Re: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt platform
On 06/29/2012 11:18 AM, Li Yang-R58472 wrote: > > >> -Original Message- From: Wood Scott-B07421 Sent: Friday, >> June 29, 2012 11:02 AM To: Kumar Gala Cc: Jia Hongtao-B38951; Wood >> Scott-B07421; Li Yang-R58472; ag...@suse.de; >> linuxppc-dev@lists.ozlabs.org Subject: Re: [PATCH 0/3] powerpc/fsl: >> PCI refactoring and QEMU paravirt platform >> >> On 06/29/2012 10:57 AM, Kumar Gala wrote: >>> Pretty sure the boards have ISA, if you see the .dts has >>> references to >> 'ISA bridge' & 'i8259' PIC. >> >> OK. How about looking for an i8259 node as well? > > That could work, but looks hackish. Our proposal for adding a new > device tree property is a generic solution. Yes, all *new* boards should have an isa node. But we want to remain compatible with existing device trees. > The only problem is that > new kernels would work with old device trees. I think we can use > your solution for transitional period. And go for a well defined > device tree binding for this in long run. The "transitional period" is until we no longer care about these specific boards, or any out-of-tree derivatives. -Scott ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
RE: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt platform
> -Original Message- > From: Wood Scott-B07421 > Sent: Friday, June 29, 2012 11:02 AM > To: Kumar Gala > Cc: Jia Hongtao-B38951; Wood Scott-B07421; Li Yang-R58472; ag...@suse.de; > linuxppc-dev@lists.ozlabs.org > Subject: Re: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt > platform > > On 06/29/2012 10:57 AM, Kumar Gala wrote: > > > > On Jun 28, 2012, at 9:36 PM, Jia Hongtao-B38951 wrote: > > > >> > >> > >>> -Original Message- > >>> From: Wood Scott-B07421 > >>> Sent: Friday, June 29, 2012 12:31 AM > >>> To: Jia Hongtao-B38951 > >>> Cc: Wood Scott-B07421; ga...@kernel.crashing.org; Li Yang-R58472; > >>> ag...@suse.de; linuxppc-dev@lists.ozlabs.org > >>> Subject: Re: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU > >>> paravirt platform > >>> > >>> On 06/27/2012 11:06 PM, Jia Hongtao-B38951 wrote: > >>>> > >>>> > >>>>> -Original Message- > >>>>> From: Wood Scott-B07421 > >>>>> Sent: Thursday, June 28, 2012 7:49 AM > >>>>> To: ga...@kernel.crashing.org > >>>>> Cc: ag...@suse.de; linuxppc-dev@lists.ozlabs.org; Jia > >>>>> Hongtao-B38951 > >>>>> Subject: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU > >>>>> paravirt platform > >>>>> > >>>>> The QEMU stuff is related to the PCI refactoring because currently > >>>>> we have a hard time selecting a primary bus under QEMU, and also > >>> because > >>>>> the generic qemu e500 platform wants a full list of FSL PCI > >>> compatibles > >>>>> to check. > >>>>> > >>>> > >>>> It seems that not all primary bus has "isa" node like 8541 and 8555. > >>> > >>> Do those boards (it's the boards that matter, not chips...) have > >>> legacy ISA? If they do, and it's not in the device tree, then we > >>> should fix the device tree for consistency, but also retain some > >>> sort of hack to remain compatible with old device trees. > >>> > >>> A board can refrain from using the new common infrastructure if it > >>> has a good reason to. > >> > >> I'm not sure that MPC8541CDS (or 8555) has legacy ISA. I just checked > >> in kernel and dts which implies the board has primary bus and no "isa" > node. > >> I will find out the facts later. > > > > Pretty sure the boards have ISA, if you see the .dts has references to > 'ISA bridge' & 'i8259' PIC. > > OK. How about looking for an i8259 node as well? That could work, but looks hackish. Our proposal for adding a new device tree property is a generic solution. The only problem is that new kernels would work with old device trees. I think we can use your solution for transitional period. And go for a well defined device tree binding for this in long run. - Leo ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
RE: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt platform
> -Original Message- > From: Linuxppc-dev [mailto:linuxppc-dev-bounces+tie- > fei.zang=freescale@lists.ozlabs.org] On Behalf Of Kumar Gala > Sent: Friday, June 29, 2012 23:58 PM > To: Jia Hongtao-B38951 > Cc: Wood Scott-B07421; linuxppc-dev@lists.ozlabs.org; Li Yang-R58472; > ag...@suse.de > Subject: Re: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt > platform > > > On Jun 28, 2012, at 9:36 PM, Jia Hongtao-B38951 wrote: > > > > > > >> -Original Message- > >> From: Wood Scott-B07421 > >> Sent: Friday, June 29, 2012 12:31 AM > >> To: Jia Hongtao-B38951 > >> Cc: Wood Scott-B07421; ga...@kernel.crashing.org; Li Yang-R58472; > >> ag...@suse.de; linuxppc-dev@lists.ozlabs.org > >> Subject: Re: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt > >> platform > >> > >> On 06/27/2012 11:06 PM, Jia Hongtao-B38951 wrote: > >>> > >>> > >>>> -Original Message- > >>>> From: Wood Scott-B07421 > >>>> Sent: Thursday, June 28, 2012 7:49 AM > >>>> To: ga...@kernel.crashing.org > >>>> Cc: ag...@suse.de; linuxppc-dev@lists.ozlabs.org; Jia Hongtao-B38951 > >>>> Subject: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt > >>>> platform > >>>> > >>>> The QEMU stuff is related to the PCI refactoring because currently > >>>> we have a hard time selecting a primary bus under QEMU, and also > >> because > >>>> the generic qemu e500 platform wants a full list of FSL PCI > >> compatibles > >>>> to check. > >>>> > >>> > >>> It seems that not all primary bus has "isa" node like 8541 and 8555. > >> > >> Do those boards (it's the boards that matter, not chips...) have legacy > >> ISA? If they do, and it's not in the device tree, then we should fix > >> the device tree for consistency, but also retain some sort of hack to > >> remain compatible with old device trees. > >> > >> A board can refrain from using the new common infrastructure if it has a > >> good reason to. > > > > I'm not sure that MPC8541CDS (or 8555) has legacy ISA. I just checked in > > kernel and dts which implies the board has primary bus and no "isa" node. > > I will find out the facts later. > > Pretty sure the boards have ISA, if you see the .dts has references to 'ISA > bridge' & 'i8259' PIC. It is on the VIA south bridge. Roy ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
Re: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt platform
On Jun 29, 2012, at 11:01 AM, Scott Wood wrote: > On 06/29/2012 10:57 AM, Kumar Gala wrote: >> >> On Jun 28, 2012, at 9:36 PM, Jia Hongtao-B38951 wrote: >> >>> >>> >>>> -Original Message- >>>> From: Wood Scott-B07421 >>>> Sent: Friday, June 29, 2012 12:31 AM >>>> To: Jia Hongtao-B38951 >>>> Cc: Wood Scott-B07421; ga...@kernel.crashing.org; Li Yang-R58472; >>>> ag...@suse.de; linuxppc-dev@lists.ozlabs.org >>>> Subject: Re: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt >>>> platform >>>> >>>> On 06/27/2012 11:06 PM, Jia Hongtao-B38951 wrote: >>>>> >>>>> >>>>>> -Original Message- >>>>>> From: Wood Scott-B07421 >>>>>> Sent: Thursday, June 28, 2012 7:49 AM >>>>>> To: ga...@kernel.crashing.org >>>>>> Cc: ag...@suse.de; linuxppc-dev@lists.ozlabs.org; Jia Hongtao-B38951 >>>>>> Subject: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt >>>>>> platform >>>>>> >>>>>> The QEMU stuff is related to the PCI refactoring because currently >>>>>> we have a hard time selecting a primary bus under QEMU, and also >>>> because >>>>>> the generic qemu e500 platform wants a full list of FSL PCI >>>> compatibles >>>>>> to check. >>>>>> >>>>> >>>>> It seems that not all primary bus has "isa" node like 8541 and 8555. >>>> >>>> Do those boards (it's the boards that matter, not chips...) have legacy >>>> ISA? If they do, and it's not in the device tree, then we should fix >>>> the device tree for consistency, but also retain some sort of hack to >>>> remain compatible with old device trees. >>>> >>>> A board can refrain from using the new common infrastructure if it has a >>>> good reason to. >>> >>> I'm not sure that MPC8541CDS (or 8555) has legacy ISA. I just checked in >>> kernel and dts which implies the board has primary bus and no "isa" node. >>> I will find out the facts later. >> >> Pretty sure the boards have ISA, if you see the .dts has references to 'ISA >> bridge' & 'i8259' PIC. > > OK. How about looking for an i8259 node as well? > > -Scott Works for me, saves us having to waste time on figuring out how to update .dts for the boards. - k ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
Re: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt platform
On 06/29/2012 10:57 AM, Kumar Gala wrote: > > On Jun 28, 2012, at 9:36 PM, Jia Hongtao-B38951 wrote: > >> >> >>> -Original Message- >>> From: Wood Scott-B07421 >>> Sent: Friday, June 29, 2012 12:31 AM >>> To: Jia Hongtao-B38951 >>> Cc: Wood Scott-B07421; ga...@kernel.crashing.org; Li Yang-R58472; >>> ag...@suse.de; linuxppc-dev@lists.ozlabs.org >>> Subject: Re: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt >>> platform >>> >>> On 06/27/2012 11:06 PM, Jia Hongtao-B38951 wrote: >>>> >>>> >>>>> -Original Message- >>>>> From: Wood Scott-B07421 >>>>> Sent: Thursday, June 28, 2012 7:49 AM >>>>> To: ga...@kernel.crashing.org >>>>> Cc: ag...@suse.de; linuxppc-dev@lists.ozlabs.org; Jia Hongtao-B38951 >>>>> Subject: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt >>>>> platform >>>>> >>>>> The QEMU stuff is related to the PCI refactoring because currently >>>>> we have a hard time selecting a primary bus under QEMU, and also >>> because >>>>> the generic qemu e500 platform wants a full list of FSL PCI >>> compatibles >>>>> to check. >>>>> >>>> >>>> It seems that not all primary bus has "isa" node like 8541 and 8555. >>> >>> Do those boards (it's the boards that matter, not chips...) have legacy >>> ISA? If they do, and it's not in the device tree, then we should fix >>> the device tree for consistency, but also retain some sort of hack to >>> remain compatible with old device trees. >>> >>> A board can refrain from using the new common infrastructure if it has a >>> good reason to. >> >> I'm not sure that MPC8541CDS (or 8555) has legacy ISA. I just checked in >> kernel and dts which implies the board has primary bus and no "isa" node. >> I will find out the facts later. > > Pretty sure the boards have ISA, if you see the .dts has references to 'ISA > bridge' & 'i8259' PIC. OK. How about looking for an i8259 node as well? -Scott ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
Re: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt platform
On Jun 28, 2012, at 9:36 PM, Jia Hongtao-B38951 wrote: > > >> -Original Message- >> From: Wood Scott-B07421 >> Sent: Friday, June 29, 2012 12:31 AM >> To: Jia Hongtao-B38951 >> Cc: Wood Scott-B07421; ga...@kernel.crashing.org; Li Yang-R58472; >> ag...@suse.de; linuxppc-dev@lists.ozlabs.org >> Subject: Re: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt >> platform >> >> On 06/27/2012 11:06 PM, Jia Hongtao-B38951 wrote: >>> >>> >>>> -Original Message- >>>> From: Wood Scott-B07421 >>>> Sent: Thursday, June 28, 2012 7:49 AM >>>> To: ga...@kernel.crashing.org >>>> Cc: ag...@suse.de; linuxppc-dev@lists.ozlabs.org; Jia Hongtao-B38951 >>>> Subject: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt >>>> platform >>>> >>>> The QEMU stuff is related to the PCI refactoring because currently >>>> we have a hard time selecting a primary bus under QEMU, and also >> because >>>> the generic qemu e500 platform wants a full list of FSL PCI >> compatibles >>>> to check. >>>> >>> >>> It seems that not all primary bus has "isa" node like 8541 and 8555. >> >> Do those boards (it's the boards that matter, not chips...) have legacy >> ISA? If they do, and it's not in the device tree, then we should fix >> the device tree for consistency, but also retain some sort of hack to >> remain compatible with old device trees. >> >> A board can refrain from using the new common infrastructure if it has a >> good reason to. > > I'm not sure that MPC8541CDS (or 8555) has legacy ISA. I just checked in > kernel and dts which implies the board has primary bus and no "isa" node. > I will find out the facts later. Pretty sure the boards have ISA, if you see the .dts has references to 'ISA bridge' & 'i8259' PIC. - k ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
RE: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt platform
> -Original Message- > From: Wood Scott-B07421 > Sent: Friday, June 29, 2012 12:31 AM > To: Jia Hongtao-B38951 > Cc: Wood Scott-B07421; ga...@kernel.crashing.org; Li Yang-R58472; > ag...@suse.de; linuxppc-dev@lists.ozlabs.org > Subject: Re: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt > platform > > On 06/27/2012 11:06 PM, Jia Hongtao-B38951 wrote: > > > > > >> -Original Message- > >> From: Wood Scott-B07421 > >> Sent: Thursday, June 28, 2012 7:49 AM > >> To: ga...@kernel.crashing.org > >> Cc: ag...@suse.de; linuxppc-dev@lists.ozlabs.org; Jia Hongtao-B38951 > >> Subject: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt > >> platform > >> > >> The QEMU stuff is related to the PCI refactoring because currently > >> we have a hard time selecting a primary bus under QEMU, and also > because > >> the generic qemu e500 platform wants a full list of FSL PCI > compatibles > >> to check. > >> > > > > It seems that not all primary bus has "isa" node like 8541 and 8555. > > Do those boards (it's the boards that matter, not chips...) have legacy > ISA? If they do, and it's not in the device tree, then we should fix > the device tree for consistency, but also retain some sort of hack to > remain compatible with old device trees. > > A board can refrain from using the new common infrastructure if it has a > good reason to. I'm not sure that MPC8541CDS (or 8555) has legacy ISA. I just checked in kernel and dts which implies the board has primary bus and no "isa" node. I will find out the facts later. > > > Without PM support for pci controllers I totally agree with this > refactoring > > for pci init. But in linux mechanism PM ops should be registered to a > driver. > > Do you have any ideas to add PM support for pci controllers under this > patchset? > > This isn't meant to be instead of making it a platform device. It's > just meant to get us away from board-specific code and primary-bus > hardcoding now, rather than once we sort out all the issues with > platform devices. > > -Scott Thanks, I'll look into the issues. -Jia Hongtao. ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
Re: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt platform
On 06/27/2012 11:06 PM, Jia Hongtao-B38951 wrote: > > >> -Original Message- >> From: Wood Scott-B07421 >> Sent: Thursday, June 28, 2012 7:49 AM >> To: ga...@kernel.crashing.org >> Cc: ag...@suse.de; linuxppc-dev@lists.ozlabs.org; Jia Hongtao-B38951 >> Subject: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt >> platform >> >> The QEMU stuff is related to the PCI refactoring because currently >> we have a hard time selecting a primary bus under QEMU, and also because >> the generic qemu e500 platform wants a full list of FSL PCI compatibles >> to check. >> > > It seems that not all primary bus has "isa" node like 8541 and 8555. Do those boards (it's the boards that matter, not chips...) have legacy ISA? If they do, and it's not in the device tree, then we should fix the device tree for consistency, but also retain some sort of hack to remain compatible with old device trees. A board can refrain from using the new common infrastructure if it has a good reason to. > Without PM support for pci controllers I totally agree with this refactoring > for pci init. But in linux mechanism PM ops should be registered to a driver. > Do you have any ideas to add PM support for pci controllers under this > patchset? This isn't meant to be instead of making it a platform device. It's just meant to get us away from board-specific code and primary-bus hardcoding now, rather than once we sort out all the issues with platform devices. -Scott ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev
RE: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt platform
> -Original Message- > From: Wood Scott-B07421 > Sent: Thursday, June 28, 2012 7:49 AM > To: ga...@kernel.crashing.org > Cc: ag...@suse.de; linuxppc-dev@lists.ozlabs.org; Jia Hongtao-B38951 > Subject: [PATCH 0/3] powerpc/fsl: PCI refactoring and QEMU paravirt > platform > > The QEMU stuff is related to the PCI refactoring because currently > we have a hard time selecting a primary bus under QEMU, and also because > the generic qemu e500 platform wants a full list of FSL PCI compatibles > to check. > It seems that not all primary bus has "isa" node like 8541 and 8555. Without PM support for pci controllers I totally agree with this refactoring for pci init. But in linux mechanism PM ops should be registered to a driver. Do you have any ideas to add PM support for pci controllers under this patchset? Thanks. -Jia Hongtao. ___ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev