK
        select NEED_SG_DMA_LENGTH
        select NO_IOPORT_MAP
-       select GENERIC_IOMAP
        select GENERIC_IOREMAP
        select GENERIC_SMP_IDLE_THREAD
        select STACKTRACE_SUPPORT
Quoted 11 lines
Quoted 3 lines

Thanks for the R-b, which tree should take this? If you have other hexagon changes for v6.10 I think it would be easiest to take this together with those. As mentioned in the cover letter the per-subsystem

I don't have other changes planned for v6.10 but I'm happy to take the change in my tree.

Quoted 5 lines

changes are all independent until HAS_IOPORT=n actually disables inb()/outb() which Arnd will take once all the subsystems are covered.

Thanks,
Niklas

Hi Brian,

With 2 more of my HAS_IOPORT patches having gone into v6.12-rc1 I just noticed that this one seems to have fallen into the cracks. This is probably my fault for not properly confirming that I'd be happy for you to take it. Do you want me to resend a new version rebased on v6.12- rc1?

Thanks,
Niklas

#6 lore raw

RE: [PATCH 1/1] hexagon: Don't select GENERIC_IOMAP without HAS_IOPORT support

Quoted 88 lines

-----Original Message-----
From: Niklas Schnelle <[email protected]>
Sent: Monday, September 30, 2024 10:51 AM
To: Brian Cain <[email protected]>
Cc: [email protected]; Arnd Bergmann <[email protected]>; Heiko Carstens <[email protected]>; [email protected]
Subject: Re: [PATCH 1/1] hexagon: Don't select GENERIC_IOMAP without HAS_IOPORT support

WARNING: This email originated from outside of Qualcomm. Please be wary of any links or attachments, and do not enable macros.

Quoted 60 lines · Brian Cain
Quoted 60 lines

-----Original Message-----
From: Niklas Schnelle <[email protected]>
Sent: Wednesday, April 3, 2024 10:22 AM
To: Brian Cain <[email protected]>
Cc: [email protected]; Arnd Bergmann <[email protected]>; Heiko Carstens <[email protected]>; [email protected]
Subject: Re: [PATCH 1/1] hexagon: Don't select GENERIC_IOMAP without HAS_IOPORT support

WARNING: This email originated from outside of Qualcomm. Please be wary of any links or attachments, and do not enable macros.

Quoted 7 lines · Brian Cain
Quoted 7 lines

-----Original Message-----
From: Niklas Schnelle <[email protected]>
Sent: Wednesday, April 3, 2024 8:25 AM
To: Brian Cain <[email protected]>
Cc: [email protected]; Arnd Bergmann <[email protected]>; Heiko Carstens <[email protected]>; [email protected];

Niklas

Quoted 5 lines
Quoted 5 lines

Schnelle <[email protected]>
Subject: [PATCH 1/1] hexagon: Don't select GENERIC_IOMAP without HAS_IOPORT support

WARNING: This email originated from outside of Qualcomm. Please be

wary

Quoted 5 lines
Quoted 5 lines

of any links or attachments, and do not enable macros.

In a future patch HAS_IOPORT=n will disable inb()/outb() and friends at compile time. As hexagon does not support I/O port access it also the GENERIC_IOMAP mechanism of dynamically choosing between I/O

port

Quoted 25 lines
Quoted 15 lines

and
MMIO access doesn't work so don't select it.

---
 arch/hexagon/Kconfig | 1 -
 1 file changed, 1 deletion(-)
arch/hexagon/Kconfig +0 -0
index e233b5efa276..5ea1bf4b7d4f 100644

arch/hexagon/Kconfig config HEXAGON · line 36

Quoted 7 lines
arch/hexagon/Kconfig +0 -1
@@ -31,7 +31,6 @@ config HEXAGON
        select HAVE_ARCH_TRACEHOOK
        select NEED_SG_DMA_LENGTH
        select NO_IOPORT_MAP
-       select GENERIC_IOMAP
        select GENERIC_IOREMAP
        select GENERIC_SMP_IDLE_THREAD
        select STACKTRACE_SUPPORT
Quoted 5 lines
Quoted 3 lines

Thanks for the R-b, which tree should take this? If you have other hexagon changes for v6.10 I think it would be easiest to take this together with those. As mentioned in the cover letter the per-subsystem

I don't have other changes planned for v6.10 but I'm happy to take the

change in my tree.

Quoted 6 lines
Quoted 6 lines

changes are all independent until HAS_IOPORT=n actually disables inb()/outb() which Arnd will take once all the subsystems are covered.

Thanks,
Niklas

Quoted 7 lines

Hi Brian,

With 2 more of my HAS_IOPORT patches having gone into v6.12-rc1 I just noticed that this one seems to have fallen into the cracks. This is probably my fault for not properly confirming that I'd be happy for you to take it. Do you want me to resend a new version rebased on v6.12- rc1?

Oh, sorry! Yes, if you could resend a rebased one I'll make sure I am able to land it.

-Brian

#7 lore raw

Re: [PATCH 1/1] hexagon: Don't select GENERIC_IOMAP without HAS_IOPORT support

Quoted 95 lines · Brian Cain
Quoted 87 lines

-----Original Message-----
From: Niklas Schnelle <[email protected]>
Sent: Monday, September 30, 2024 10:51 AM
To: Brian Cain <[email protected]>
Cc: [email protected]; Arnd Bergmann <[email protected]>; Heiko Carstens <[email protected]>; [email protected]
Subject: Re: [PATCH 1/1] hexagon: Don't select GENERIC_IOMAP without HAS_IOPORT support

WARNING: This email originated from outside of Qualcomm. Please be wary of any links or attachments, and do not enable macros.

Quoted 60 lines · Brian Cain
Quoted 60 lines

-----Original Message-----
From: Niklas Schnelle <[email protected]>
Sent: Wednesday, April 3, 2024 10:22 AM
To: Brian Cain <[email protected]>
Cc: [email protected]; Arnd Bergmann <[email protected]>; Heiko Carstens <[email protected]>; [email protected]
Subject: Re: [PATCH 1/1] hexagon: Don't select GENERIC_IOMAP without HAS_IOPORT support

WARNING: This email originated from outside of Qualcomm. Please be wary of any links or attachments, and do not enable macros.

Quoted 7 lines · Brian Cain
Quoted 7 lines

-----Original Message-----
From: Niklas Schnelle <[email protected]>
Sent: Wednesday, April 3, 2024 8:25 AM
To: Brian Cain <[email protected]>
Cc: [email protected]; Arnd Bergmann <[email protected]>; Heiko Carstens <[email protected]>; [email protected];

Niklas

Quoted 5 lines
Quoted 5 lines

Schnelle <[email protected]>
Subject: [PATCH 1/1] hexagon: Don't select GENERIC_IOMAP without HAS_IOPORT support

WARNING: This email originated from outside of Qualcomm. Please be

wary

Quoted 5 lines
Quoted 5 lines

of any links or attachments, and do not enable macros.

In a future patch HAS_IOPORT=n will disable inb()/outb() and friends at compile time. As hexagon does not support I/O port access it also the GENERIC_IOMAP mechanism of dynamically choosing between I/O

port

Quoted 25 lines
Quoted 15 lines

and
MMIO access doesn't work so don't select it.

---
 arch/hexagon/Kconfig | 1 -
 1 file changed, 1 deletion(-)
arch/hexagon/Kconfig +0 -0
index e233b5efa276..5ea1bf4b7d4f 100644

arch/hexagon/Kconfig config HEXAGON · line 36

Quoted 7 lines
arch/hexagon/Kconfig +0 -1
@@ -31,7 +31,6 @@ config HEXAGON
        select HAVE_ARCH_TRACEHOOK
        select NEED_SG_DMA_LENGTH
        select NO_IOPORT_MAP
-       select GENERIC_IOMAP
        select GENERIC_IOREMAP
        select GENERIC_SMP_IDLE_THREAD
        select STACKTRACE_SUPPORT
Quoted 5 lines
Quoted 3 lines

Thanks for the R-b, which tree should take this? If you have other hexagon changes for v6.10 I think it would be easiest to take this together with those. As mentioned in the cover letter the per-subsystem

I don't have other changes planned for v6.10 but I'm happy to take the

change in my tree.

Quoted 6 lines
Quoted 6 lines

changes are all independent until HAS_IOPORT=n actually disables inb()/outb() which Arnd will take once all the subsystems are covered.

Thanks,
Niklas

Quoted 7 lines

Hi Brian,

With 2 more of my HAS_IOPORT patches having gone into v6.12-rc1 I just noticed that this one seems to have fallen into the cracks. This is probably my fault for not properly confirming that I'd be happy for you to take it. Do you want me to resend a new version rebased on v6.12- rc1?

Quoted 3 lines

Oh, sorry! Yes, if you could resend a rebased one I'll make sure I am able to land it.

-Brian

Hi Brian,

I just tried with

b4 shazam '<[email protected]>'

and this patch actually still applies cleanly and matches the result I got rebasing the whole series[0]. So on second thought taking it from here directly might still be the easiest. That said I'll have to send out the other 4 outstanding patches required for the final inb() compile time switch, so if you prefer I can still resend too.

Thanks
Niklas

[0]
https://git.kernel.org/pub/scm/linux/kernel/git/niks/linux.git/log/?h=has_ioport

#8 lore raw

Re: [PATCH 1/1] hexagon: Don't select GENERIC_IOMAP without HAS_IOPORT support

Quoted 16 lines · Niklas Schnelle
Quoted 10 lines · Brian Cain
Quoted 6 lines

With 2 more of my HAS_IOPORT patches having gone into v6.12-rc1 I just noticed that this one seems to have fallen into the cracks. This is probably my fault for not properly confirming that I'd be happy for you to take it. Do you want me to resend a new version rebased on v6.12- rc1?

Oh, sorry! Yes, if you could resend a rebased one I'll make sure I am able to land it.

-Brian

Hi Brian,

I just tried with

Quoted 7 lines
   b4 shazam '<[email protected]>'

and this patch actually still applies cleanly and matches the result I got rebasing the whole series[0]. So on second thought taking it from here directly might still be the easiest. That said I'll have to send out the other 4 outstanding patches required for the final inb() compile time switch, so if you prefer I can still resend too.

Hi Niklas,

For simplicity I would suggest you keep all the remaining patches as a series now and I'll take them through the asm-generic tree in order to get them in more easily along with the final patch.

Arnd

#9 lore raw

Re: [PATCH 1/1] hexagon: Don't select GENERIC_IOMAP without HAS_IOPORT support

Quoted 25 lines · Arnd Bergmann
Quoted 16 lines · Niklas Schnelle
Quoted 10 lines · Brian Cain
Quoted 6 lines

With 2 more of my HAS_IOPORT patches having gone into v6.12-rc1 I just noticed that this one seems to have fallen into the cracks. This is probably my fault for not properly confirming that I'd be happy for you to take it. Do you want me to resend a new version rebased on v6.12- rc1?

Oh, sorry! Yes, if you could resend a rebased one I'll make sure I am able to land it.

-Brian

Hi Brian,

I just tried with

Quoted 7 lines
   b4 shazam '<[email protected]>'

and this patch actually still applies cleanly and matches the result I got rebasing the whole series[0]. So on second thought taking it from here directly might still be the easiest. That said I'll have to send out the other 4 outstanding patches required for the final inb() compile time switch, so if you prefer I can still resend too.

Quoted 7 lines

Hi Niklas,

For simplicity I would suggest you keep all the remaining patches as a series now and I'll take them through the asm-generic tree in order to get them in more easily along with the final patch.

Arnd

Hi Arnd,

Works for me. So, I think apart from this one that was simply missed and has an Ack the Bluetooth patch hasn't seen any feedback, for the drm one I have a few minor todos and then there is the tty/serial one.

So maybe I can just sent a version of all the remaining ones and add all the remaining maintainers for the whole series. I'll include a note that we plan on taking this through asm-generic.

Thanks,
Niklas