* Re: help requesting for mounting eldk tools on AMCC PPC 405EP Taihu kit
From: Wolfgang Denk @ 2006-06-18 20:49 UTC (permalink / raw)
To: jagadeesh kalisetty; +Cc: linuxppc-dev
In-Reply-To: <20060618152344.13300.qmail@webmail18.rediffmail.com>
In message <20060618152344.13300.qmail@webmail18.rediffmail.com> you wrote:
>
> but the kernel is failed to mount on /dev/nfs as below please can you
> provide necessary steps to fix the problem.
...
> => setenv setboot setenv bootargs root=/dev/pts nfsroot=192.168.2.18:/opt/eldk/ppc_4xx ip=192.168.1.62:::::eth0:off
What's /dev/pts? You mentioned you intended to use NFS as root file
system. Then I'd expect to see /dev/nfs here. [Note: 'n' != 'p']
> bootcmd=run setboot;bootm 0xfc000000 0xfc300000
You mentioned you intended to use NFS as root file system. Then why
do you use a ramdisk?
> ## Loading RAMDisk Image at fc300000 ...
> Image Name: Taihu Ramdisk
> Image Type: PowerPC Linux RAMDisk Image (gzip compressed)
> Data Size: 15447973 Bytes = 14.7 MB
Your ramdisk image is much bigger than the default maximum kernel
ramdisk size (4 MB). But you don't pass any ramdisk size information
with your boot arguments:
> Kernel command line: root=/dev/pts nfsroot=192.168.2.18:/opt/eldk/ppc_4xx ip=192.168.1.62:::::eth0:off
If you really want to use this ramdisk, you should set the root
device and the ramdisk size appropriately.
Best regards,
Wolfgang Denk
--
Software Engineering: Embedded and Realtime Systems, Embedded Linux
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd@denx.de
NOTE: The Most Fundamental Particles in This Product Are Held
Together by a "Gluing" Force About Which Little is Currently Known
and Whose Adhesive Power Can Therefore Not Be Permanently Guaranteed.
^ permalink raw reply
* Linux Audio Xilinx board
From: Prakash N @ 2006-06-19 4:49 UTC (permalink / raw)
To: linuxppc-embedded
[-- Attachment #1: Type: text/plain, Size: 14184 bytes --]
Hi all,
I'm developing an IP Phone which has to be optimized for cost. I thought of prototyping, the application on Xilinx Board (http://www.xilinx.com/univ/xupv2p.html (or) www.digilentinc.com/xupv2p). Xilinx had given the OS support thro' Montavista, VxWorks. The montavista Linux which is supported & "open source" is available with the kernel 2.4.26. I used www.eecs.wsu.edu/~jamie/research/LinuxPort/linux_port.htm as a reference for porting linux on xilinx board. Since I'm more into VLSI than embedded designs I dont know much about the later.
Now coming to problem, Xilinx had given BSPs for Ethernet, UART, Compact Flash etc , which I can replace it in the Montavista distribution and I can configure and compile the kernel for the same. What I mean here is the montavista linux distribution doesnt have the xilinx BSPs, which I need to put in their distribution, before configuration & compilation. But the Xilinx had not given the BSP for Audio Codec, which is available in the board.
But as per the work done by Jamie, http://www.eecs.wsu.edu/~jamie/research/designs/audio_test /Audio%20Test.doc
http://www.eecs.wsu.edu/~jamie/research/designs/audio_test/7.1i.zip
Coming to my application, Linphone (www.linphone.org), an open source
VoIP phone, I could compile and run it on FC4 Linux x86 computer. Also
I cross compiled the same on powerPC-405 platform using cross
compiling tools. Since audio support/device driver is not provided
under montavista (compatible to Xilinx board) I couldnt test the
audio, which is very important in phone kind of products.
I could see the ac97 Codec (LM4550 chip) in the board. Also I could
see ac97 related files in two places,
1) Under Xilinx supplied software
2) Under Montavista Linux distribution
Probably I may need to change the ac97 files in montavista with "some
modifications" in Xilinx provided files. Here "some modifications" is
the advice given by Jamie (transcript is given below), which she had
stopped replying after some time.
I\'d attached the
1.Xilinx provided BSPs which I replace in monta vista distribution.
(attached as temp.zip)
2.ac97 files provided by Xilinx (xac97_l.c & xac97_l.h)
3.ac97 related files by Montavista (ac97*.*) as given in
linuxppc_2_4_devel folder
The linphone is trying to check /dev/dsp for audio i/o and I just need
OSS rather ALSA, as later is not supported in Linux 2.4.
Now if you can say how can I make a driver / adapter in linux for
Linphone, so that it can access the xilinx board specific Audio Codec.
------------------------------------------------------------------------------------------------------------------------
Transcripts of mails with Jamie
------------------------------------------------------------------------------------------------------------------------Prakash,
- Hide quoted text -
",1] ); //-->I could able to test the Mic (Mike, Speaker input) & speaker (speaker output), as a stand alone application (not on linux) running on board.
Coming to my application, Linphone (www.linphone.org), an open source VoIP phone, I could compile and run it on FC4 Linux x86 computer. Also I cross compiled the same on powerPC-405 platform using cross compiling tools. Since audio support/device driver is not provided under montavista (compatible to Xilinx board) I couldnt test the audio, which is very important in phone kind of products.
I could see the ac97 Codec (LM4550 chip) in the board. Also I could see ac97 related files in two places,
1) Under Xilinx supplied software
2) Under Montavista Linux distribution
Probably I may need to change the ac97 files in montavista with "some modifications" in Xilinx provided files. Here "some modifications" is the advice given by Jamie (transcript is given below), which she had stopped replying after some time.
The linphone is trying to check /dev/dsp for audio i/o and I just need OSS rather ALSA, as later is not supported in Linux 2.4.
Now if you can say how can I make a driver / adapter in linux for Linphone, so that it can access the xilinx board specific Audio Codec.
Also I want to know whether there is Linux 2.6 version available for xilinx board, which does have the BSPs for the same.
------------------------------------------------------------------------------------------
Transcripts of mails with Jamie
------------------------------ ------------------------------------------------------------
Prakash,
Sorry about the delay. I was out of town for a while.
If you track down Xilinx device driver functions, you\'ll
see that they use their own IO functions. These functions
are specific to Xilinx products, and so they can\'t be used directly
under Linux. The memory mapped IO driver I wrote can be used to
replace these functions. You basically create a character device on
your own, so that the device can be read/write in Linux. I assume you
have a good
programming knowledge, so just look into the AC97
device driver, and track down to the lowest level IO.
>From there, use the Linux IO driver to replace these functions. I hope
this helps.
-Jamie
----- Original Message -----
From: "Prakash N" <nprakash@cedt.iisc.ernet.in>
To: "Jamie Lin" <jamiehl@mail.wsu.edu>
Sent: Thursday, May 04, 2006 11:16 PM
Subject: RE: XILINX board-Audio Codec-Linux driver support
> Jamie,
> Could you please brief me how can I proceed in developing the
> driver/adapter for linux-xilinx combination. Just give me some steps
> for this audio support. Like, just outline what I need to do, then I
> can proceed. I feel I may have to modify/add xilinx hardware related
> stuff to AC97 general linux drivers.
> Please guide me in this.
> Sincerely,
> Prakash
>
> -----Original Message-----
> From: Jamie Lin [mailto:jamiehl@mail.wsu.edu]
> Sent: Monday, May 01, 2006 10:25 PM
> To: Prakash N
> Subject: Re: XILINX board-Audio Codec-Linux driver support
>
>
> No, you can\'t just use these files under Linux. Xilinx has its way of
",1] ); //-->
Sorry about the delay. I was out of town for a while.
If you track down Xilinx device driver functions, you'll
see that they use their own IO functions. These functions
are specific to Xilinx products, and so they can't be used directly
under Linux. The memory mapped IO driver I wrote can be used to
replace these functions. You basically create a character device on
your own, so that the device can be read/write in Linux. I assume you
have a good
programming knowledge, so just look into the AC97
device driver, and track down to the lowest level IO.
>From there, use the Linux IO driver to replace these functions. I hope
this helps.
-Jamie
----- Original Message -----
From: "Prakash N" <nprakash@cedt.iisc.ernet.in>
To: "Jamie Lin" <jamiehl@mail.wsu.edu>
Sent: Thursday, May 04, 2006 11:16 PM
Subject: RE: XILINX board-Audio Codec-Linux driver support
> Jamie,
> Could you please brief me how can I proceed in developing the
> driver/adapter for linux-xilinx combination. Just give me some steps
> for this audio support. Like, just outline what I need to do, then I
> can proceed. I feel I may have to modify/add xilinx hardware related
> stuff to AC97 general linux drivers.
> Please guide me in this.
> Sincerely,
> Prakash
>
> -----Original Message-----
> From: Jamie Lin [mailto:jamiehl@mail.wsu.edu]
> Sent: Monday, May 01, 2006 10:25 PM
> To: Prakash N
> Subject: Re: XILINX board-Audio Codec-Linux driver support
>
>
> No, you can't just use these files under Linux. Xilinx has its way of
> be used to make Xilinx drivers. All drivers in EDK only work under
> specific Xilinx platforms. That\'s why when we want to run Ethernet in
> Linux, we need to copy BSP to Linux, so that it knows a specific type
> of Ethernet needs to be made. We only replace files that comes with
> BSP in Linux. The rest of driver files in Linux are used to adapt the
> device. If mv_linux BSP doesn\'t generate the device drivers for some
> of your devices in the base system, it means that device is not
> supported under Linux. To conclude, you need to write a custom AC97
> adapter in order to make it work in Linux.
>
> -Jamie
>
>
> ----- Original Message -----
> From: "Prakash N" <nprakash@cedt.iisc.ernet.in>
> To: "Jamie Lin" <jamiehl@mail.wsu.edu>
> Sent: Sunday, April 30, 2006 11:58 PM
> Subject: RE: XILINX board-Audio Codec-Linux driver support
>
>
> > Jamie,
> > Iam confused with bsp folders. I just updated EDK with this,
> > http://www.xilinx.com/univ/XUPV2P/lib/lib_rev_1_1.zip
> > In that I could see EDK\\sw\\XilinxProcessorIPLib\\drivers\\ac97_v2_00_a
> > having data and src dirs which contains,
> > Data --> ac97_v2_1_0.tcl, ac97_v2_1_0.mdd
> > Src --> xac97_l.h ,xac97_l.c , Makefile
> > Is it not possible to use these drivers, by copying and pasting in
> > EDK\\sw\\ThirdParty\\bsp\\linux_mvl31_v1_01_a\\drivers
",1] ); //--> > making device drivers to adapt their platforms. The Makefiles can only
> be used to make Xilinx drivers. All drivers in EDK only work under
> specific Xilinx platforms. That's why when we want to run Ethernet in
> Linux, we need to copy BSP to Linux, so that it knows a specific type
> of Ethernet needs to be made. We only replace files that comes with
> BSP in Linux. The rest of driver files in Linux are used to adapt the
> device. If mv_linux BSP doesn't generate the device drivers for some
> of your devices in the base system, it means that device is not
> supported under Linux. To conclude, you need to write a custom AC97
> adapter in order to make it work in Linux.
>
> -Jamie
>
>
> ----- Original Message -----
> From: "Prakash N" <nprakash@cedt.iisc.ernet.in>
> To: "Jamie Lin" <jamiehl@mail.wsu.edu>
> Sent: Sunday, April 30, 2006 11:58 PM
> Subject: RE: XILINX board-Audio Codec-Linux driver support
>
>
> > Jamie,
> > Iam confused with bsp folders. I just updated EDK with this,
> > http://www.xilinx.com/univ/XUPV2P/lib/lib_rev_1_1.zip
> > In that I could see EDK\sw\XilinxProcessorIPLib\drivers\ac97_v2_00_a
> > having data and src dirs which contains,
> > Data --> ac97_v2_1_0.tcl, ac97_v2_1_0.mdd
> > Src --> xac97_l.h ,xac97_l.c , Makefile
> > Is it not possible to use these drivers, by copying and pasting in
> > EDK\sw\ThirdParty\bsp\linux_mvl31_v1_01_a\drivers
> > Please clarify,
> > Prakash
> >
> >
> > -----Original Message-----
> > From: Jamie Lin [mailto:jamiehl@mail.wsu.edu]
> > Sent: Monday, May 01, 2006 9:49 AM
> > To: Prakash N
> > Subject: Re: XILINX board-Audio Codec-Linux driver support
> >
> >
> > Prakash,
> >
> > The BSP only offers a limit number of necessary drivers. You\'ll have
> > to write custom drivers on your own if they\'re not included. One thing
>
> > you can do is to use the sample IO driver I\'ve written to replace
> > low
> > level IOs in the Xilinx AC97 codec driver. This will take you some
> > time, but is a way to adapt AC97 codec in Linux that I can think of.
> > Other than this approach, I haven\'t seen any ready to use driver for
> > this codec.
> >
> > -Jamie
> >
> >
> > ----- Original Message -----
> > From: <nprakash@cedt.iisc.ernet.in>
> > To: <jamiehl@mail.wsu.edu>
> > Sent: Friday, April 28, 2006 5:23 AM
> > Subject: XILINX board-Audio Codec-Linux driver support
> >
> >
> > > Jamie Lin,
> > > I could follow your doc and successfully ported the Linux in the
> > > board. I
> > just
> > > want to know how can I add the audio codec drivers in linux. I
> > > mean,
>
> > > I
> > couldnt
> > > see the Audio Codec in "Connected Peripherals" in Library & OS
",1] ); //--> > > and run libgen .....
> > Please clarify,
> > Prakash
> >
> >
> > -----Original Message-----
> > From: Jamie Lin [mailto:jamiehl@mail.wsu.edu]
> > Sent: Monday, May 01, 2006 9:49 AM
> > To: Prakash N
> > Subject: Re: XILINX board-Audio Codec-Linux driver support
> >
> >
> > Prakash,
> >
> > The BSP only offers a limit number of necessary drivers. You'll have
> > to write custom drivers on your own if they're not included. One thing
>
> > you can do is to use the sample IO driver I've written to replace
> > low
> > level IOs in the Xilinx AC97 codec driver. This will take you some
> > time, but is a way to adapt AC97 codec in Linux that I can think of.
> > Other than this approach, I haven't seen any ready to use driver for
> > this codec.
> >
> > -Jamie
> >
> >
> > ----- Original Message -----
> > From: <nprakash@cedt.iisc.ernet.in>
> > To: <jamiehl@mail.wsu.edu>
> > Sent: Friday, April 28, 2006 5:23 AM
> > Subject: XILINX board-Audio Codec-Linux driver support
> >
> >
> > > Jamie Lin,
> > > I could follow your doc and successfully ported the Linux in the
> > > board. I
> > just
> > > want to know how can I add the audio codec drivers in linux. I
> > > mean,
>
> > > I
> > couldnt
> > > see the Audio Codec in "Connected Peripherals" in Library & OS
>
> > > see its there
> > with
> > > all version, base address... defined. How can I add this in that
> > > so
> > > that
> > thro\'
> > > libgen I can add drivers for that also.
> > > Regards
> > > Prakash
------------------------------------------------------------------------------------------------------------------------
I feel I\'d given whatever possible details, still If something is
missing, Please ask the same.
Prakash
",1] ); D(["mb","
--
N.Prakash
Bangalore
+919448787632
",0] ); D(["ma",[1,"2 attachments — Scanning for viruses...temp.zip
294K linuxaudio.zip
22K ","10bd14d0c40a4b1e"] ] ); //--> > > > parameters (software platform settings). But in the hardware I could
>
> > > see its there
> > with
> > > all version, base address... defined. How can I add this in that
> > > so
> > > that
> > thro'
> > > libgen I can add drivers for that also.
> > > Regards
> > > Prakash
---------------------------------------------------------------------------------------------------
I feel I'd given whatever possible details, still If something is
missing, Please ask the same.
N.Prakash
ME Final Year- Microelectronics
CEDT-IISc
Bangalore
India
---------------------------------
Yahoo! India Answers: Share what you know. Learn something new Click here
Send free SMS to your Friends on Mobile from your Yahoo! Messenger Download now
[-- Attachment #2: Type: text/html, Size: 21498 bytes --]
^ permalink raw reply
* RE: Using bestcomm in an external module (MPC5200B to be exact)
From: Trueskew @ 2006-06-19 7:28 UTC (permalink / raw)
To: linuxppc-embedded
I should've let you all know, I'm more of a 2.4 person than a 2.6 person.
So I read what I could on 2.6 difference with 2.4, and it looked like my
easiest choice was to modify the bestcomm code to export a couple tables and
its API functions. I found that kallsyms has almost all of the functions I
needed to resolve (with 'T' next to them), but they were still not working.
I added EXPORT_MODULE(TaskBDReset) and similar declarations in the bestcomm
api code, but that still didn't do the trick for me. So what is the trick?
I think I'm on the right track, I can insmod a module that my driver needs,
see kallsyms get updated, and access the functions directly in my driver. I
hope I'm close to doing something similar with the bestcomm functions, can
anyone offer me some advice on what else I need to do?
Thanks.
________________________________
From: Trueskew [mailto:trueskew@gmail.com]
Sent: Friday, June 16, 2006 9:47 PM
To: 'linuxppc-embedded@ozlabs.org'
Subject: Using bestcomm in an external module (MPC5200B to be exact)
I'm writing a driver for a TI codec attached to the MPC5200B PSC3 codec
lines on a Lite5200b. It's an external module, and I'd like to use Bestcomm
with it. I've already tried a bunch of horrible things to get it to work.
I could say it's close, but by close I mean the FEC still runs, my stuff
doesn't.
1. Can the Bestcomm I/F be used by an external module, or is the module
required to be built in if built-in drivers use Bestcomm? TasksInitAPI (no
so bad) and TasksLoadImage (yeah, bad) aren't really supposed to be called
more than once from what I've read, and the kernel loads them for ethernet
and ATA (no disk in my system if that matters).
2. If it can be used by an external module, how would I do it? I've tried
compiling the source into my module and replicating the initialization the
kernel does in addition to my own stuff. It's not too bad, though I don't
get any interrupts yet.
The LTIB from Freescale includes AIC23 and AC97 drivers that use Bestcomm.
I built my kernel with them set as modules, and they had the same unresolved
problems that my driver has on insmod, so I'm pessimistic. If someone could
put me out of my misery one way or another, I'd appreciate it.
Thanks.
Sal
^ permalink raw reply
* could linux support to configure the PLB-PCI bridge of PPC440 into adapter mode
From: 陈 云 @ 2006-06-19 7:57 UTC (permalink / raw)
To: linuxppc-embedded
hi,now I want to use PPC440GX as PCI adapter card bridge. according to user
manual, the PLB-PCI bridge should be configured to Adapter mode. after
check some code of PCI of linux, i find the linux only support configuring
the PLB-PCI bridge into HOST mode. Is my understanding correct ? if the
normal version of linux don't support the adapter mode, could i found the
patch?
best regards and thank you.
chen
_________________________________________________________________
与联机的朋友进行交流，请使用 MSN Messenger: http://messenger.msn.com/cn
^ permalink raw reply
* Re: Using bestcomm in an external module (MPC5200B to be exact)
From: Andrey Volkov @ 2006-06-19 13:15 UTC (permalink / raw)
To: Trueskew; +Cc: linuxppc-embedded
In-Reply-To: <44965226.33b91fca.1abe.ffffd120@mx.gmail.com>
Trueskew wrote:
> I should've let you all know, I'm more of a 2.4 person than a 2.6 person.
> So I read what I could on 2.6 difference with 2.4, and it looked like my
> easiest choice was to modify the bestcomm code to export a couple tables and
> its API functions. I found that kallsyms has almost all of the functions I
> needed to resolve (with 'T' next to them), but they were still not working.
> I added EXPORT_MODULE(TaskBDReset) and similar declarations in the bestcomm
> api code, but that still didn't do the trick for me. So what is the trick?
> I think I'm on the right track, I can insmod a module that my driver needs,
> see kallsyms get updated, and access the functions directly in my driver. I
> hope I'm close to doing something similar with the bestcomm functions, can
> anyone offer me some advice on what else I need to do?
>
> Thanks.
Please, check Sylvain's patchset (http://www.246tNt.com/mpc52xx/)
_before_ you will do anithing. If be more shortly - Freescale's
BestcommAPI _deprecated_ and will _not_ included to 2.6.xxx.
--
Regards
Andrey Volkov
>
> ________________________________
>
> From: Trueskew [mailto:trueskew@gmail.com]
> Sent: Friday, June 16, 2006 9:47 PM
> To: 'linuxppc-embedded@ozlabs.org'
> Subject: Using bestcomm in an external module (MPC5200B to be exact)
>
>
> I'm writing a driver for a TI codec attached to the MPC5200B PSC3 codec
> lines on a Lite5200b. It's an external module, and I'd like to use Bestcomm
> with it. I've already tried a bunch of horrible things to get it to work.
> I could say it's close, but by close I mean the FEC still runs, my stuff
> doesn't.
>
> 1. Can the Bestcomm I/F be used by an external module, or is the module
> required to be built in if built-in drivers use Bestcomm? TasksInitAPI (no
> so bad) and TasksLoadImage (yeah, bad) aren't really supposed to be called
> more than once from what I've read, and the kernel loads them for ethernet
> and ATA (no disk in my system if that matters).
>
> 2. If it can be used by an external module, how would I do it? I've tried
> compiling the source into my module and replicating the initialization the
> kernel does in addition to my own stuff. It's not too bad, though I don't
> get any interrupts yet.
>
> The LTIB from Freescale includes AIC23 and AC97 drivers that use Bestcomm.
> I built my kernel with them set as modules, and they had the same unresolved
> problems that my driver has on insmod, so I'm pessimistic. If someone could
> put me out of my misery one way or another, I'd appreciate it.
>
> Thanks.
> Sal
>
> _______________________________________________
> Linuxppc-embedded mailing list
> Linuxppc-embedded@ozlabs.org
> https://ozlabs.org/mailman/listinfo/linuxppc-embedded
>
^ permalink raw reply
* Re: [PATCH 5/9 v3] Add the MPC8641 HPCN platform files.
From: Jon Loeliger @ 2006-06-19 14:04 UTC (permalink / raw)
To: Benjamin Herrenschmidt; +Cc: linuxppc-dev@ozlabs.org, Jon Loeliger
In-Reply-To: <1150600056.23600.152.camel@localhost.localdomain>
On Sat, 2006-06-17 at 22:07, Benjamin Herrenschmidt wrote:
>
> You don't have any way to freeze the timebase ? (No GPIO controlling
> TBEN ?) It's pretty nasty to rely on software sync,
I will investigate this option and follow up with a patch,
assuming that there is a HW solution for your suggestion.
> especially if it's supposed to be an example design ...
Who said that? ;-)
jdl
^ permalink raw reply
* Re: [PATCH 4/9 v3] Add starting of secondary 86xx CPUs.
From: Jon Loeliger @ 2006-06-19 14:06 UTC (permalink / raw)
To: Benjamin Herrenschmidt; +Cc: linuxppc-dev@ozlabs.org, Jon Loeliger
In-Reply-To: <1150599918.23600.150.camel@localhost.localdomain>
On Sat, 2006-06-17 at 22:05, Benjamin Herrenschmidt wrote:
> > +BEGIN_FTR_SECTION
> > + LOAD_BAT(4,r3,r4,r5)
> > + LOAD_BAT(5,r3,r4,r5)
> > + LOAD_BAT(6,r3,r4,r5)
> > + LOAD_BAT(7,r3,r4,r5)
> > +END_FTR_SECTION_IFSET(CPU_FTR_HAS_HIGH_BATS)
> > blr
>
> Looks good.
Excellent. Thanks.
jdl
^ permalink raw reply
* Re: [PATCH 5/9 v3] Add the MPC8641 HPCN platform files.
From: Jon Loeliger @ 2006-06-19 14:09 UTC (permalink / raw)
To: Benjamin Herrenschmidt; +Cc: linuxppc-dev@ozlabs.org, Jon Loeliger
In-Reply-To: <1150600470.23600.158.camel@localhost.localdomain>
On Sat, 2006-06-17 at 22:14, Benjamin Herrenschmidt wrote:
> I'm not too fan of the big table in the kernel however. I'd rather have
> it in the device-tree.
We are waiting for this release to get into the mainline,
and for your proposed rewrite to be accepted. We have folks
here who are on deck to totally modify this to accommodate
your proposal via the Device Tree file.
But we're not going to read your mind to get there. :-)
jdl
^ permalink raw reply
* Re: could linux support to configure the PLB-PCI bridge of PPC440 into adapter mode
From: Matthias Fuchs @ 2006-06-19 14:11 UTC (permalink / raw)
To: hilichen2000; +Cc: linuxppc-embedded
In-Reply-To: <BAY110-F4002B7619F83B2E70939DEC3860@phx.gbl>
Hi chen,
it's not easy to setup the PCI-to-PLB mapping in a generic way so that
is gets into the kernel. Also it might be too late when the adapter OS
configures the bridge. The bridge must be setup quickly after reset so
that the host OS can assign resources to the card. The bootloader of
your board might be a good place to setup of the adapter stuff. At least
number and sizes of the needed address spaces should be setup early.
The PLB address can be setup or modified later (e.g. from the target OS).
> hi,now I want to use PPC440GX as PCI adapter card bridge. according to user
> manual, the PLB-PCI bridge should be configured to Adapter mode. after
> check some code of PCI of linux, i find the linux only support configuring
> the PLB-PCI bridge into HOST mode. Is my understanding correct ? if the
> normal version of linux don't support the adapter mode, could i found the
> patch?
Probably not.
Matthias
^ permalink raw reply
* L3 cache on Apple dual processor (7450 r 2.1) question
From: Till Wimmer @ 2006-06-19 14:17 UTC (permalink / raw)
To: linuxppc-dev
Hello,
i tried to enable the L3 cache on my dual 800MHz processor board (taken
from Quicksilver). I'm running the 2.6.8-3-powerpc-smp kernel from
Debian.
I changed the code for in core99_init_caches() in
arch/ppc/platforms/pmac_smp.c because the cache wasn't enabled by
default:
if (cpu == 0){
_set_L3CR(0);
_set_L3CR(0x9F424340);
core99_l3_cache = _get_L3CR();
printk("CPU0: L3CR is %lx\n", core99_l3_cache);
} else {
printk("CPU%d: L3CR was %lx\n", cpu, _get_L3CR());
_set_L3CR(0);
_set_L3CR(core99_l3_cache);
printk("CPU%d: L3CR is %lx\n", cpu, _get_L3CR());
}
Now my questions are:
Is there anything else to do for getting the L3 cache working?
How can i check if the L3 is enabled? E.g. is there a benchmark program
which accounts for cache settings?
Thanx
Till
^ permalink raw reply
* Re: L3 cache on Apple dual processor (7450 r 2.1) question
From: Olof Johansson @ 2006-06-19 14:36 UTC (permalink / raw)
To: Till Wimmer; +Cc: linuxppc-dev
In-Reply-To: <1150726624.4711.16.camel@localhost.localdomain>
On Mon, Jun 19, 2006 at 04:17:04PM +0200, Till Wimmer wrote:
> Now my questions are:
> Is there anything else to do for getting the L3 cache working?
> How can i check if the L3 is enabled? E.g. is there a benchmark program
> which accounts for cache settings?
lmbench has testcases for memory latency testing.
http://www.bitmover.com/lmbench/
-Olof
^ permalink raw reply
* RE: Using bestcomm in an external module (MPC5200B to be exact)
From: Trueskew @ 2006-06-19 14:50 UTC (permalink / raw)
To: 'Andrey Volkov'; +Cc: linuxppc-embedded
In-Reply-To: <4496A363.8080802@varma-el.com>
Hi Andrey,
Thanks for the reply. I did notice that a few weeks ago and was waiting for
the patch state to be changed from "New" to "Accepted". Can you give me
more details on how you found out bestcomm won't be included in 2.6? The
latest downloadable LTIB from Freescale for the Lite5200b includes bestcomm
in several places, and the mpc52xx fec driver uses it, along with the ATA
driver that Sylvain provided a patch for.
As for http://www.246tNt.com/mpc52xx/, it's been coming up "page not found"
for me for those same few weeks when I've tried it off an on, and
www.246tNt.com just has a Welcome line in it.
Sal
-----Original Message-----
From: Andrey Volkov [mailto:avolkov@varma-el.com]
Sent: Monday, June 19, 2006 6:15 AM
To: Trueskew
Cc: linuxppc-embedded@ozlabs.org
Subject: Re: Using bestcomm in an external module (MPC5200B to be exact)
Trueskew wrote:
> I should've let you all know, I'm more of a 2.4 person than a 2.6 person.
> So I read what I could on 2.6 difference with 2.4, and it looked like
> my easiest choice was to modify the bestcomm code to export a couple
> tables and its API functions. I found that kallsyms has almost all of
> the functions I needed to resolve (with 'T' next to them), but they were
still not working.
> I added EXPORT_MODULE(TaskBDReset) and similar declarations in the
> bestcomm api code, but that still didn't do the trick for me. So what is
the trick?
> I think I'm on the right track, I can insmod a module that my driver
> needs, see kallsyms get updated, and access the functions directly in
> my driver. I hope I'm close to doing something similar with the
> bestcomm functions, can anyone offer me some advice on what else I need to
do?
>
> Thanks.
Please, check Sylvain's patchset (http://www.246tNt.com/mpc52xx/) _before_
you will do anithing. If be more shortly - Freescale's BestcommAPI
_deprecated_ and will _not_ included to 2.6.xxx.
--
Regards
Andrey Volkov
>
> ________________________________
>
> From: Trueskew [mailto:trueskew@gmail.com]
> Sent: Friday, June 16, 2006 9:47 PM
> To: 'linuxppc-embedded@ozlabs.org'
> Subject: Using bestcomm in an external module (MPC5200B to be exact)
>
>
> I'm writing a driver for a TI codec attached to the MPC5200B PSC3
> codec lines on a Lite5200b. It's an external module, and I'd like to
> use Bestcomm with it. I've already tried a bunch of horrible things to
get it to work.
> I could say it's close, but by close I mean the FEC still runs, my
> stuff doesn't.
>
> 1. Can the Bestcomm I/F be used by an external module, or is the
> module required to be built in if built-in drivers use Bestcomm?
> TasksInitAPI (no so bad) and TasksLoadImage (yeah, bad) aren't really
> supposed to be called more than once from what I've read, and the
> kernel loads them for ethernet and ATA (no disk in my system if that
matters).
>
> 2. If it can be used by an external module, how would I do it? I've
> tried compiling the source into my module and replicating the
> initialization the kernel does in addition to my own stuff. It's not
> too bad, though I don't get any interrupts yet.
>
> The LTIB from Freescale includes AIC23 and AC97 drivers that use Bestcomm.
> I built my kernel with them set as modules, and they had the same
> unresolved problems that my driver has on insmod, so I'm pessimistic.
> If someone could put me out of my misery one way or another, I'd
appreciate it.
>
> Thanks.
> Sal
>
> _______________________________________________
> Linuxppc-embedded mailing list
> Linuxppc-embedded@ozlabs.org
> https://ozlabs.org/mailman/listinfo/linuxppc-embedded
>
^ permalink raw reply
* OFW node names
From: Doug Maxey @ 2006-06-19 17:19 UTC (permalink / raw)
To: linuxppc-dev
Howdy!
I'm looking at page 164 in the 1275-1994 at the "name" entry.
Does any platform known to use yaboot (or grub for that matter) use the
preferred encoding of the manufacturer portion of the name property?
How about the "ONNNNN" OUI from the RAC, or the uppercase VWXYZ company stock symbol?
My view has been limited to the IBM platform. According to the
examples, this is a recommended practice. I know IBM uses the
lowercase variant.
If anyone has an example, please point me at tarball of the device tree.
++doug
^ permalink raw reply
* [patch 00/20] cell: patches for 2.6.18
From: arnd @ 2006-06-19 18:33 UTC (permalink / raw)
To: paulus; +Cc: linuxppc-dev, cbe-oss-dev, linux-kernel
Here are the cell patches I promised to submit ages ago for
inclusion in 2.6.18. I refreshed them to apply on the
union of 2.6.17 and powerpc.git.
Arnd <><
--
^ permalink raw reply
* [patch 01/20] cell: add RAS support
From: arnd @ 2006-06-19 18:33 UTC (permalink / raw)
To: paulus; +Cc: linuxppc-dev, cbe-oss-dev, linux-kernel
In-Reply-To: <20060619183315.653672000@klappe.arndb.de>
From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This is a first version of support for the Cell BE "Reliability,
Availability and Serviceability" features.
It doesn't yet handle some of the RAS interrupts (the ones described in
iic_is/iic_irr), I'm still working on a proper way to expose these. They
are essentially a cascaded controller by themselves (sic !) though I may
just handle them locally to the iic driver. I need also to sync with
David Erb on the way he hooked in the performance monitor interrupt.
So that's all for 2.6.17 and I'll do more work on that with my rework of
the powerpc interrupt layer that I'm hacking on at the moment.
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Index: powerpc.git/arch/powerpc/kernel/head_64.S
===================================================================
--- powerpc.git.orig/arch/powerpc/kernel/head_64.S
+++ powerpc.git/arch/powerpc/kernel/head_64.S
@@ -316,6 +316,21 @@ label##_pSeries: \
mtspr SPRN_SPRG1,r13; /* save r13 */ \
EXCEPTION_PROLOG_PSERIES(PACA_EXGEN, label##_common)
+#define HSTD_EXCEPTION_PSERIES(n, label) \
+ . = n; \
+ .globl label##_pSeries; \
+label##_pSeries: \
+ HMT_MEDIUM; \
+ mtspr SPRN_SPRG1,r20; /* save r20 */ \
+ mfspr r20,SPRN_HSRR0; /* copy HSRR0 to SRR0 */ \
+ mtspr SPRN_SRR0,r20; \
+ mfspr r20,SPRN_HSRR1; /* copy HSRR0 to SRR0 */ \
+ mtspr SPRN_SRR1,r20; \
+ mfspr r20,SPRN_SPRG1; /* restore r20 */ \
+ mtspr SPRN_SPRG1,r13; /* save r13 */ \
+ EXCEPTION_PROLOG_PSERIES(PACA_EXGEN, label##_common)
+
+
#define STD_EXCEPTION_ISERIES(n, label, area) \
.globl label##_iSeries; \
label##_iSeries: \
@@ -544,8 +559,17 @@ system_call_pSeries:
STD_EXCEPTION_PSERIES(0xf20, altivec_unavailable)
+#ifdef CONFIG_CBE_RAS
+ HSTD_EXCEPTION_PSERIES(0x1200, cbe_system_error)
+#endif /* CONFIG_CBE_RAS */
STD_EXCEPTION_PSERIES(0x1300, instruction_breakpoint)
+#ifdef CONFIG_CBE_RAS
+ HSTD_EXCEPTION_PSERIES(0x1600, cbe_maintenance)
+#endif /* CONFIG_CBE_RAS */
STD_EXCEPTION_PSERIES(0x1700, altivec_assist)
+#ifdef CONFIG_CBE_RAS
+ HSTD_EXCEPTION_PSERIES(0x1800, cbe_thermal)
+#endif /* CONFIG_CBE_RAS */
. = 0x3000
@@ -827,6 +851,11 @@ machine_check_common:
#else
STD_EXCEPTION_COMMON(0x1700, altivec_assist, .unknown_exception)
#endif
+#ifdef CONFIG_CBE_RAS
+ STD_EXCEPTION_COMMON(0x1200, cbe_system_error, .cbe_system_error_exception)
+ STD_EXCEPTION_COMMON(0x1600, cbe_maintenance, .cbe_maintenance_exception)
+ STD_EXCEPTION_COMMON(0x1800, cbe_thermal, .cbe_thermal_exception)
+#endif /* CONFIG_CBE_RAS */
/*
* Here we have detected that the kernel stack pointer is bad.
Index: powerpc.git/arch/powerpc/platforms/cell/Kconfig
===================================================================
--- powerpc.git.orig/arch/powerpc/platforms/cell/Kconfig
+++ powerpc.git/arch/powerpc/platforms/cell/Kconfig
@@ -16,4 +16,8 @@ config SPUFS_MMAP
select MEMORY_HOTPLUG
default y
+config CBE_RAS
+ bool "RAS features for bare metal Cell BE"
+ default y
+
endmenu
Index: powerpc.git/arch/powerpc/platforms/cell/Makefile
===================================================================
--- powerpc.git.orig/arch/powerpc/platforms/cell/Makefile
+++ powerpc.git/arch/powerpc/platforms/cell/Makefile
@@ -1,5 +1,6 @@
obj-y += interrupt.o iommu.o setup.o spider-pic.o
-obj-y += pervasive.o
+obj-y += cbe_regs.o pervasive.o
+obj-$(CONFIG_CBE_RAS) += ras.o
obj-$(CONFIG_SMP) += smp.o
obj-$(CONFIG_SPU_FS) += spu-base.o spufs/
Index: powerpc.git/arch/powerpc/platforms/cell/pervasive.c
===================================================================
--- powerpc.git.orig/arch/powerpc/platforms/cell/pervasive.c
+++ powerpc.git/arch/powerpc/platforms/cell/pervasive.c
@@ -37,36 +37,28 @@
#include <asm/reg.h>
#include "pervasive.h"
+#include "cbe_regs.h"
static DEFINE_SPINLOCK(cbe_pervasive_lock);
-struct cbe_pervasive {
- struct pmd_regs __iomem *regs;
- unsigned int thread;
-};
-
-/* can't use per_cpu from setup_arch */
-static struct cbe_pervasive cbe_pervasive[NR_CPUS];
static void __init cbe_enable_pause_zero(void)
{
unsigned long thread_switch_control;
unsigned long temp_register;
- struct cbe_pervasive *p;
- int thread;
+ struct cbe_pmd_regs __iomem *pregs;
spin_lock_irq(&cbe_pervasive_lock);
- p = &cbe_pervasive[smp_processor_id()];
-
- if (!cbe_pervasive->regs)
+ pregs = cbe_get_cpu_pmd_regs(smp_processor_id());
+ if (pregs == NULL)
goto out;
pr_debug("Power Management: CPU %d\n", smp_processor_id());
/* Enable Pause(0) control bit */
- temp_register = in_be64(&p->regs->pm_control);
+ temp_register = in_be64(&pregs->pm_control);
- out_be64(&p->regs->pm_control,
- temp_register|PMD_PAUSE_ZERO_CONTROL);
+ out_be64(&pregs->pm_control,
+ temp_register | CBE_PMD_PAUSE_ZERO_CONTROL);
/* Enable DEC and EE interrupt request */
thread_switch_control = mfspr(SPRN_TSC_CELL);
@@ -75,25 +67,16 @@ static void __init cbe_enable_pause_zero
switch ((mfspr(SPRN_CTRLF) & CTRL_CT)) {
case CTRL_CT0:
thread_switch_control |= TSC_CELL_DEC_ENABLE_0;
- thread = 0;
break;
case CTRL_CT1:
thread_switch_control |= TSC_CELL_DEC_ENABLE_1;
- thread = 1;
break;
default:
printk(KERN_WARNING "%s: unknown configuration\n",
__FUNCTION__);
- thread = -1;
break;
}
- if (p->thread != thread)
- printk(KERN_WARNING "%s: device tree inconsistant, "
- "cpu %i: %d/%d\n", __FUNCTION__,
- smp_processor_id(),
- p->thread, thread);
-
mtspr(SPRN_TSC_CELL, thread_switch_control);
out:
@@ -104,6 +87,11 @@ static void cbe_idle(void)
{
unsigned long ctrl;
+ /* Why do we do that on every idle ? Couldn't that be done once for
+ * all or do we lose the state some way ? Also, the pm_control
+ * register setting, that can't be set once at boot ? We really want
+ * to move that away in order to implement a simple powersave
+ */
cbe_enable_pause_zero();
while (1) {
@@ -152,8 +140,15 @@ static int cbe_system_reset_exception(st
timer_interrupt(regs);
break;
case SRR1_WAKEMT:
- /* no action required */
break;
+#ifdef CONFIG_CBE_RAS
+ case SRR1_WAKESYSERR:
+ cbe_system_error_exception(regs);
+ break;
+ case SRR1_WAKETHERM:
+ cbe_thermal_exception(regs);
+ break;
+#endif /* CONFIG_CBE_RAS */
default:
/* do system reset */
return 0;
@@ -162,68 +157,11 @@ static int cbe_system_reset_exception(st
return 1;
}
-static int __init cbe_find_pmd_mmio(int cpu, struct cbe_pervasive *p)
+void __init cbe_pervasive_init(void)
{
- struct device_node *node;
- unsigned int *int_servers;
- char *addr;
- unsigned long real_address;
- unsigned int size;
-
- struct pmd_regs __iomem *pmd_mmio_area;
- int hardid, thread;
- int proplen;
-
- pmd_mmio_area = NULL;
- hardid = get_hard_smp_processor_id(cpu);
- for (node = NULL; (node = of_find_node_by_type(node, "cpu"));) {
- int_servers = (void *) get_property(node,
- "ibm,ppc-interrupt-server#s", &proplen);
- if (!int_servers) {
- printk(KERN_WARNING "%s misses "
- "ibm,ppc-interrupt-server#s property",
- node->full_name);
- continue;
- }
- for (thread = 0; thread < proplen / sizeof (int); thread++) {
- if (hardid == int_servers[thread]) {
- addr = get_property(node, "pervasive", NULL);
- goto found;
- }
- }
- }
-
- printk(KERN_WARNING "%s: CPU %d not found\n", __FUNCTION__, cpu);
- return -EINVAL;
-
-found:
- real_address = *(unsigned long*) addr;
- addr += sizeof (unsigned long);
- size = *(unsigned int*) addr;
-
- pr_debug("pervasive area for CPU %d at %lx, size %x\n",
- cpu, real_address, size);
- p->regs = ioremap(real_address, size);
- p->thread = thread;
- return 0;
-}
-
-void __init cell_pervasive_init(void)
-{
- struct cbe_pervasive *p;
- int cpu;
- int ret;
-
if (!cpu_has_feature(CPU_FTR_PAUSE_ZERO))
return;
- for_each_possible_cpu(cpu) {
- p = &cbe_pervasive[cpu];
- ret = cbe_find_pmd_mmio(cpu, p);
- if (ret)
- return;
- }
-
ppc_md.idle_loop = cbe_idle;
ppc_md.system_reset_exception = cbe_system_reset_exception;
}
Index: powerpc.git/arch/powerpc/platforms/cell/ras.c
===================================================================
--- /dev/null
+++ powerpc.git/arch/powerpc/platforms/cell/ras.c
@@ -0,0 +1,112 @@
+#define DEBUG
+
+#include <linux/config.h>
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <linux/smp.h>
+
+#include <asm/reg.h>
+#include <asm/io.h>
+#include <asm/prom.h>
+#include <asm/machdep.h>
+
+#include "ras.h"
+#include "cbe_regs.h"
+
+
+static void dump_fir(int cpu)
+{
+ struct cbe_pmd_regs __iomem *pregs = cbe_get_cpu_pmd_regs(cpu);
+ struct cbe_iic_regs __iomem *iregs = cbe_get_cpu_iic_regs(cpu);
+
+ if (pregs == NULL)
+ return;
+
+ /* Todo: do some nicer parsing of bits and based on them go down
+ * to other sub-units FIRs and not only IIC
+ */
+ printk(KERN_ERR "Global Checkstop FIR : 0x%016lx\n",
+ in_be64(&pregs->checkstop_fir));
+ printk(KERN_ERR "Global Recoverable FIR : 0x%016lx\n",
+ in_be64(&pregs->checkstop_fir));
+ printk(KERN_ERR "Global MachineCheck FIR : 0x%016lx\n",
+ in_be64(&pregs->spec_att_mchk_fir));
+
+ if (iregs == NULL)
+ return;
+ printk(KERN_ERR "IOC FIR : 0x%016lx\n",
+ in_be64(&iregs->ioc_fir));
+
+}
+
+void cbe_system_error_exception(struct pt_regs *regs)
+{
+ int cpu = smp_processor_id();
+
+ printk(KERN_ERR "System Error Interrupt on CPU %d !\n", cpu);
+ dump_fir(cpu);
+ dump_stack();
+}
+
+void cbe_maintenance_exception(struct pt_regs *regs)
+{
+ int cpu = smp_processor_id();
+
+ /*
+ * Nothing implemented for the maintenance interrupt at this point
+ */
+
+ printk(KERN_ERR "Unhandled Maintenance interrupt on CPU %d !\n", cpu);
+ dump_stack();
+}
+
+void cbe_thermal_exception(struct pt_regs *regs)
+{
+ int cpu = smp_processor_id();
+
+ /*
+ * Nothing implemented for the thermal interrupt at this point
+ */
+
+ printk(KERN_ERR "Unhandled Thermal interrupt on CPU %d !\n", cpu);
+ dump_stack();
+}
+
+static int cbe_machine_check_handler(struct pt_regs *regs)
+{
+ int cpu = smp_processor_id();
+
+ printk(KERN_ERR "Machine Check Interrupt on CPU %d !\n", cpu);
+ dump_fir(cpu);
+
+ /* No recovery from this code now, lets continue */
+ return 0;
+}
+
+void __init cbe_ras_init(void)
+{
+ unsigned long hid0;
+
+ /*
+ * Enable System Error & thermal interrupts and wakeup conditions
+ */
+
+ hid0 = mfspr(SPRN_HID0);
+ hid0 |= HID0_CBE_THERM_INT_EN | HID0_CBE_THERM_WAKEUP |
+ HID0_CBE_SYSERR_INT_EN | HID0_CBE_SYSERR_WAKEUP;
+ mtspr(SPRN_HID0, hid0);
+ mb();
+
+ /*
+ * Install machine check handler. Leave setting of precise mode to
+ * what the firmware did for now
+ */
+ ppc_md.machine_check_exception = cbe_machine_check_handler;
+ mb();
+
+ /*
+ * For now, we assume that IOC_FIR is already set to forward some
+ * error conditions to the System Error handler. If that is not true
+ * then it will have to be fixed up here.
+ */
+}
Index: powerpc.git/arch/powerpc/platforms/cell/ras.h
===================================================================
--- /dev/null
+++ powerpc.git/arch/powerpc/platforms/cell/ras.h
@@ -0,0 +1,9 @@
+#ifndef RAS_H
+#define RAS_H
+
+extern void cbe_system_error_exception(struct pt_regs *regs);
+extern void cbe_maintenance_exception(struct pt_regs *regs);
+extern void cbe_thermal_exception(struct pt_regs *regs);
+extern void cbe_ras_init(void);
+
+#endif /* RAS_H */
Index: powerpc.git/arch/powerpc/platforms/cell/setup.c
===================================================================
--- powerpc.git.orig/arch/powerpc/platforms/cell/setup.c
+++ powerpc.git/arch/powerpc/platforms/cell/setup.c
@@ -52,7 +52,9 @@
#include "interrupt.h"
#include "iommu.h"
+#include "cbe_regs.h"
#include "pervasive.h"
+#include "ras.h"
#ifdef DEBUG
#define DBG(fmt...) udbg_printf(fmt)
@@ -82,6 +84,12 @@ static void __init cell_setup_arch(void)
ppc_md.init_IRQ = iic_init_IRQ;
ppc_md.get_irq = iic_get_irq;
+ cbe_regs_init();
+
+#ifdef CONFIG_CBE_RAS
+ cbe_ras_init();
+#endif
+
#ifdef CONFIG_SMP
smp_init_cell();
#endif
@@ -98,7 +106,7 @@ static void __init cell_setup_arch(void)
init_pci_config_tokens();
find_and_init_phbs();
spider_init_IRQ();
- cell_pervasive_init();
+ cbe_pervasive_init();
#ifdef CONFIG_DUMMY_CONSOLE
conswitchp = &dummy_con;
#endif
Index: powerpc.git/include/asm-powerpc/reg.h
===================================================================
--- powerpc.git.orig/include/asm-powerpc/reg.h
+++ powerpc.git/include/asm-powerpc/reg.h
@@ -386,6 +386,8 @@
#define SRR1_WAKEMT 0x00280000 /* mtctrl */
#define SRR1_WAKEDEC 0x00180000 /* Decrementer interrupt */
#define SRR1_WAKETHERM 0x00100000 /* Thermal management interrupt */
+#define SPRN_HSRR0 0x13A /* Save/Restore Register 0 */
+#define SPRN_HSRR1 0x13B /* Save/Restore Register 1 */
#ifndef SPRN_SVR
#define SPRN_SVR 0x11E /* System Version Register */
Index: powerpc.git/arch/powerpc/platforms/cell/pervasive.h
===================================================================
--- powerpc.git.orig/arch/powerpc/platforms/cell/pervasive.h
+++ powerpc.git/arch/powerpc/platforms/cell/pervasive.h
@@ -25,38 +25,9 @@
#ifndef PERVASIVE_H
#define PERVASIVE_H
-struct pmd_regs {
- u8 pad_0x0000_0x0800[0x0800 - 0x0000]; /* 0x0000 */
-
- /* Thermal Sensor Registers */
- u64 ts_ctsr1; /* 0x0800 */
- u64 ts_ctsr2; /* 0x0808 */
- u64 ts_mtsr1; /* 0x0810 */
- u64 ts_mtsr2; /* 0x0818 */
- u64 ts_itr1; /* 0x0820 */
- u64 ts_itr2; /* 0x0828 */
- u64 ts_gitr; /* 0x0830 */
- u64 ts_isr; /* 0x0838 */
- u64 ts_imr; /* 0x0840 */
- u64 tm_cr1; /* 0x0848 */
- u64 tm_cr2; /* 0x0850 */
- u64 tm_simr; /* 0x0858 */
- u64 tm_tpr; /* 0x0860 */
- u64 tm_str1; /* 0x0868 */
- u64 tm_str2; /* 0x0870 */
- u64 tm_tsr; /* 0x0878 */
-
- /* Power Management */
- u64 pm_control; /* 0x0880 */
-#define PMD_PAUSE_ZERO_CONTROL 0x10000
- u64 pm_status; /* 0x0888 */
-
- /* Time Base Register */
- u64 tbr; /* 0x0890 */
-
- u8 pad_0x0898_0x1000 [0x1000 - 0x0898]; /* 0x0898 */
-};
-
-void __init cell_pervasive_init(void);
+extern void cbe_pervasive_init(void);
+extern void cbe_system_error_exception(struct pt_regs *regs);
+extern void cbe_maintenance_exception(struct pt_regs *regs);
+extern void cbe_thermal_exception(struct pt_regs *regs);
#endif
Index: powerpc.git/arch/powerpc/platforms/cell/cbe_regs.h
===================================================================
--- /dev/null
+++ powerpc.git/arch/powerpc/platforms/cell/cbe_regs.h
@@ -0,0 +1,129 @@
+/*
+ * cbe_regs.h
+ *
+ * This file is intended to hold the various register definitions for CBE
+ * on-chip system devices (memory controller, IO controller, etc...)
+ *
+ * (c) 2006 Benjamin Herrenschmidt <benh@kernel.crashing.org>, IBM Corp.
+ */
+
+#ifndef CBE_REGS_H
+#define CBE_REGS_H
+
+/*
+ *
+ * Some HID register definitions
+ *
+ */
+
+/* CBE specific HID0 bits */
+#define HID0_CBE_THERM_WAKEUP 0x0000020000000000ul
+#define HID0_CBE_SYSERR_WAKEUP 0x0000008000000000ul
+#define HID0_CBE_THERM_INT_EN 0x0000000400000000ul
+#define HID0_CBE_SYSERR_INT_EN 0x0000000200000000ul
+
+
+/*
+ *
+ * Pervasive unit register definitions
+ *
+ */
+
+struct cbe_pmd_regs {
+ u8 pad_0x0000_0x0800[0x0800 - 0x0000]; /* 0x0000 */
+
+ /* Thermal Sensor Registers */
+ u64 ts_ctsr1; /* 0x0800 */
+ u64 ts_ctsr2; /* 0x0808 */
+ u64 ts_mtsr1; /* 0x0810 */
+ u64 ts_mtsr2; /* 0x0818 */
+ u64 ts_itr1; /* 0x0820 */
+ u64 ts_itr2; /* 0x0828 */
+ u64 ts_gitr; /* 0x0830 */
+ u64 ts_isr; /* 0x0838 */
+ u64 ts_imr; /* 0x0840 */
+ u64 tm_cr1; /* 0x0848 */
+ u64 tm_cr2; /* 0x0850 */
+ u64 tm_simr; /* 0x0858 */
+ u64 tm_tpr; /* 0x0860 */
+ u64 tm_str1; /* 0x0868 */
+ u64 tm_str2; /* 0x0870 */
+ u64 tm_tsr; /* 0x0878 */
+
+ /* Power Management */
+ u64 pm_control; /* 0x0880 */
+#define CBE_PMD_PAUSE_ZERO_CONTROL 0x10000
+ u64 pm_status; /* 0x0888 */
+
+ /* Time Base Register */
+ u64 tbr; /* 0x0890 */
+
+ u8 pad_0x0898_0x0c00 [0x0c00 - 0x0898]; /* 0x0898 */
+
+ /* Fault Isolation Registers */
+ u64 checkstop_fir; /* 0x0c00 */
+ u64 recoverable_fir;
+ u64 spec_att_mchk_fir;
+ u64 fir_mode_reg;
+ u64 fir_enable_mask;
+
+ u8 pad_0x0c28_0x1000 [0x1000 - 0x0c28]; /* 0x0c28 */
+};
+
+extern struct cbe_pmd_regs __iomem *cbe_get_pmd_regs(struct device_node *np);
+extern struct cbe_pmd_regs __iomem *cbe_get_cpu_pmd_regs(int cpu);
+
+/*
+ *
+ * IIC unit register definitions
+ *
+ */
+
+struct cbe_iic_pending_bits {
+ u32 data;
+ u8 flags;
+ u8 class;
+ u8 source;
+ u8 prio;
+};
+
+#define CBE_IIC_IRQ_VALID 0x80
+#define CBE_IIC_IRQ_IPI 0x40
+
+struct cbe_iic_thread_regs {
+ struct cbe_iic_pending_bits pending;
+ struct cbe_iic_pending_bits pending_destr;
+ u64 generate;
+ u64 prio;
+};
+
+struct cbe_iic_regs {
+ u8 pad_0x0000_0x0400[0x0400 - 0x0000]; /* 0x0000 */
+
+ /* IIC interrupt registers */
+ struct cbe_iic_thread_regs thread[2]; /* 0x0400 */
+ u64 iic_ir; /* 0x0440 */
+ u64 iic_is; /* 0x0448 */
+
+ u8 pad_0x0450_0x0500[0x0500 - 0x0450]; /* 0x0450 */
+
+ /* IOC FIR */
+ u64 ioc_fir_reset; /* 0x0500 */
+ u64 ioc_fir_set;
+ u64 ioc_checkstop_enable;
+ u64 ioc_fir_error_mask;
+ u64 ioc_syserr_enable;
+ u64 ioc_fir;
+
+ u8 pad_0x0530_0x1000[0x1000 - 0x0530]; /* 0x0530 */
+};
+
+extern struct cbe_iic_regs __iomem *cbe_get_iic_regs(struct device_node *np);
+extern struct cbe_iic_regs __iomem *cbe_get_cpu_iic_regs(int cpu);
+
+
+/* Init this module early */
+extern void cbe_regs_init(void);
+
+
+#endif /* CBE_REGS_H */
Index: powerpc.git/arch/powerpc/platforms/cell/interrupt.c
===================================================================
--- powerpc.git.orig/arch/powerpc/platforms/cell/interrupt.c
+++ powerpc.git/arch/powerpc/platforms/cell/interrupt.c
@@ -33,29 +33,10 @@
#include <asm/ptrace.h>
#include "interrupt.h"
-
-struct iic_pending_bits {
- u32 data;
- u8 flags;
- u8 class;
- u8 source;
- u8 prio;
-};
-
-enum iic_pending_flags {
- IIC_VALID = 0x80,
- IIC_IPI = 0x40,
-};
-
-struct iic_regs {
- struct iic_pending_bits pending;
- struct iic_pending_bits pending_destr;
- u64 generate;
- u64 prio;
-};
+#include "cbe_regs.h"
struct iic {
- struct iic_regs __iomem *regs;
+ struct cbe_iic_thread_regs __iomem *regs;
u8 target_id;
};
@@ -115,7 +96,7 @@ static struct hw_interrupt_type iic_pic
.end = iic_end,
};
-static int iic_external_get_irq(struct iic_pending_bits pending)
+static int iic_external_get_irq(struct cbe_iic_pending_bits pending)
{
int irq;
unsigned char node, unit;
@@ -168,15 +149,15 @@ int iic_get_irq(struct pt_regs *regs)
{
struct iic *iic;
int irq;
- struct iic_pending_bits pending;
+ struct cbe_iic_pending_bits pending;
iic = &__get_cpu_var(iic);
*(unsigned long *) &pending =
in_be64((unsigned long __iomem *) &iic->regs->pending_destr);
irq = -1;
- if (pending.flags & IIC_VALID) {
- if (pending.flags & IIC_IPI) {
+ if (pending.flags & CBE_IIC_IRQ_VALID) {
+ if (pending.flags & CBE_IIC_IRQ_IPI) {
irq = IIC_IPI_OFFSET + (pending.prio >> 4);
/*
if (irq > 0x80)
@@ -226,7 +207,7 @@ static int setup_iic_hardcoded(void)
regs += 0x20;
printk(KERN_INFO "IIC for CPU %d at %lx\n", cpu, regs);
- iic->regs = ioremap(regs, sizeof(struct iic_regs));
+ iic->regs = ioremap(regs, sizeof(struct cbe_iic_thread_regs));
iic->target_id = (nodeid << 4) + ((cpu & 1) ? 0xf : 0xe);
}
@@ -267,12 +248,12 @@ static int setup_iic(void)
}
iic = &per_cpu(iic, np[0]);
- iic->regs = ioremap(regs[0], sizeof(struct iic_regs));
+ iic->regs = ioremap(regs[0], sizeof(struct cbe_iic_thread_regs));
iic->target_id = ((np[0] & 2) << 3) + ((np[0] & 1) ? 0xf : 0xe);
printk("IIC for CPU %d at %lx mapped to %p\n", np[0], regs[0], iic->regs);
iic = &per_cpu(iic, np[1]);
- iic->regs = ioremap(regs[2], sizeof(struct iic_regs));
+ iic->regs = ioremap(regs[2], sizeof(struct cbe_iic_thread_regs));
iic->target_id = ((np[1] & 2) << 3) + ((np[1] & 1) ? 0xf : 0xe);
printk("IIC for CPU %d at %lx mapped to %p\n", np[1], regs[2], iic->regs);
Index: powerpc.git/arch/powerpc/kernel/prom.c
===================================================================
--- powerpc.git.orig/arch/powerpc/kernel/prom.c
+++ powerpc.git/arch/powerpc/kernel/prom.c
@@ -2096,3 +2096,46 @@ int prom_update_property(struct device_n
return 0;
}
+
+/* Find the device node for a given logical cpu number, also returns the cpu
+ * local thread number (index in ibm,interrupt-server#s) if relevant and
+ * asked for (non NULL)
+ */
+struct device_node *of_get_cpu_node(int cpu, unsigned int *thread)
+{
+ int hardid;
+ struct device_node *np;
+
+ hardid = get_hard_smp_processor_id(cpu);
+
+ for_each_node_by_type(np, "cpu") {
+ u32 *intserv;
+ unsigned int plen, t;
+
+ /* Check for ibm,ppc-interrupt-server#s. If it doesn't exist
+ * fallback to "reg" property and assume no threads
+ */
+ intserv = (u32 *)get_property(np, "ibm,ppc-interrupt-server#s",
+ &plen);
+ if (intserv == NULL) {
+ u32 *reg = (u32 *)get_property(np, "reg", NULL);
+ if (reg == NULL)
+ continue;
+ if (*reg == hardid) {
+ if (thread)
+ *thread = 0;
+ return np;
+ }
+ } else {
+ plen /= sizeof(u32);
+ for (t = 0; t < plen; t++) {
+ if (hardid == intserv[t]) {
+ if (thread)
+ *thread = t;
+ return np;
+ }
+ }
+ }
+ }
+ return NULL;
+}
Index: powerpc.git/arch/powerpc/platforms/cell/cbe_regs.c
===================================================================
--- /dev/null
+++ powerpc.git/arch/powerpc/platforms/cell/cbe_regs.c
@@ -0,0 +1,128 @@
+/*
+ * cbe_regs.c
+ *
+ * Accessor routines for the various MMIO register blocks of the CBE
+ *
+ * (c) 2006 Benjamin Herrenschmidt <benh@kernel.crashing.org>, IBM Corp.
+ */
+
+
+#include <linux/config.h>
+#include <linux/percpu.h>
+#include <linux/types.h>
+
+#include <asm/io.h>
+#include <asm/pgtable.h>
+#include <asm/prom.h>
+#include <asm/ptrace.h>
+
+#include "cbe_regs.h"
+
+#define MAX_CBE 2
+
+/*
+ * Current implementation uses "cpu" nodes. We build our own mapping
+ * array of cpu numbers to cpu nodes locally for now to allow interrupt
+ * time code to have a fast path rather than call of_get_cpu_node(). If
+ * we implement cpu hotplug, we'll have to install an appropriate norifier
+ * in order to release references to the cpu going away
+ */
+static struct cbe_regs_map
+{
+ struct device_node *cpu_node;
+ struct cbe_pmd_regs __iomem *pmd_regs;
+ struct cbe_iic_regs __iomem *iic_regs;
+} cbe_regs_maps[MAX_CBE];
+static int cbe_regs_map_count;
+
+static struct cbe_thread_map
+{
+ struct device_node *cpu_node;
+ struct cbe_regs_map *regs;
+} cbe_thread_map[NR_CPUS];
+
+static struct cbe_regs_map *cbe_find_map(struct device_node *np)
+{
+ int i;
+
+ for (i = 0; i < cbe_regs_map_count; i++)
+ if (cbe_regs_maps[i].cpu_node == np)
+ return &cbe_regs_maps[i];
+ return NULL;
+}
+
+struct cbe_pmd_regs __iomem *cbe_get_pmd_regs(struct device_node *np)
+{
+ struct cbe_regs_map *map = cbe_find_map(np);
+ if (map == NULL)
+ return NULL;
+ return map->pmd_regs;
+}
+
+struct cbe_pmd_regs __iomem *cbe_get_cpu_pmd_regs(int cpu)
+{
+ struct cbe_regs_map *map = cbe_thread_map[cpu].regs;
+ if (map == NULL)
+ return NULL;
+ return map->pmd_regs;
+}
+
+
+struct cbe_iic_regs __iomem *cbe_get_iic_regs(struct device_node *np)
+{
+ struct cbe_regs_map *map = cbe_find_map(np);
+ if (map == NULL)
+ return NULL;
+ return map->iic_regs;
+}
+struct cbe_iic_regs __iomem *cbe_get_cpu_iic_regs(int cpu)
+{
+ struct cbe_regs_map *map = cbe_thread_map[cpu].regs;
+ if (map == NULL)
+ return NULL;
+ return map->iic_regs;
+}
+
+void __init cbe_regs_init(void)
+{
+ int i;
+ struct device_node *cpu;
+
+ /* Build local fast map of CPUs */
+ for_each_cpu(i)
+ cbe_thread_map[i].cpu_node = of_get_cpu_node(i, NULL);
+
+ /* Find maps for each device tree CPU */
+ for_each_node_by_type(cpu, "cpu") {
+ struct cbe_regs_map *map = &cbe_regs_maps[cbe_regs_map_count++];
+
+ /* That hack must die die die ! */
+ struct address_prop {
+ unsigned long address;
+ unsigned int len;
+ } __attribute__((packed)) *prop;
+
+
+ if (cbe_regs_map_count > MAX_CBE) {
+ printk(KERN_ERR "cbe_regs: More BE chips than supported"
+ "!\n");
+ cbe_regs_map_count--;
+ return;
+ }
+ map->cpu_node = cpu;
+ for_each_cpu(i)
+ if (cbe_thread_map[i].cpu_node == cpu)
+ cbe_thread_map[i].regs = map;
+
+ prop = (struct address_prop *)get_property(cpu, "pervasive",
+ NULL);
+ if (prop != NULL)
+ map->pmd_regs = ioremap(prop->address, prop->len);
+
+ prop = (struct address_prop *)get_property(cpu, "iic",
+ NULL);
+ if (prop != NULL)
+ map->iic_regs = ioremap(prop->address, prop->len);
+ }
+}
+
Index: powerpc.git/include/asm-powerpc/prom.h
===================================================================
--- powerpc.git.orig/include/asm-powerpc/prom.h
+++ powerpc.git/include/asm-powerpc/prom.h
@@ -238,5 +238,8 @@ void of_parse_dma_window(struct device_n
extern void kdump_move_device_tree(void);
+/* CPU OF node matching */
+struct device_node *of_get_cpu_node(int cpu, unsigned int *thread);
+
#endif /* __KERNEL__ */
#endif /* _POWERPC_PROM_H */
Index: powerpc.git/arch/powerpc/configs/cell_defconfig
===================================================================
--- powerpc.git.orig/arch/powerpc/configs/cell_defconfig
+++ powerpc.git/arch/powerpc/configs/cell_defconfig
@@ -133,6 +133,7 @@ CONFIG_CELL_IIC=y
#
CONFIG_SPU_FS=m
CONFIG_SPUFS_MMAP=y
+CONFIG_CBE_RAS=y
#
# Kernel options
--
^ permalink raw reply
* [patch 02/20] cell: fix interrupt priority handling
From: arnd @ 2006-06-19 18:33 UTC (permalink / raw)
To: paulus; +Cc: linuxppc-dev, cbe-oss-dev, linux-kernel
In-Reply-To: <20060619183315.653672000@klappe.arndb.de>
Checking the priority field to test for irq validity is
completely bogus and breaks with future external interrupt
controllers.
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
---
Index: powerpc.git/arch/powerpc/platforms/cell/interrupt.c
===================================================================
--- powerpc.git.orig/arch/powerpc/platforms/cell/interrupt.c
+++ powerpc.git/arch/powerpc/platforms/cell/interrupt.c
@@ -117,8 +117,7 @@ static int iic_external_get_irq(struct c
* One of these units can be connected
* to an external interrupt controller.
*/
- if (pending.prio > 0x3f ||
- pending.class != 2)
+ if (pending.class != 2)
break;
irq = IIC_EXT_OFFSET
+ spider_get_irq(node)
--
^ permalink raw reply
* [patch 03/20] cell: update defconfig
From: arnd @ 2006-06-19 18:33 UTC (permalink / raw)
To: paulus; +Cc: linuxppc-dev, cbe-oss-dev, linux-kernel
In-Reply-To: <20060619183315.653672000@klappe.arndb.de>
Enable some of the most requested features in defconfig
and refresh with the latest powerpc.git Kconfig files.
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
---
Index: powerpc.git/arch/powerpc/configs/cell_defconfig
===================================================================
--- powerpc.git.orig/arch/powerpc/configs/cell_defconfig
+++ powerpc.git/arch/powerpc/configs/cell_defconfig
@@ -1,7 +1,7 @@
#
# Automatically generated make config: don't edit
-# Linux kernel version: 2.6.16
-# Thu Mar 23 20:48:09 2006
+# Linux kernel version: 2.6.17
+# Mon Jun 19 17:23:03 2006
#
CONFIG_PPC64=y
CONFIG_64BIT=y
@@ -11,6 +11,7 @@ CONFIG_GENERIC_HARDIRQS=y
CONFIG_RWSEM_XCHGADD_ALGORITHM=y
CONFIG_GENERIC_HWEIGHT=y
CONFIG_GENERIC_CALIBRATE_DELAY=y
+CONFIG_GENERIC_FIND_NEXT_BIT=y
CONFIG_PPC=y
CONFIG_EARLY_PRINTK=y
CONFIG_COMPAT=y
@@ -55,7 +56,7 @@ CONFIG_SYSCTL=y
# CONFIG_AUDIT is not set
CONFIG_IKCONFIG=y
CONFIG_IKCONFIG_PROC=y
-# CONFIG_CPUSETS is not set
+CONFIG_CPUSETS=y
# CONFIG_RELAY is not set
CONFIG_INITRAMFS_SOURCE=""
CONFIG_CC_OPTIMIZE_FOR_SIZE=y
@@ -116,6 +117,7 @@ CONFIG_PPC_MULTIPLATFORM=y
# CONFIG_PPC_PMAC is not set
# CONFIG_PPC_MAPLE is not set
CONFIG_PPC_CELL=y
+CONFIG_PPC_SYSTEMSIM=y
# CONFIG_U3_DART is not set
CONFIG_PPC_RTAS=y
# CONFIG_RTAS_ERROR_LOGGING is not set
@@ -153,20 +155,24 @@ CONFIG_FORCE_MAX_ZONEORDER=13
CONFIG_KEXEC=y
# CONFIG_CRASH_DUMP is not set
CONFIG_IRQ_ALL_CPUS=y
-# CONFIG_NUMA is not set
+CONFIG_NUMA=y
+CONFIG_NODES_SHIFT=4
CONFIG_ARCH_SELECT_MEMORY_MODEL=y
-CONFIG_ARCH_FLATMEM_ENABLE=y
CONFIG_ARCH_SPARSEMEM_ENABLE=y
CONFIG_SELECT_MEMORY_MODEL=y
# CONFIG_FLATMEM_MANUAL is not set
# CONFIG_DISCONTIGMEM_MANUAL is not set
CONFIG_SPARSEMEM_MANUAL=y
CONFIG_SPARSEMEM=y
+CONFIG_NEED_MULTIPLE_NODES=y
CONFIG_HAVE_MEMORY_PRESENT=y
# CONFIG_SPARSEMEM_STATIC is not set
CONFIG_SPARSEMEM_EXTREME=y
-# CONFIG_MEMORY_HOTPLUG is not set
+CONFIG_MEMORY_HOTPLUG=y
CONFIG_SPLIT_PTLOCK_CPUS=4
+CONFIG_MIGRATION=y
+CONFIG_HAVE_ARCH_EARLY_PFN_TO_NID=y
+CONFIG_ARCH_MEMORY_PROBE=y
# CONFIG_PPC_64K_PAGES is not set
CONFIG_SCHED_SMT=y
CONFIG_PROC_DEVICETREE=y
@@ -183,6 +189,7 @@ CONFIG_GENERIC_ISA_DMA=y
# CONFIG_PPC_INDIRECT_PCI is not set
CONFIG_PCI=y
CONFIG_PCI_DOMAINS=y
+CONFIG_PCIEPORTBUS=y
# CONFIG_PCI_DEBUG is not set
#
@@ -477,7 +484,7 @@ CONFIG_DM_MULTIPATH=m
#
CONFIG_NETDEVICES=y
# CONFIG_DUMMY is not set
-# CONFIG_BONDING is not set
+CONFIG_BONDING=y
# CONFIG_EQUALIZER is not set
# CONFIG_TUN is not set
@@ -625,6 +632,7 @@ CONFIG_SERIAL_NONSTANDARD=y
# CONFIG_N_HDLC is not set
# CONFIG_SPECIALIX is not set
# CONFIG_SX is not set
+# CONFIG_RIO is not set
# CONFIG_STALDRV is not set
#
@@ -767,6 +775,7 @@ CONFIG_I2C_ALGOBIT=y
# Multimedia devices
#
# CONFIG_VIDEO_DEV is not set
+CONFIG_VIDEO_V4L2=y
#
# Digital Video Broadcasting Devices
@@ -1055,11 +1064,7 @@ CONFIG_DEBUGGER=y
# CONFIG_XMON is not set
CONFIG_IRQSTACKS=y
# CONFIG_BOOTX_TEXT is not set
-# CONFIG_PPC_EARLY_DEBUG_LPAR is not set
-# CONFIG_PPC_EARLY_DEBUG_G5 is not set
-# CONFIG_PPC_EARLY_DEBUG_RTAS is not set
-# CONFIG_PPC_EARLY_DEBUG_MAPLE is not set
-# CONFIG_PPC_EARLY_DEBUG_ISERIES is not set
+# CONFIG_PPC_EARLY_DEBUG is not set
#
# Security options
--
^ permalink raw reply
* [patch 04/20] cell: register SPUs as sysdevs
From: arnd @ 2006-06-19 18:33 UTC (permalink / raw)
To: paulus; +Cc: linuxppc-dev, cbe-oss-dev, linux-kernel
In-Reply-To: <20060619183315.653672000@klappe.arndb.de>
From: Jeremy Kerr <jk@ozlabs.org>
SPUs are registered as system devices, exposing attributes through
sysfs. Since the sysdev includes a kref, we can remove the one in
struct spu (it isn't used at the moment anyway).
Currently only the interrupt source and numa node attributes are added.
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Index: powerpc.git/arch/powerpc/platforms/cell/spu_base.c
===================================================================
--- powerpc.git.orig/arch/powerpc/platforms/cell/spu_base.c
+++ powerpc.git/arch/powerpc/platforms/cell/spu_base.c
@@ -649,6 +649,46 @@ out:
return ret;
}
+struct sysdev_class spu_sysdev_class = {
+ set_kset_name("spu")
+};
+
+static ssize_t spu_show_isrc(struct sys_device *sysdev, char *buf)
+{
+ struct spu *spu = container_of(sysdev, struct spu, sysdev);
+ return sprintf(buf, "%d\n", spu->isrc);
+
+}
+static SYSDEV_ATTR(isrc, 0400, spu_show_isrc, NULL);
+
+extern int attach_sysdev_to_node(struct sys_device *dev, int nid);
+
+static int spu_create_sysdev(struct spu *spu)
+{
+ int ret;
+
+ spu->sysdev.id = spu->number;
+ spu->sysdev.cls = &spu_sysdev_class;
+ ret = sysdev_register(&spu->sysdev);
+ if (ret) {
+ printk(KERN_ERR "Can't register SPU %d with sysfs\n",
+ spu->number);
+ return ret;
+ }
+
+ sysdev_create_file(&spu->sysdev, &attr_isrc);
+ sysfs_add_device_to_node(&spu->sysdev, spu->nid);
+
+ return 0;
+}
+
+static void spu_destroy_sysdev(struct spu *spu)
+{
+ sysdev_remove_file(&spu->sysdev, &attr_isrc);
+ sysfs_remove_device_from_node(&spu->sysdev, spu->nid);
+ sysdev_unregister(&spu->sysdev);
+}
+
static int __init create_spu(struct device_node *spe)
{
struct spu *spu;
@@ -695,6 +735,10 @@ static int __init create_spu(struct devi
if (ret)
goto out_unmap;
+ ret = spu_create_sysdev(spu);
+ if (ret)
+ goto out_free_irqs;
+
list_add(&spu->list, &spu_list);
mutex_unlock(&spu_mutex);
@@ -703,6 +747,9 @@ static int __init create_spu(struct devi
spu->problem, spu->priv1, spu->priv2, spu->number);
goto out;
+out_free_irqs:
+ spu_free_irqs(spu);
+
out_unmap:
mutex_unlock(&spu_mutex);
spu_unmap(spu);
@@ -716,6 +763,7 @@ static void destroy_spu(struct spu *spu)
{
list_del_init(&spu->list);
+ spu_destroy_sysdev(spu);
spu_free_irqs(spu);
spu_unmap(spu);
kfree(spu);
@@ -728,6 +776,7 @@ static void cleanup_spu_base(void)
list_for_each_entry_safe(spu, tmp, &spu_list, list)
destroy_spu(spu);
mutex_unlock(&spu_mutex);
+ sysdev_class_unregister(&spu_sysdev_class);
}
module_exit(cleanup_spu_base);
@@ -736,6 +785,11 @@ static int __init init_spu_base(void)
struct device_node *node;
int ret;
+ /* create sysdev class for spus */
+ ret = sysdev_class_register(&spu_sysdev_class);
+ if (ret)
+ return ret;
+
ret = -ENODEV;
for (node = of_find_node_by_type(NULL, "spe");
node; node = of_find_node_by_type(node, "spe")) {
Index: powerpc.git/include/asm-powerpc/spu.h
===================================================================
--- powerpc.git.orig/include/asm-powerpc/spu.h
+++ powerpc.git/include/asm-powerpc/spu.h
@@ -25,8 +25,8 @@
#ifdef __KERNEL__
#include <linux/config.h>
-#include <linux/kref.h>
#include <linux/workqueue.h>
+#include <linux/sysdev.h>
#define LS_SIZE (256 * 1024)
#define LS_ADDR_MASK (LS_SIZE - 1)
@@ -123,7 +123,6 @@ struct spu {
u64 flags;
u64 dar;
u64 dsisr;
- struct kref kref;
size_t ls_size;
unsigned int slb_replace;
struct mm_struct *mm;
@@ -144,6 +143,8 @@ struct spu {
char irq_c0[8];
char irq_c1[8];
char irq_c2[8];
+
+ struct sys_device sysdev;
};
struct spu *spu_alloc(void);
--
^ permalink raw reply
* [patch 05/20] cell: always build spu base into the kernel
From: arnd @ 2006-06-19 18:33 UTC (permalink / raw)
To: paulus; +Cc: linuxppc-dev, cbe-oss-dev, linux-kernel
In-Reply-To: <20060619183315.653672000@klappe.arndb.de>
The spu_base module is rather deeply intermixed with the
core kernel, so it makes sense to have that built-in.
This will let us extend the base in the future without
having to export more core symbols just for it.
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Index: powerpc.git/arch/powerpc/platforms/cell/Makefile
===================================================================
--- powerpc.git.orig/arch/powerpc/platforms/cell/Makefile
+++ powerpc.git/arch/powerpc/platforms/cell/Makefile
@@ -3,15 +3,11 @@ obj-y += cbe_regs.o pervasive.o
obj-$(CONFIG_CBE_RAS) += ras.o
obj-$(CONFIG_SMP) += smp.o
-obj-$(CONFIG_SPU_FS) += spu-base.o spufs/
-
-spu-base-y += spu_base.o spu_priv1.o
# needed only when building loadable spufs.ko
spufs-modular-$(CONFIG_SPU_FS) += spu_syscalls.o
obj-y += $(spufs-modular-m)
# always needed in kernel
-spufs-builtin-$(CONFIG_SPU_FS) += spu_callbacks.o
+spufs-builtin-$(CONFIG_SPU_FS) += spu_callbacks.o spu_base.o spu_priv1.o spufs/
obj-y += $(spufs-builtin-y) $(spufs-builtin-m)
-
Index: powerpc.git/arch/powerpc/platforms/cell/spufs/Makefile
===================================================================
--- powerpc.git.orig/arch/powerpc/platforms/cell/spufs/Makefile
+++ powerpc.git/arch/powerpc/platforms/cell/spufs/Makefile
@@ -1,5 +1,7 @@
+obj-y += switch.o
+
obj-$(CONFIG_SPU_FS) += spufs.o
-spufs-y += inode.o file.o context.o switch.o syscalls.o
+spufs-y += inode.o file.o context.o syscalls.o
spufs-y += sched.o backing_ops.o hw_ops.o run.o
# Rules to build switch.o with the help of SPU tool chain
Index: powerpc.git/arch/powerpc/platforms/cell/spufs/switch.c
===================================================================
--- powerpc.git.orig/arch/powerpc/platforms/cell/spufs/switch.c
+++ powerpc.git/arch/powerpc/platforms/cell/spufs/switch.c
@@ -2074,6 +2074,7 @@ int spu_save(struct spu_state *prev, str
}
return rc;
}
+EXPORT_SYMBOL_GPL(spu_save);
/**
* spu_restore - SPU context restore, with harvest and locking.
@@ -2103,6 +2104,7 @@ int spu_restore(struct spu_state *new, s
}
return rc;
}
+EXPORT_SYMBOL_GPL(spu_restore);
/**
* spu_harvest - SPU harvest (reset) operation
@@ -2193,6 +2195,7 @@ void spu_init_csa(struct spu_state *csa)
init_priv1(csa);
init_priv2(csa);
}
+EXPORT_SYMBOL_GPL(spu_init_csa);
void spu_fini_csa(struct spu_state *csa)
{
@@ -2203,3 +2206,4 @@ void spu_fini_csa(struct spu_state *csa)
vfree(csa->lscsa);
}
+EXPORT_SYMBOL_GPL(spu_fini_csa);
--
^ permalink raw reply
* [patch 06/20] spufs: restore mapping of mssync register
From: arnd @ 2006-06-19 18:33 UTC (permalink / raw)
To: paulus; +Cc: linuxppc-dev, cbe-oss-dev, linux-kernel
In-Reply-To: <20060619183315.653672000@klappe.arndb.de>
A recent change to the way that the mfc file gets mapped made it
impossible to map the SPE Multi-Source Synchronization register
into user space, but that may be needed by some applications.
This restores the missing functionality.
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
---
Index: powerpc.git/arch/powerpc/platforms/cell/spufs/file.c
===================================================================
--- powerpc.git.orig/arch/powerpc/platforms/cell/spufs/file.c
+++ powerpc.git/arch/powerpc/platforms/cell/spufs/file.c
@@ -825,6 +825,55 @@ DEFINE_SIMPLE_ATTRIBUTE(spufs_signal2_ty
spufs_signal2_type_set, "%llu");
#ifdef CONFIG_SPUFS_MMAP
+static struct page *spufs_mss_mmap_nopage(struct vm_area_struct *vma,
+ unsigned long address, int *type)
+{
+ return spufs_ps_nopage(vma, address, type, 0x0000);
+}
+
+static struct vm_operations_struct spufs_mss_mmap_vmops = {
+ .nopage = spufs_mss_mmap_nopage,
+};
+
+/*
+ * mmap support for problem state MFC DMA area [0x0000 - 0x0fff].
+ * Mapping this area requires that the application have CAP_SYS_RAWIO,
+ * as these registers require special care when read/writing.
+ */
+static int spufs_mss_mmap(struct file *file, struct vm_area_struct *vma)
+{
+ if (!(vma->vm_flags & VM_SHARED))
+ return -EINVAL;
+
+ if (!capable(CAP_SYS_RAWIO))
+ return -EPERM;
+
+ vma->vm_flags |= VM_RESERVED;
+ vma->vm_page_prot = __pgprot(pgprot_val(vma->vm_page_prot)
+ | _PAGE_NO_CACHE);
+
+ vma->vm_ops = &spufs_mss_mmap_vmops;
+ return 0;
+}
+#endif
+
+static int spufs_mss_open(struct inode *inode, struct file *file)
+{
+ struct spufs_inode_info *i = SPUFS_I(inode);
+
+ file->private_data = i->i_ctx;
+ return nonseekable_open(inode, file);
+}
+
+static struct file_operations spufs_mss_fops = {
+ .open = spufs_mss_open,
+#ifdef CONFIG_SPUFS_MMAP
+ .mmap = spufs_mss_mmap,
+#endif
+};
+
+
+#ifdef CONFIG_SPUFS_MMAP
static struct page *spufs_mfc_mmap_nopage(struct vm_area_struct *vma,
unsigned long address, int *type)
{
@@ -1292,6 +1341,7 @@ struct tree_descr spufs_dir_contents[] =
{ "signal2", &spufs_signal2_fops, 0666, },
{ "signal1_type", &spufs_signal1_type, 0666, },
{ "signal2_type", &spufs_signal2_type, 0666, },
+ { "mss", &spufs_mss_fops, 0666, },
{ "mfc", &spufs_mfc_fops, 0666, },
{ "cntl", &spufs_cntl_fops, 0666, },
{ "npc", &spufs_npc_ops, 0666, },
--
^ permalink raw reply
* [patch 07/20] spufs: fix deadlock in spu_create error path
From: arnd @ 2006-06-19 18:33 UTC (permalink / raw)
To: paulus; +Cc: linuxppc-dev, cbe-oss-dev, linux-kernel
In-Reply-To: <20060619183315.653672000@klappe.arndb.de>
From: Michael Ellerman <michael@ellerman.id.au>
spufs_rmdir tries to acquire the spufs root
i_mutex, which is already held by spufs_create_thread.
This was tracked as Bug #H9512.
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
---
Index: powerpc.git/arch/powerpc/platforms/cell/spufs/inode.c
===================================================================
--- powerpc.git.orig/arch/powerpc/platforms/cell/spufs/inode.c
+++ powerpc.git/arch/powerpc/platforms/cell/spufs/inode.c
@@ -157,20 +157,12 @@ static void spufs_prune_dir(struct dentr
mutex_unlock(&dir->d_inode->i_mutex);
}
+/* Caller must hold root->i_mutex */
static int spufs_rmdir(struct inode *root, struct dentry *dir_dentry)
{
- struct spu_context *ctx;
-
/* remove all entries */
- mutex_lock(&root->i_mutex);
spufs_prune_dir(dir_dentry);
- mutex_unlock(&root->i_mutex);
- /* We have to give up the mm_struct */
- ctx = SPUFS_I(dir_dentry->d_inode)->i_ctx;
- spu_forget(ctx);
-
- /* XXX Do we need to hold i_mutex here ? */
return simple_rmdir(root, dir_dentry);
}
@@ -199,16 +191,23 @@ out:
static int spufs_dir_close(struct inode *inode, struct file *file)
{
+ struct spu_context *ctx;
struct inode *dir;
struct dentry *dentry;
int ret;
dentry = file->f_dentry;
dir = dentry->d_parent->d_inode;
+ ctx = SPUFS_I(dentry->d_inode)->i_ctx;
+ mutex_lock(&dir->i_mutex);
ret = spufs_rmdir(dir, dentry);
+ mutex_unlock(&dir->i_mutex);
WARN_ON(ret);
+ /* We have to give up the mm_struct */
+ spu_forget(ctx);
+
return dcache_dir_close(inode, file);
}
@@ -324,8 +323,13 @@ long spufs_create_thread(struct nameidat
* in error path of *_open().
*/
ret = spufs_context_open(dget(dentry), mntget(nd->mnt));
- if (ret < 0)
- spufs_rmdir(nd->dentry->d_inode, dentry);
+ if (ret < 0) {
+ WARN_ON(spufs_rmdir(nd->dentry->d_inode, dentry));
+ mutex_unlock(&nd->dentry->d_inode->i_mutex);
+ spu_forget(SPUFS_I(dentry->d_inode)->i_ctx);
+ dput(dentry);
+ goto out;
+ }
out_dput:
dput(dentry);
--
^ permalink raw reply
* [patch 08/20] spufs: set up correct SLB entries for 64k pages
From: arnd @ 2006-06-19 18:33 UTC (permalink / raw)
To: paulus; +Cc: linuxppc-dev, cbe-oss-dev, linux-kernel
In-Reply-To: <20060619183315.653672000@klappe.arndb.de>
spufs currently knows only 4k pages and 16M hugetlb
pages. Make it use the regular methods for deciding on
the SLB bits.
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
---
Index: powerpc.git/arch/powerpc/platforms/cell/spu_base.c
===================================================================
--- powerpc.git.orig/arch/powerpc/platforms/cell/spu_base.c
+++ powerpc.git/arch/powerpc/platforms/cell/spu_base.c
@@ -71,7 +71,7 @@ static int __spu_trap_data_seg(struct sp
{
struct spu_priv2 __iomem *priv2 = spu->priv2;
struct mm_struct *mm = spu->mm;
- u64 esid, vsid;
+ u64 esid, vsid, llp;
pr_debug("%s\n", __FUNCTION__);
@@ -91,9 +91,14 @@ static int __spu_trap_data_seg(struct sp
}
esid = (ea & ESID_MASK) | SLB_ESID_V;
- vsid = (get_vsid(mm->context.id, ea) << SLB_VSID_SHIFT) | SLB_VSID_USER;
+#ifdef CONFIG_HUGETLB_PAGE
if (in_hugepage_area(mm->context, ea))
- vsid |= SLB_VSID_L;
+ llp = mmu_psize_defs[mmu_huge_psize].sllp;
+ else
+#endif
+ llp = mmu_psize_defs[mmu_virtual_psize].sllp;
+ vsid = (get_vsid(mm->context.id, ea) << SLB_VSID_SHIFT) |
+ SLB_VSID_USER | llp;
out_be64(&priv2->slb_index_W, spu->slb_replace);
out_be64(&priv2->slb_vsid_RW, vsid);
Index: powerpc.git/arch/powerpc/platforms/cell/spufs/switch.c
===================================================================
--- powerpc.git.orig/arch/powerpc/platforms/cell/spufs/switch.c
+++ powerpc.git/arch/powerpc/platforms/cell/spufs/switch.c
@@ -718,13 +718,15 @@ static inline void invalidate_slbs(struc
static inline void get_kernel_slb(u64 ea, u64 slb[2])
{
- slb[0] = (get_kernel_vsid(ea) << SLB_VSID_SHIFT) | SLB_VSID_KERNEL;
- slb[1] = (ea & ESID_MASK) | SLB_ESID_V;
+ u64 llp;
- /* Large pages are used for kernel text/data, but not vmalloc. */
- if (cpu_has_feature(CPU_FTR_16M_PAGE)
- && REGION_ID(ea) == KERNEL_REGION_ID)
- slb[0] |= SLB_VSID_L;
+ if (REGION_ID(ea) == KERNEL_REGION_ID)
+ llp = mmu_psize_defs[mmu_linear_psize].sllp;
+ else
+ llp = mmu_psize_defs[mmu_virtual_psize].sllp;
+ slb[0] = (get_kernel_vsid(ea) << SLB_VSID_SHIFT) |
+ SLB_VSID_KERNEL | llp;
+ slb[1] = (ea & ESID_MASK) | SLB_ESID_V;
}
static inline void load_mfc_slb(struct spu *spu, u64 slb[2], int slbe)
--
^ permalink raw reply
* [patch 09/20] spufs: add a phys-id attribute to each SPU context
From: arnd @ 2006-06-19 18:33 UTC (permalink / raw)
To: paulus; +Cc: linuxppc-dev, cbe-oss-dev, linux-kernel
In-Reply-To: <20060619183315.653672000@klappe.arndb.de>
For performance analysis, it is often interesting to know
which physical SPE a thread is currently running on, and,
more importantly, if it is running at all.
This patch adds a simple attribute to each SPU directory
with that information.
The attribute is read-only and called 'phys-id'. It contains
an ascii string with the number of the physical SPU (e.g.
"0x5"), or alternatively the string "0xffffffff" (32 bit -1)
when it is not running at all at the time that the file
is read.
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
---
Index: powerpc.git/arch/powerpc/platforms/cell/spufs/file.c
===================================================================
--- powerpc.git.orig/arch/powerpc/platforms/cell/spufs/file.c
+++ powerpc.git/arch/powerpc/platforms/cell/spufs/file.c
@@ -1328,6 +1328,22 @@ static u64 spufs_srr0_get(void *data)
DEFINE_SIMPLE_ATTRIBUTE(spufs_srr0_ops, spufs_srr0_get, spufs_srr0_set,
"%llx\n")
+static u64 spufs_id_get(void *data)
+{
+ struct spu_context *ctx = data;
+ u64 num;
+
+ spu_acquire(ctx);
+ if (ctx->state == SPU_STATE_RUNNABLE)
+ num = ctx->spu->number;
+ else
+ num = (unsigned int)-1;
+ spu_release(ctx);
+
+ return num;
+}
+DEFINE_SIMPLE_ATTRIBUTE(spufs_id_ops, spufs_id_get, 0, "0x%llx\n")
+
struct tree_descr spufs_dir_contents[] = {
{ "mem", &spufs_mem_fops, 0666, },
{ "regs", &spufs_regs_fops, 0666, },
@@ -1351,5 +1367,6 @@ struct tree_descr spufs_dir_contents[] =
{ "spu_tag_mask", &spufs_spu_tag_mask_ops, 0666, },
{ "event_mask", &spufs_event_mask_ops, 0666, },
{ "srr0", &spufs_srr0_ops, 0666, },
+ { "phys-id", &spufs_id_ops, 0666, },
{},
};
--
^ permalink raw reply
* [patch 10/20] spufs: fix initial state of wbox file
From: arnd @ 2006-06-19 18:33 UTC (permalink / raw)
To: paulus; +Cc: linuxppc-dev, cbe-oss-dev, linux-kernel
In-Reply-To: <20060619183315.653672000@klappe.arndb.de>
The wbox channel count of an spu is now initialized
to four for the saved context. This makes it possible
to write to the mailbox right away without waiting
for the SPE to become scheduled first.
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
---
Index: powerpc.git/arch/powerpc/platforms/cell/spufs/switch.c
===================================================================
--- powerpc.git.orig/arch/powerpc/platforms/cell/spufs/switch.c
+++ powerpc.git/arch/powerpc/platforms/cell/spufs/switch.c
@@ -2129,6 +2129,7 @@ static void init_prob(struct spu_state *
csa->spu_chnlcnt_RW[28] = 1;
csa->spu_chnlcnt_RW[30] = 1;
csa->prob.spu_runcntl_RW = SPU_RUNCNTL_STOP;
+ csa->prob.mb_stat_R = 0x000400;
}
static void init_priv1(struct spu_state *csa)
--
^ permalink raw reply
* [patch 11/20] spufs: use kzalloc in create_spu
From: arnd @ 2006-06-19 18:33 UTC (permalink / raw)
To: paulus; +Cc: linuxppc-dev, cbe-oss-dev, linux-kernel
In-Reply-To: <20060619183315.653672000@klappe.arndb.de>
From: Jeremy Kerr <jk@ozlabs.org>
Clean up create_spu() a little by using kzalloc instead of kmalloc +
assignments.
Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
---
arch/powerpc/platforms/cell/spu_base.c | 17 +----------------
1 file changed, 1 insertion(+), 16 deletions(-)
Index: powerpc.git/arch/powerpc/platforms/cell/spu_base.c
===================================================================
--- powerpc.git.orig/arch/powerpc/platforms/cell/spu_base.c
+++ powerpc.git/arch/powerpc/platforms/cell/spu_base.c
@@ -701,7 +701,7 @@ static int __init create_spu(struct devi
static int number;
ret = -ENOMEM;
- spu = kmalloc(sizeof (*spu), GFP_KERNEL);
+ spu = kzalloc(sizeof (*spu), GFP_KERNEL);
if (!spu)
goto out;
@@ -713,28 +713,11 @@ static int __init create_spu(struct devi
spu->nid = of_node_to_nid(spe);
if (spu->nid == -1)
spu->nid = 0;
-
- spu->stop_code = 0;
- spu->slb_replace = 0;
- spu->mm = NULL;
- spu->ctx = NULL;
- spu->rq = NULL;
- spu->pid = 0;
- spu->class_0_pending = 0;
- spu->flags = 0UL;
- spu->dar = 0UL;
- spu->dsisr = 0UL;
spin_lock_init(&spu->register_lock);
-
spu_mfc_sdr_set(spu, mfspr(SPRN_SDR1));
spu_mfc_sr1_set(spu, 0x33);
-
- spu->ibox_callback = NULL;
- spu->wbox_callback = NULL;
- spu->stop_callback = NULL;
- spu->mfc_callback = NULL;
-
mutex_lock(&spu_mutex);
+
spu->number = number++;
ret = spu_request_irqs(spu);
if (ret)
--
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox