* Re: [2.6 patch] don't allow users to set CONFIG_BROKEN=y
From: Jesper Juhl @ 2005-12-13 18:59 UTC (permalink / raw)
To: Geert Uytterhoeven
Cc: tony.luck, linux-ia64, grundler, matthew,
Linux Kernel Development, Adrian Bunk, Linux/PPC Development,
lethal, linux-mtd, dwmw2, kkojima, parisc-linux
In-Reply-To: <Pine.LNX.4.62.0512131926530.17990@pademelon.sonytel.be>
On 12/13/05, Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> On Tue, 13 Dec 2005, Adrian Bunk wrote:
> > Do not allow people to create configurations with CONFIG_BROKEN=3Dy.
> >
> > The sole reason for CONFIG_BROKEN=3Dy would be if you are working on
> > fixing a broken driver, but in this case editing the Kconfig file is
> > trivial.
> >
> > Never ever should a user enable CONFIG_BROKEN.
> ^^^^
> OK, a user, not an expert. Let's assume users don't enable EXPERIMENTAL.
>
> But I'd like to at least have the possibility to enable broken drivers, e=
ven if
> it's just for compile regression tests.
>
I agree, and it's very convenient to be able to enable it in
menuconfig etc. Perhaps CONFIG_BROKEN should just be moved to kernel
hacking instead...?
--
Jesper Juhl <jesper.juhl@gmail.com>
Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please http://www.expita.com/nomime.html
^ permalink raw reply
* Re: [2.6 patch] don't allow users to set CONFIG_BROKEN=y
From: Adrian Bunk @ 2005-12-13 18:51 UTC (permalink / raw)
To: Geert Uytterhoeven
Cc: tony.luck, linux-ia64, grundler, matthew,
Linux Kernel Development, Linux/PPC Development, lethal,
linux-mtd, dwmw2, kkojima, parisc-linux
In-Reply-To: <Pine.LNX.4.62.0512131926530.17990@pademelon.sonytel.be>
On Tue, Dec 13, 2005 at 07:28:41PM +0100, Geert Uytterhoeven wrote:
> On Tue, 13 Dec 2005, Adrian Bunk wrote:
> > Do not allow people to create configurations with CONFIG_BROKEN=y.
> >
> > The sole reason for CONFIG_BROKEN=y would be if you are working on
> > fixing a broken driver, but in this case editing the Kconfig file is
> > trivial.
> >
> > Never ever should a user enable CONFIG_BROKEN.
> ^^^^
> OK, a user, not an expert. Let's assume users don't enable EXPERIMENTAL.
Let's assume users don't have any hardware (e.g. several SATA drivers)
that requires EXPERIMENTAL...
> But I'd like to at least have the possibility to enable broken drivers, even if
> it's just for compile regression tests.
If a kernel developer really wants to enable BROKEN drivers, my patch
still gives them the possibility to do so with a trivial edit of
init/Kconfig.
> Gr{oetje,eeting}s,
>
> Geert
cu
Adrian
--
"Is there not promise of rain?" Ling Tan asked suddenly out
of the darkness. There had been need of rain for many days.
"Only a promise," Lao Er said.
Pearl S. Buck - Dragon Seed
^ permalink raw reply
* Re: [2.6 patch] don't allow users to set CONFIG_BROKEN=y
From: Geert Uytterhoeven @ 2005-12-13 18:28 UTC (permalink / raw)
To: Adrian Bunk
Cc: tony.luck, linux-ia64, grundler, matthew,
Linux Kernel Development, Linux/PPC Development, lethal,
linux-mtd, dwmw2, kkojima, parisc-linux
In-Reply-To: <20051213180551.GN23349@stusta.de>
On Tue, 13 Dec 2005, Adrian Bunk wrote:
> Do not allow people to create configurations with CONFIG_BROKEN=y.
>
> The sole reason for CONFIG_BROKEN=y would be if you are working on
> fixing a broken driver, but in this case editing the Kconfig file is
> trivial.
>
> Never ever should a user enable CONFIG_BROKEN.
^^^^
OK, a user, not an expert. Let's assume users don't enable EXPERIMENTAL.
But I'd like to at least have the possibility to enable broken drivers, even if
it's just for compile regression tests.
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
^ permalink raw reply
* MPC5200 RTC
From: bennett78 @ 2005-12-13 18:10 UTC (permalink / raw)
To: Wolfgang Denk; +Cc: linuxppc-embedded
In-Reply-To: <20051207195112.06457353A66@atlas.denx.de>
[-- Attachment #1.1: Type: text/plain, Size: 1147 bytes --]
* The /dev/rtc interface will block on reads until an interrupt
* has been received. If a RTC interrupt has already happened,
* it will output an unsigned long and then block.
Paul Gortmaker's rtctest (included below) works really well on my x86
machine (driver/char/rtc.c) as a nice way for a user process to wait on
ticks
of a programable frequency (via ioctl).
Under Denx linuxppc_2_4_devel MPC5200 I've tried:
driver/char/rtc.c - kernel panics
seems to need a RTC_IRQ to be defined
include/linux/mc14681rtc.h mentions #define RTC_IRQ 8 ?
CONFIG_PPC_RTC also panics
driver/char/genrtc.c (with #define CONFIG_GEN_RTC_X)
sort of works but I get the following:
gen_rtc_interrupt: system delaying clock ticks?
The internal MPC5200 RTC doesn't seem supported.
I see a lot of references to external RTCs, probably because
the standby power for the 5200 is high. I would like to
use a timer as the RTC interrupt and add an external RTC
later as a time reference (none on Lite5200 dev brd).
How best can I assign a GPTx as an interrupt? other than re-write
genrtc.c?
thanks,
Frank Bennett
*//*
[-- Attachment #1.2: Type: text/html, Size: 1820 bytes --]
[-- Attachment #2: rtctest.c --]
[-- Type: text/x-c, Size: 4562 bytes --]
/*
* Real Time Clock Driver Test/Example Program
*
* Compile with:
* gcc -s -Wall -Wstrict-prototypes rtctest.c -o rtctest
*
* Copyright (C) 1996, Paul Gortmaker.
*
* Released under the GNU General Public License, version 2,
* included herein by reference.
*
*/
#define spinlock_t int
#include <stdio.h>
#include <linux/mc146818rtc.h>
#include <sys/ioctl.h>
#include <sys/time.h>
#include <sys/types.h>
#include <fcntl.h>
#include <unistd.h>
#include <errno.h>
void main(void) {
int i, fd, retval, irqcount = 0;
unsigned long tmp, data;
struct rtc_time rtc_tm;
fd = open ("/dev/rtc", O_RDONLY);
if (fd == -1) {
perror("/dev/rtc");
exit(errno);
}
fprintf(stderr, "\n\t\t\tRTC Driver Test Example.\n\n");
/* Turn on update interrupts (one per second) */
retval = ioctl(fd, RTC_UIE_ON, 0);
if (retval == -1) {
perror("ioctl");
exit(errno);
}
fprintf(stderr, "Counting 5 update (1/sec) interrupts from reading /dev/rtc:");
fflush(stderr);
for (i=1; i<6; i++) {
/* This read will block */
retval = read(fd, &data, sizeof(unsigned long));
if (retval == -1) {
perror("read");
exit(errno);
}
fprintf(stderr, " %d",i);
fflush(stderr);
irqcount++;
}
fprintf(stderr, "\nAgain, from using select(2) on /dev/rtc:");
fflush(stderr);
for (i=1; i<6; i++) {
struct timeval tv = {5, 0}; /* 5 second timeout on select */
fd_set readfds;
FD_ZERO(&readfds);
FD_SET(fd, &readfds);
/* The select will wait until an RTC interrupt happens. */
retval = select(fd+1, &readfds, NULL, NULL, &tv);
if (retval == -1) {
perror("select");
exit(errno);
}
/* This read won't block unlike the select-less case above. */
retval = read(fd, &data, sizeof(unsigned long));
if (retval == -1) {
perror("read");
exit(errno);
}
fprintf(stderr, " %d",i);
fflush(stderr);
irqcount++;
}
/* Turn off update interrupts */
retval = ioctl(fd, RTC_UIE_OFF, 0);
if (retval == -1) {
perror("ioctl");
exit(errno);
}
/* Read the RTC time/date */
retval = ioctl(fd, RTC_RD_TIME, &rtc_tm);
if (retval == -1) {
perror("ioctl");
exit(errno);
}
fprintf(stderr, "\n\nCurrent RTC date/time is %d-%d-%d, %02d:%02d:%02d.\n",
rtc_tm.tm_mday, rtc_tm.tm_mon + 1, rtc_tm.tm_year + 1900,
rtc_tm.tm_hour, rtc_tm.tm_min, rtc_tm.tm_sec);
/* Set the alarm to 5 sec in the future, and check for rollover */
rtc_tm.tm_sec += 5;
if (rtc_tm.tm_sec >= 60) {
rtc_tm.tm_sec %= 60;
rtc_tm.tm_min++;
}
if (rtc_tm.tm_min == 60) {
rtc_tm.tm_min = 0;
rtc_tm.tm_hour++;
}
if (rtc_tm.tm_hour == 24)
rtc_tm.tm_hour = 0;
retval = ioctl(fd, RTC_ALM_SET, &rtc_tm);
if (retval == -1) {
perror("ioctl");
exit(errno);
}
/* Read the current alarm settings */
retval = ioctl(fd, RTC_ALM_READ, &rtc_tm);
if (retval == -1) {
perror("ioctl");
exit(errno);
}
fprintf(stderr, "Alarm time now set to %02d:%02d:%02d.\n",
rtc_tm.tm_hour, rtc_tm.tm_min, rtc_tm.tm_sec);
/* Enable alarm interrupts */
retval = ioctl(fd, RTC_AIE_ON, 0);
if (retval == -1) {
perror("ioctl");
exit(errno);
}
fprintf(stderr, "Waiting 5 seconds for alarm...");
fflush(stderr);
/* This blocks until the alarm ring causes an interrupt */
retval = read(fd, &data, sizeof(unsigned long));
if (retval == -1) {
perror("read");
exit(errno);
}
irqcount++;
fprintf(stderr, " okay. Alarm rang.\n");
/* Disable alarm interrupts */
retval = ioctl(fd, RTC_AIE_OFF, 0);
if (retval == -1) {
perror("ioctl");
exit(errno);
}
/* Read periodic IRQ rate */
retval = ioctl(fd, RTC_IRQP_READ, &tmp);
if (retval == -1) {
perror("ioctl");
exit(errno);
}
fprintf(stderr, "\nPeriodic IRQ rate was %ldHz.\n", tmp);
fprintf(stderr, "Counting 20 interrupts at:");
fflush(stderr);
/* The frequencies 128Hz, 256Hz, ... 8192Hz are only allowed for root. */
for (tmp=2; tmp<=64; tmp*=2) {
retval = ioctl(fd, RTC_IRQP_SET, tmp);
if (retval == -1) {
perror("ioctl");
exit(errno);
}
fprintf(stderr, "\n%ldHz:\t", tmp);
fflush(stderr);
/* Enable periodic interrupts */
retval = ioctl(fd, RTC_PIE_ON, 0);
if (retval == -1) {
perror("ioctl");
exit(errno);
}
for (i=1; i<21; i++) {
/* This blocks */
retval = read(fd, &data, sizeof(unsigned long));
if (retval == -1) {
perror("read");
exit(errno);
}
fprintf(stderr, " %d",i);
fflush(stderr);
irqcount++;
}
/* Disable periodic interrupts */
retval = ioctl(fd, RTC_PIE_OFF, 0);
if (retval == -1) {
perror("ioctl");
exit(errno);
}
}
fprintf(stderr, "\n\n\t\t\t *** Test complete ***\n");
fprintf(stderr, "\nTyping \"cat /proc/interrupts\" will show %d more events on IRQ 8.\n\n",
irqcount);
close(fd);
} /* end main */
^ permalink raw reply
* [2.6 patch] don't allow users to set CONFIG_BROKEN=y
From: Adrian Bunk @ 2005-12-13 18:05 UTC (permalink / raw)
To: Simon Richter, linux-kernel, tony.luck, linux-ia64, matthew,
grundler, parisc-linux, paulus, linuxppc-dev, lethal, kkojima,
dwmw2, linux-mtd
In-Reply-To: <20051213173112.GA24094@flint.arm.linux.org.uk>
On Tue, Dec 13, 2005 at 05:31:12PM +0000, Russell King wrote:
> On Tue, Dec 13, 2005 at 03:00:01PM +0100, Adrian Bunk wrote:
> > defconfig files are virtually never a configuration for the kernel they
> > are shipped with since they aren't updated every time some configuration
> > option is changed.
> >
> > Consider a defconfig with CONFIG_BROKEN=n, and a driver that is enabled
> > in this defconfig gets for some reason marked as broken in the Kconfig
> > file - this will give exactly the same result as the one you describe.
>
> Adrian,
Hi Russell,
> The defconfig files in arch/arm/configs are for platform configurations
> and are provided by the platform maintainers as a _working_ configuration
> for their platform. They're not "defconfigs". They got called
> "defconfigs" as a result of the kbuild "cleanups". Please don't confuse
> them as such.
>
> If, in order to have a working platform configuration, they deem that
> CONFIG_BROKEN must be enabled, then that's the way it is.
if a working platform configuration configuration requires
CONFIG_BROKEN=y, the problem is a bug that should be fixed properly.
We are talking about a class of bugs that can usually be easily fixed if
reported - so why aren't they reported?
The MTD_SHARP case is a good example, because otherwise I might have
soon sent a patch that would have removed this driver with the rationale
"both marked as obsolete and BROKEN can clearly be removed".
> Therefore, I request that either you leave the ARM platform configurations
> well alone, or follow the advice I've given so that we can _properly_
> assess the impact of your changes.
Unless someone can tell me a valid case for enabling BROKEN that does
both create a working configuration and not hide real issues it seems
the approch below might be the way to go.
Yes, you might dislike this at the first sight.
But if you consider that although this might result in a short-term
breakage of some configurations, this will also result in proper bug
reports and fixing of the wrong BROKEN dependency bugs, I hope you agree
that this will actually improve the situation.
> Thanks.
cu
Adrian
<-- snip -->
Do not allow people to create configurations with CONFIG_BROKEN=y.
The sole reason for CONFIG_BROKEN=y would be if you are working on
fixing a broken driver, but in this case editing the Kconfig file is
trivial.
Never ever should a user enable CONFIG_BROKEN.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
--- linux-2.6.15-rc5-mm2-full/init/Kconfig.old 2005-12-13 18:48:40.000000000 +0100
+++ linux-2.6.15-rc5-mm2-full/init/Kconfig 2005-12-13 18:48:52.000000000 +0100
@@ -31,19 +31,8 @@
you say Y here, you will be offered the choice of using features or
drivers that are currently considered to be in the alpha-test phase.
-config CLEAN_COMPILE
- bool "Select only drivers expected to compile cleanly" if EXPERIMENTAL
- default y
- help
- Select this option if you don't even want to see the option
- to configure known-broken drivers.
-
- If unsure, say Y
-
config BROKEN
bool
- depends on !CLEAN_COMPILE
- default y
config BROKEN_ON_SMP
bool
^ permalink raw reply
* Re: ibm emac driver update
From: Simon Haynes @ 2005-12-13 18:44 UTC (permalink / raw)
To: Eugene Surovegin; +Cc: linuxppc-embedded
In-Reply-To: <20051212181457.GA30505@gate.ebshome.net>
On Monday 12 December 2005 18:14, Eugene Surovegin wrote:
> On Mon, Dec 12, 2005 at 05:26:34PM +0000, Simon Haynes wrote:
> > I am running a Ebony board with linux-2.6.12-rc1 and I amd having
> > intermittent network problems. Having searched the archives it would seem
> > that the driver for the IBM EMAC is buggy ?
> >
> > I have tried to patch my kernel with
> >
> > http://kernel.ebsname.net/emac/4xx_napi_emac.diff
> >
> > I believe this is almost a total driver re-write. Is this correct ?
> >
> > After having a couple of problems applying the patch, Kconfig and
> > ibm_emac_core.c. I fixed up the original code and applied the patch.
> >
> > On compilation the build fails with missing references to kzalloc,
> > which I guess I could easily fix with kmalloc and memset.
> >
> > However there is also an undefined field, dcr_base, in the
> > ocp_func_mal_data structure. This structure appears to be defined in
> > include/asm/ibm_ocp.h and this file does not appear to be changed by the
> > patch. I guess this patch is for a later kernel.
> >
> > Can anyone point me in the direction of a suitable patch for this kernel
> > ?
>
> I'm not sure about 2.6.12-rc1, but you can try GIT tree/patches at
> http://kernel.ebshome.net.
>
> Hint: you can assume that in addition to emac_3-2.6-28.diff, there are
> also emac_3-2.6-NN.diff (NN < 28) on my website which were made for
> earlier kernel versions :).
>
> Alternatively, you can backport all relevant stuff from current 2.6 or
> you can pay someone to do this for you.
I have eventually got git going. I am not familiar with it but following a
tutorial I did
git clone rsync://kernel.ebshome.net/linux-2.6-emac.git
cd linux-2.6-emac
git checkout
I then ran find looking for a patches dir and could not find one.
Could you please tell me how to locate the emac_3-2.6-NN.diff files.
Cheers
Simon
^ permalink raw reply
* Re: [2.6 patch] defconfig's shouldn't set CONFIG_BROKEN=y
From: Russell King @ 2005-12-13 17:31 UTC (permalink / raw)
To: Adrian Bunk
Cc: tony.luck, linux-ia64, grundler, matthew, linux-kernel,
linuxppc-dev, lethal, linux-mtd, dwmw2, kkojima, parisc-linux
In-Reply-To: <20051213140001.GG23349@stusta.de>
On Tue, Dec 13, 2005 at 03:00:01PM +0100, Adrian Bunk wrote:
> defconfig files are virtually never a configuration for the kernel they
> are shipped with since they aren't updated every time some configuration
> option is changed.
>
> Consider a defconfig with CONFIG_BROKEN=n, and a driver that is enabled
> in this defconfig gets for some reason marked as broken in the Kconfig
> file - this will give exactly the same result as the one you describe.
Adrian,
The defconfig files in arch/arm/configs are for platform configurations
and are provided by the platform maintainers as a _working_ configuration
for their platform. They're not "defconfigs". They got called
"defconfigs" as a result of the kbuild "cleanups". Please don't confuse
them as such.
If, in order to have a working platform configuration, they deem that
CONFIG_BROKEN must be enabled, then that's the way it is.
Therefore, I request that either you leave the ARM platform configurations
well alone, or follow the advice I've given so that we can _properly_
assess the impact of your changes.
Thanks.
--
Russell King
Linux kernel 2.6 ARM Linux - http://www.arm.linux.org.uk/
maintainer of: 2.6 Serial core
^ permalink raw reply
* how to define BASE_BAUD
From: HappyPhot @ 2005-12-13 17:19 UTC (permalink / raw)
To: linuxppc-embedded
Hi,
Can anybody tell me how to define the "BASE_BAUD" value
for the "struct serial_state rs_table[]" ?
It seems has different value for each platform !!
thank you,
/HappyPhot
^ permalink raw reply
* Re: [PATCH] powerpc: Update MPIC workarounds
From: Olof Johansson @ 2005-12-13 16:58 UTC (permalink / raw)
To: Benjamin Herrenschmidt; +Cc: linuxppc64-dev, linuxppc-dev list
In-Reply-To: <1134457469.6989.189.camel@gaston>
On Tue, Dec 13, 2005 at 06:04:29PM +1100, Benjamin Herrenschmidt wrote:
> From: Segher Boessenkool <segher@kernel.crashing.org>
>
> Cleanup the MPIC IO-APIC workarounds, make them a bit more generic,
> smaller and faster.
I really don't like all the hand-coded constants in this code. They're
all over the place, and there's no descriptions of what they are there
for. Lots of hardcoded offsets, etc. Since this is a cleanup, wouldn't
it be a good time to use symbolic constands and/or comment them up a
bit?
> Index: linux-work/arch/powerpc/sysdev/mpic.c
> ===================================================================
> --- linux-work.orig/arch/powerpc/sysdev/mpic.c 2005-12-06 16:17:43.000000000 +1100
> +++ linux-work/arch/powerpc/sysdev/mpic.c 2005-12-07 13:30:45.000000000 +1100
> @@ -175,57 +175,57 @@ static inline int mpic_is_ht_interrupt(s
> return mpic->fixups[source_no].base != NULL;
> }
>
> +
> static inline void mpic_apic_end_irq(struct mpic *mpic, unsigned int source_no)
> {
> struct mpic_irq_fixup *fixup = &mpic->fixups[source_no];
> - u32 tmp;
>
> spin_lock(&mpic->fixup_lock);
> - writeb(0x11 + 2 * fixup->irq, fixup->base);
> - tmp = readl(fixup->base + 2);
> - writel(tmp | 0x80000000ul, fixup->base + 2);
> - /* config writes shouldn't be posted but let's be safe ... */
> - (void)readl(fixup->base + 2);
> + writeb(0x11 + 2 * fixup->irq, fixup->base + 2);
> + writel(fixup->data, fixup->base + 4);
This seems like a functional change: Previous code wrote at base, new at
base+2?
-Olof
^ permalink raw reply
* Re: chrdevs[] explained
From: Kumar Gala @ 2005-12-13 14:40 UTC (permalink / raw)
To: Gabe.Johnson; +Cc: linuxppc-embedded
In-Reply-To: <OFDEA31089.DE015378-ON862570D5.005B4F12-862570D5.005B7940@microware.com>
On Dec 12, 2005, at 10:38 AM, Gabe.Johnson@radisys.com wrote:
>
> Could someone explain to me the chrdevs array found in fs/
> char_dev.c how does that get populated and with what does it get
> populated?
This is a pretty generic kernel question, I would suggest posting to
linux-kernel as it not really powerpc specific.
- kumar
^ permalink raw reply
* Re: [2.6 patch] defconfig's shouldn't set CONFIG_BROKEN=y
From: Adrian Bunk @ 2005-12-13 14:00 UTC (permalink / raw)
To: Simon Richter
Cc: tony.luck, linux-ia64, grundler, matthew, linux-kernel,
linuxppc-dev, lethal, linux-mtd, dwmw2, kkojima, parisc-linux
In-Reply-To: <439ECDCC.80707@hogyros.de>
On Tue, Dec 13, 2005 at 02:34:04PM +0100, Simon Richter wrote:
> Hi,
Hi Simon,
> Adrian Bunk wrote:
>
> >>It's a problem introduced by your patch because the resulting defconfig
> >>file becomes _wrong_ by your change, and other changes in the defconfig
> >>are thereby hidden.
> >>...
>
> >No, CONFIG_BROKEN=y in a defconfig file is a bug.
>
> Indeed, but that's not the point. A defconfig file should be the result
> of running one of the various configuration targets; yours are
> hand-patched. If you run the defconfig target, it will copy the config
> file and run oldconfig, thus resulting in a different configuration file
> (because options may now be gone and hence disabled) than what was in
> the defconfig, and thus people may come to the wrong conclusion that if
> a driver is enabled in a defconfig file, it will be built.
defconfig files are virtually never a configuration for the kernel they
are shipped with since they aren't updated every time some configuration
option is changed.
Consider a defconfig with CONFIG_BROKEN=n, and a driver that is enabled
in this defconfig gets for some reason marked as broken in the Kconfig
file - this will give exactly the same result as the one you describe.
> Simon
cu
Adrian
--
"Is there not promise of rain?" Ling Tan asked suddenly out
of the darkness. There had been need of rain for many days.
"Only a promise," Lao Er said.
Pearl S. Buck - Dragon Seed
^ permalink raw reply
* Re: ibm emac driver update
From: Simon Haynes @ 2005-12-13 14:35 UTC (permalink / raw)
To: linuxppc-embedded
In-Reply-To: <20051212181457.GA30505@gate.ebshome.net>
Eugine,
Thanks for the reply. I am not familiar with git and I do not
have it installed on my system. I have looked for rpms and
tried to build git. It seems that I will need to make a lot of changes
to my system to get this going.
In a response from Wolfgang he lists a url which can be used to access
his git tree via a browser. I can not find these patch files on his git tree
but I was wondering if you made the files available in a similar way.
Alternatively can you mail the file to me ?
Or can you tell me if the files are somewhere on
http://www.denx.de/cgi-bin/gitweb.cgi?p=linux-2.6-denx.git
If not I guess I will just have to get all the bits together to run
git and learn how to drive it.
Many Thanks
Simon.
On Monday 12 December 2005 18:14, Eugene Surovegin wrote:
> On Mon, Dec 12, 2005 at 05:26:34PM +0000, Simon Haynes wrote:
> > I am running a Ebony board with linux-2.6.12-rc1 and I amd having
> > intermittent network problems. Having searched the archives it would seem
> > that the driver for the IBM EMAC is buggy ?
> >
> > I have tried to patch my kernel with
> >
> > http://kernel.ebsname.net/emac/4xx_napi_emac.diff
> >
> > I believe this is almost a total driver re-write. Is this correct ?
> >
> > After having a couple of problems applying the patch, Kconfig and
> > ibm_emac_core.c. I fixed up the original code and applied the patch.
> >
> > On compilation the build fails with missing references to kzalloc,
> > which I guess I could easily fix with kmalloc and memset.
> >
> > However there is also an undefined field, dcr_base, in the
> > ocp_func_mal_data structure. This structure appears to be defined in
> > include/asm/ibm_ocp.h and this file does not appear to be changed by the
> > patch. I guess this patch is for a later kernel.
> >
> > Can anyone point me in the direction of a suitable patch for this kernel
> > ?
>
> I'm not sure about 2.6.12-rc1, but you can try GIT tree/patches at
> http://kernel.ebshome.net.
>
> Hint: you can assume that in addition to emac_3-2.6-28.diff, there are
> also emac_3-2.6-NN.diff (NN < 28) on my website which were made for
> earlier kernel versions :).
>
> Alternatively, you can backport all relevant stuff from current 2.6 or
> you can pay someone to do this for you.
^ permalink raw reply
* Re: [2.6 patch] defconfig's shouldn't set CONFIG_BROKEN=y
From: Simon Richter @ 2005-12-13 13:34 UTC (permalink / raw)
To: Adrian Bunk
Cc: tony.luck, linux-ia64, grundler, matthew, linux-kernel,
linuxppc-dev, lethal, linux-mtd, dwmw2, kkojima, parisc-linux
In-Reply-To: <20051213001028.GS23349@stusta.de>
[-- Attachment #1: Type: text/plain, Size: 752 bytes --]
Hi,
Adrian Bunk wrote:
>>It's a problem introduced by your patch because the resulting defconfig
>>file becomes _wrong_ by your change, and other changes in the defconfig
>>are thereby hidden.
>>...
> No, CONFIG_BROKEN=y in a defconfig file is a bug.
Indeed, but that's not the point. A defconfig file should be the result
of running one of the various configuration targets; yours are
hand-patched. If you run the defconfig target, it will copy the config
file and run oldconfig, thus resulting in a different configuration file
(because options may now be gone and hence disabled) than what was in
the defconfig, and thus people may come to the wrong conclusion that if
a driver is enabled in a defconfig file, it will be built.
Simon
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 307 bytes --]
^ permalink raw reply
* Re: HZ manipulation
From: Wojciech Kromer @ 2005-12-13 13:36 UTC (permalink / raw)
To: linuxppc-embedded
In-Reply-To: <439E86DC.4010901@dgt.com.pl>
Found it!
Sorry, It was my error.
I have few #ifdef-s to determine HZ value depend on selected board,
but in some (for example itimer.c and timer.c) files <linux/config.h> is
NOT included
before <asm/param.h> is included sowhere in the dark....
So some code was compiled with default value
^ permalink raw reply
* HZ manipulation
From: Wojciech Kromer @ 2005-12-13 8:31 UTC (permalink / raw)
To: linuxppc-embedded
[-- Attachment #1: Type: text/plain, Size: 1771 bytes --]
I have 2.4.25 kernel from denx.de and latest uClibc running on mpc8248.
I'm trying to change HZ value, but still have strange results with
simple test
Recompiling whole uClibc does not change the results.
What could be wrong?
with HZ==1000:
#
alarm_test
alarm fired after
0.099092s
slept 0.999769s
with HZ==500:
#
alarm_test
alarm fired after
0.199622s
slept 0.999761s
with HZ==100:
# alarm_test
alarm fired after 0.999693s
slept 0.999760s
- here is procesor id :
... by kernel:
# cat /proc/cpuinfo
processor : 0
cpu : 8280
core clock : 400 MHz
CPM clock : 200 MHz
bus clock : 100 MHz
revision : 32.20 (pvr 8082 2014)
bogomips : 266.24
... and by u-boot:
MPC8248 Reset Status: External Soft, External
Hard
MPC8248 Clock
Configuration
- Bus-to-Core Mult 4x, VCO Div 2, 60x Bus Freq 25-75 , Core Freq
100-300
- dfbrg 1, corecnf 0x1a, busdf 3, cpmdf 1, plldf 0, pllmf
3
- vco_out 400000000, scc_clk 100000000, brg_clk
25000000
- cpu_clk 400000000, cpm_clk 200000000, bus_clk
100000000
- pci_clk
66666666
CPU: MPC8248 (HiP7 Rev 14, Mask 1.0 1K50M) at 400 MHz
- test code included
[-- Attachment #2: alarm_test.c --]
[-- Type: text/x-csrc, Size: 908 bytes --]
#include <stdio.h>
#include <signal.h>
#include <unistd.h>
#include <sys/time.h>
static void sighand(int sig)
{
}
int main()
{
struct timeval tv1, tv2;
sigset_t set;
sigfillset(&set);
sigdelset(&set, SIGALRM);
sigdelset(&set, SIGINT);
signal(SIGALRM, sighand);
gettimeofday(&tv1, NULL);
alarm(1);
sigsuspend(&set);
gettimeofday(&tv2, NULL);
tv1.tv_sec = tv2.tv_sec - tv1.tv_sec - (tv2.tv_usec < tv1.tv_usec);
tv1.tv_usec = tv2.tv_usec - tv1.tv_usec + (tv2.tv_usec < tv1.tv_usec) * 1000000;
printf("alarm fired after %ld.%06lds\n",tv1.tv_sec, tv1.tv_usec);
//--------------------
gettimeofday(&tv1, NULL);
sleep(1);
gettimeofday(&tv2, NULL);
tv1.tv_sec = tv2.tv_sec - tv1.tv_sec - (tv2.tv_usec < tv1.tv_usec);
tv1.tv_usec = tv2.tv_usec - tv1.tv_usec + (tv2.tv_usec < tv1.tv_usec) * 1000000;
printf("slept %ld.%06lds\n",tv1.tv_sec, tv1.tv_usec);
return 0;
}
^ permalink raw reply
* Re: Can we compile them with eldk and debug them with BDI2000?
From: Wolfgang Denk @ 2005-12-13 8:18 UTC (permalink / raw)
To: zengshuai; +Cc: ppc
In-Reply-To: <31064960.1134445714325.JavaMail.postfix@mx3.mail.sohu.com>
In message <31064960.1134445714325.JavaMail.postfix@mx3.mail.sohu.com> you wrote:
> thanks a lot.Would you give me more inf,please?
I'm afraid this is off topic on this list.
> There were a test.c.
> When we use "ppc_6xx-gcc -o test test.c",we can get a "test" which can running in Linux.
> How can we get a "test" which can running in bareboard and be debuged using the BDI2000?
> Would you give a flag example?
See the U-Boot source code. U-Boot itself is just such a bare-silicon
application program (a pretty complex one, though). Also see the
examples/ directory which contains code for several simple standalone
applications within the U-Boot context.
Note that you will have to understand how the linker works, and know
exactly what you are doing.
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
Four thousand throats may be cut in one night by a running man.
-- Klingon Soldier, "Day of the Dove", stardate unknown
^ permalink raw reply
* RE: MPC8245 with its internal UART
From: Heiko Schocher @ 2005-12-13 8:01 UTC (permalink / raw)
To: HappyPhot; +Cc: linuxppc-embedded
In-Reply-To: <003201c5ffb9$fe0d6310$0760120a@photon>
Hello,
On Tuesday, December 13, 2005 8:51 AM HappyPhot wrote:
> Hi,
> I get this one "linuxppc_2_4_devel-2005-10-25-1440.tar.bz2"
> and try "make CPC45_defconfig". But it shows "No rule to make target
> ...." Did I do something wrong ?
Try "make CPC45_config"
It works fine on my PC ...
Bets regards,
Heiko
^ permalink raw reply
* Re: MPC8245 with its internal UART
From: HappyPhot @ 2005-12-13 7:50 UTC (permalink / raw)
To: hs, linuxppc-embedded
In-Reply-To: <AHEILKONAKAEJPHNMOPNIEDBCEAA.hs@denx.de>
Hi,
I get this one "linuxppc_2_4_devel-2005-10-25-1440.tar.bz2"
and try "make CPC45_defconfig". But it shows "No rule to make
target ...."
Did I do something wrong ?
thank you,
/HappyPhot
----- Original Message -----
From: "Heiko Schocher" <hs@denx.de>
To: <linuxppc-embedded@ozlabs.org>
Cc: <happyphot@gmail.com>
Sent: Wednesday, December 07, 2005 3:36 PM
Subject: MPC8245 with its internal UART
Hello,
> Who can tell me which flatform is similar to mine ? I've tried
> "make sandpoint_defconfig", but seems not the right one.
If you want to use Linux 2.4.25, you can try CPC45_defconfig
from the Denx Kernel ...
Best regards,
Heiko
^ permalink raw reply
* Re: pq2_find_bridges hangs system
From: Alex BASTOS @ 2005-12-13 8:24 UTC (permalink / raw)
To: Vitaly Bordug; +Cc: linuxppc-embedded list
In-Reply-To: <4390604D.50002@ru.mvista.com>
Vitaly,
It didn't work. So I will check pq2ads_setup_pci to check if
some board specific issue is affected.
I will say you if I find anything
Thanks
Alex.
Citando Vitaly Bordug <vbordug@ru.mvista.com>:
> Alex,
> please probe this. If it will not help, pq2ads_setup_pci should be digged for
> some board specifics.
>
> Also, check if hose->set_cfg_type is 1 as done for ads8272.
>
> -------- Original Message --------
> Subject: [patch 01/34] ppc32: Fix incorrect PCI frequency value
> Date: Thu, 01 Dec 2005 00:51:15 -0800
> From: akpm@osdl.org
> To: torvalds@osdl.org
> CC: akpm@osdl.org, vbordug@ru.mvista.com, galak@kernel.crashing.org
>
>
> From: Vitaly Bordug <vbordug@ru.mvista.com>
>
> The time to wait after deasserting PCI_RST has been counted with incorrect
> value - this patch fixes the issue.
>
> Signed-off-by: Vitaly Bordug <vbordug@ru.mvista.com>
> Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
> Signed-off-by: Andrew Morton <akpm@osdl.org>
> ---
>
> arch/ppc/syslib/m82xx_pci.c | 3 ++-
> 1 files changed, 2 insertions(+), 1 deletion(-)
>
> diff -puN arch/ppc/syslib/m82xx_pci.c~ppc32-fix-incorrect-pci-frequency-value
> arch/ppc/syslib/m82xx_pci.c
> ---
>
devel/arch/ppc/syslib/m82xx_pci.c~ppc32-fix-incorrect-pci-frequency-value 2005-11-30
> 23:58:00.000000000 -0800
> +++ devel-akpm/arch/ppc/syslib/m82xx_pci.c 2005-11-30 23:58:00.000000000
> -0800
> @@ -248,7 +248,8 @@ pq2ads_setup_pci(struct pci_controller *
> pci_div = ( (sccr & SCCR_PCI_MODCK) ? 2 : 1) *
> ( ( (sccr & SCCR_PCIDF_MSK) >> SCCR_PCIDF_SHIFT) + 1);
> freq = (uint)((2*binfo->bi_cpmfreq)/(pci_div));
> - time = (int)666666/freq;
> + time = (int)66666666/freq;
> +
> /* due to PCI Local Bus spec, some devices needs to wait such a long
> time after RST deassertion. More specifically, 0.508s for 66MHz & twice
> more for 33 */
> printk("%s: The PCI bus is %d Mhz.\nWaiting %s after deasserting
> RST...\n",__FILE__,freq,
> _
>
>
>
> --
> Sincerely,
> Vitaly
>
^ permalink raw reply
* [PATCH] powerpc: Experimental support for new G5 Macs
From: Benjamin Herrenschmidt @ 2005-12-13 7:17 UTC (permalink / raw)
To: Paul Mackerras; +Cc: linuxppc64-dev, linuxppc-dev list
This adds some very basic support for the new machines, including the
Quad G5 (tested), and other new dual core based machines and iMac G5
iSight (untested). This is still experimental ! If you have more than
2Gb of RAM, you MUST boot with mem=2G to limit it as the iommu/DART code
for the new bridge chip is not there yet (hopefully tomorrow). There is
no thermal control yet, there is no proper handing of MSIs, etc.. but it
boots, I have all 4 cores up on my machine.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Index: linux-work/arch/powerpc/platforms/powermac/feature.c
===================================================================
--- linux-work.orig/arch/powerpc/platforms/powermac/feature.c 2005-12-13 17:51:20.000000000 +1100
+++ linux-work/arch/powerpc/platforms/powermac/feature.c 2005-12-13 18:13:11.000000000 +1100
@@ -101,7 +101,8 @@ static const char *macio_names[] =
"Keylargo",
"Pangea",
"Intrepid",
- "K2"
+ "K2",
+ "Shasta",
};
@@ -119,7 +120,7 @@ static const char *macio_names[] =
static struct device_node *uninorth_node;
static u32 __iomem *uninorth_base;
static u32 uninorth_rev;
-static int uninorth_u3;
+static int uninorth_maj;
static void __iomem *u3_ht;
/*
@@ -1399,8 +1400,15 @@ static long g5_fw_enable(struct device_n
static long g5_mpic_enable(struct device_node *node, long param, long value)
{
unsigned long flags;
+ struct device_node *parent = of_get_parent(node);
+ int is_u3;
- if (node->parent == NULL || strcmp(node->parent->name, "u3"))
+ if (parent == NULL)
+ return 0;
+ is_u3 = strcmp(parent->name, "u3") == 0 ||
+ strcmp(parent->name, "u4") == 0;
+ of_node_put(parent);
+ if (!is_u3)
return 0;
LOCK(flags);
@@ -1464,7 +1472,7 @@ static long g5_i2s_enable(struct device_
},
};
- if (macio->type != macio_keylargo2 /* && macio->type != macio_shasta*/)
+ if (macio->type != macio_keylargo2 && macio->type != macio_shasta)
return -ENODEV;
if (strncmp(node->name, "i2s-", 4))
return -ENODEV;
@@ -1473,11 +1481,9 @@ static long g5_i2s_enable(struct device_
case 0:
case 1:
break;
-#if 0
case 2:
if (macio->type == macio_shasta)
break;
-#endif
default:
return -ENODEV;
}
@@ -1508,7 +1514,7 @@ static long g5_reset_cpu(struct device_n
struct device_node *np;
macio = &macio_chips[0];
- if (macio->type != macio_keylargo2)
+ if (macio->type != macio_keylargo2 && macio->type != macio_shasta)
return -ENODEV;
np = find_path_device("/cpus");
@@ -1547,7 +1553,8 @@ static long g5_reset_cpu(struct device_n
*/
void g5_phy_disable_cpu1(void)
{
- UN_OUT(U3_API_PHY_CONFIG_1, 0);
+ if (uninorth_maj == 3)
+ UN_OUT(U3_API_PHY_CONFIG_1, 0);
}
#endif /* CONFIG_POWER4 */
@@ -2462,6 +2469,14 @@ static struct pmac_mb_def pmac_mb_defs[]
PMAC_TYPE_POWERMAC_G5_U3L, g5_features,
0,
},
+ { "PowerMac11,2", "PowerMac G5 Dual Core",
+ PMAC_TYPE_POWERMAC_G5_U3L, g5_features,
+ 0,
+ },
+ { "PowerMac12,1", "iMac G5 (iSight)",
+ PMAC_TYPE_POWERMAC_G5_U3L, g5_features,
+ 0,
+ },
{ "RackMac3,1", "XServe G5",
PMAC_TYPE_XSERVE_G5, g5_features,
0,
@@ -2574,6 +2589,11 @@ static int __init probe_motherboard(void
pmac_mb.model_name = "Unknown K2-based";
pmac_mb.features = g5_features;
break;
+ case macio_shasta:
+ pmac_mb.model_id = PMAC_TYPE_UNKNOWN_SHASTA;
+ pmac_mb.model_name = "Unknown Shasta-based";
+ pmac_mb.features = g5_features;
+ break;
#endif /* CONFIG_POWER4 */
default:
return -ENODEV;
@@ -2651,7 +2671,12 @@ static void __init probe_uninorth(void)
/* Locate G5 u3 */
if (uninorth_node == NULL) {
uninorth_node = of_find_node_by_name(NULL, "u3");
- uninorth_u3 = 1;
+ uninorth_maj = 3;
+ }
+ /* Locate G5 u4 */
+ if (uninorth_node == NULL) {
+ uninorth_node = of_find_node_by_name(NULL, "u4");
+ uninorth_maj = 4;
}
if (uninorth_node == NULL)
return;
@@ -2664,12 +2689,13 @@ static void __init probe_uninorth(void)
return;
uninorth_base = ioremap(address, 0x40000);
uninorth_rev = in_be32(UN_REG(UNI_N_VERSION));
- if (uninorth_u3)
+ if (uninorth_maj == 3 || uninorth_maj == 4)
u3_ht = ioremap(address + U3_HT_CONFIG_BASE, 0x1000);
- printk(KERN_INFO "Found %s memory controller & host bridge,"
- " revision: %d\n", uninorth_u3 ? "U3" : "UniNorth",
- uninorth_rev);
+ printk(KERN_INFO "Found %s memory controller & host bridge"
+ " @ 0x%08x revision: 0x%02x\n", uninorth_maj == 3 ? "U3" :
+ uninorth_maj == 4 ? "U4" : "UniNorth",
+ (unsigned int)address, uninorth_rev);
printk(KERN_INFO "Mapped at 0x%08lx\n", (unsigned long)uninorth_base);
/* Set the arbitrer QAck delay according to what Apple does
@@ -2677,7 +2703,8 @@ static void __init probe_uninorth(void)
if (uninorth_rev < 0x11) {
actrl = UN_IN(UNI_N_ARB_CTRL) & ~UNI_N_ARB_CTRL_QACK_DELAY_MASK;
actrl |= ((uninorth_rev < 3) ? UNI_N_ARB_CTRL_QACK_DELAY105 :
- UNI_N_ARB_CTRL_QACK_DELAY) << UNI_N_ARB_CTRL_QACK_DELAY_SHIFT;
+ UNI_N_ARB_CTRL_QACK_DELAY) <<
+ UNI_N_ARB_CTRL_QACK_DELAY_SHIFT;
UN_OUT(UNI_N_ARB_CTRL, actrl);
}
@@ -2685,7 +2712,8 @@ static void __init probe_uninorth(void)
* revs 1.5 to 2.O and Pangea. Seem to toggle the UniN Maxbus/PCI
* memory timeout
*/
- if ((uninorth_rev >= 0x11 && uninorth_rev <= 0x24) || uninorth_rev == 0xc0)
+ if ((uninorth_rev >= 0x11 && uninorth_rev <= 0x24) ||
+ uninorth_rev == 0xc0)
UN_OUT(0x2160, UN_IN(0x2160) & 0x00ffffff);
}
@@ -2736,12 +2764,14 @@ static void __init probe_one_macio(const
node->full_name);
return;
}
- if (type == macio_keylargo) {
+ if (type == macio_keylargo || type == macio_keylargo2) {
u32 *did = (u32 *)get_property(node, "device-id", NULL);
if (*did == 0x00000025)
type = macio_pangea;
if (*did == 0x0000003e)
type = macio_intrepid;
+ if (*did == 0x0000004f)
+ type = macio_shasta;
}
macio_chips[i].of_node = node;
macio_chips[i].type = type;
@@ -2840,7 +2870,8 @@ set_initial_features(void)
}
#ifdef CONFIG_POWER4
- if (macio_chips[0].type == macio_keylargo2) {
+ if (macio_chips[0].type == macio_keylargo2 ||
+ macio_chips[0].type == macio_shasta) {
#ifndef CONFIG_SMP
/* On SMP machines running UP, we have the second CPU eating
* bus cycles. We need to take it off the bus. This is done
Index: linux-work/arch/powerpc/platforms/powermac/pic.c
===================================================================
--- linux-work.orig/arch/powerpc/platforms/powermac/pic.c 2005-12-13 17:51:20.000000000 +1100
+++ linux-work/arch/powerpc/platforms/powermac/pic.c 2005-12-13 18:13:11.000000000 +1100
@@ -524,18 +524,56 @@ static void __init pmac_pic_setup_mpic_n
#endif /* defined(CONFIG_XMON) && defined(CONFIG_PPC32) */
}
+static struct mpic * __init pmac_setup_one_mpic(struct device_node *np,
+ int master)
+{
+ unsigned char senses[128];
+ int offset = master ? 0 : 128;
+ int count = master ? 128 : 124;
+ const char *name = master ? " MPIC 1 " : " MPIC 2 ";
+ struct resource r;
+ struct mpic *mpic;
+ unsigned int flags = master ? MPIC_PRIMARY : 0;
+ int rc;
+
+ rc = of_address_to_resource(np, 0, &r);
+ if (rc)
+ return NULL;
+
+ pmac_call_feature(PMAC_FTR_ENABLE_MPIC, np, 0, 0);
+
+ prom_get_irq_senses(senses, offset, offset + count);
+
+ flags |= MPIC_WANTS_RESET;
+ if (get_property(np, "big-endian", NULL))
+ flags |= MPIC_BIG_ENDIAN;
+
+ /* Primary Big Endian means HT interrupts. This is quite dodgy
+ * but works until I find a better way
+ */
+ if (master && (flags & MPIC_BIG_ENDIAN))
+ flags |= MPIC_BROKEN_U3;
+
+ mpic = mpic_alloc(r.start, flags, 0, offset, count, master ? 252 : 0,
+ senses, count, name);
+ if (mpic == NULL)
+ return NULL;
+
+ mpic_init(mpic);
+
+ return mpic;
+ }
+
static int __init pmac_pic_probe_mpic(void)
{
struct mpic *mpic1, *mpic2;
struct device_node *np, *master = NULL, *slave = NULL;
- unsigned char senses[128];
- struct resource r;
/* We can have up to 2 MPICs cascaded */
for (np = NULL; (np = of_find_node_by_type(np, "open-pic"))
!= NULL;) {
if (master == NULL &&
- get_property(np, "interrupt-parent", NULL) != NULL)
+ get_property(np, "interrupts", NULL) == NULL)
master = of_node_get(np);
else if (slave == NULL)
slave = of_node_get(np);
@@ -557,13 +595,8 @@ static int __init pmac_pic_probe_mpic(vo
ppc_md.get_irq = mpic_get_irq;
/* Setup master */
- BUG_ON(of_address_to_resource(master, 0, &r));
- pmac_call_feature(PMAC_FTR_ENABLE_MPIC, master, 0, 0);
- prom_get_irq_senses(senses, 0, 128);
- mpic1 = mpic_alloc(r.start, MPIC_PRIMARY | MPIC_WANTS_RESET,
- 0, 0, 128, 252, senses, 128, " OpenPIC ");
+ mpic1 = pmac_setup_one_mpic(master, 1);
BUG_ON(mpic1 == NULL);
- mpic_init(mpic1);
/* Install NMI if any */
pmac_pic_setup_mpic_nmi(mpic1);
@@ -574,27 +607,12 @@ static int __init pmac_pic_probe_mpic(vo
if (slave == NULL || slave->n_intrs < 1)
return 0;
- /* Setup slave, failures are non-fatal */
- if (of_address_to_resource(slave, 0, &r)) {
- printk(KERN_ERR "Can't get address of MPIC %s\n",
- slave->full_name);
- return 0;
- }
- pmac_call_feature(PMAC_FTR_ENABLE_MPIC, slave, 0, 0);
- prom_get_irq_senses(senses, 128, 128 + 124);
-
- /* We don't need to set MPIC_BROKEN_U3 here since we don't have
- * hypertransport interrupts routed to it, at least not on currently
- * supported machines, that may change.
- */
- mpic2 = mpic_alloc(r.start, MPIC_BIG_ENDIAN | MPIC_WANTS_RESET,
- 0, 128, 124, 0, senses, 124, " U3-MPIC ");
+ mpic2 = pmac_setup_one_mpic(slave, 0);
if (mpic2 == NULL) {
- printk(KERN_ERR "Can't create slave MPIC %s\n",
- slave->full_name);
+ printk(KERN_ERR "Failed to setup slave MPIC\n");
+ of_node_put(slave);
return 0;
}
- mpic_init(mpic2);
mpic_setup_cascade(slave->intrs[0].line, pmac_u3_cascade, mpic2);
of_node_put(slave);
Index: linux-work/include/asm-powerpc/pmac_feature.h
===================================================================
--- linux-work.orig/include/asm-powerpc/pmac_feature.h 2005-11-24 17:18:48.000000000 +1100
+++ linux-work/include/asm-powerpc/pmac_feature.h 2005-12-13 18:13:11.000000000 +1100
@@ -121,6 +121,7 @@
#define PMAC_TYPE_IMAC_G5 0x152 /* iMac G5 */
#define PMAC_TYPE_XSERVE_G5 0x153 /* Xserve G5 */
#define PMAC_TYPE_UNKNOWN_K2 0x19f /* Any other K2 based */
+#define PMAC_TYPE_UNKNOWN_SHASTA 0x19e /* Any other Shasta based */
/*
* Motherboard flags
@@ -341,6 +342,7 @@ enum {
macio_pangea,
macio_intrepid,
macio_keylargo2,
+ macio_shasta,
};
struct macio_chip
Index: linux-work/arch/powerpc/Kconfig
===================================================================
--- linux-work.orig/arch/powerpc/Kconfig 2005-12-13 15:03:21.000000000 +1100
+++ linux-work/arch/powerpc/Kconfig 2005-12-13 18:13:11.000000000 +1100
@@ -300,6 +300,7 @@ config PPC_PMAC64
bool
depends on PPC_PMAC && POWER4
select U3_DART
+ select MPIC_BROKEN_U3
select GENERIC_TBSYNC
default y
Index: linux-work/arch/powerpc/kernel/prom.c
===================================================================
--- linux-work.orig/arch/powerpc/kernel/prom.c 2005-12-13 17:51:20.000000000 +1100
+++ linux-work/arch/powerpc/kernel/prom.c 2005-12-13 18:13:11.000000000 +1100
@@ -298,6 +298,16 @@ static int __devinit finish_node_interru
int i, j, n, sense;
unsigned int *irq, virq;
struct device_node *ic;
+ int trace = 0;
+
+ //#define TRACE(fmt...) do { if (trace) { printk(fmt); mdelay(1000); } } while(0)
+#define TRACE(fmt...)
+
+ if (!strcmp(np->name, "smu-doorbell"))
+ trace = 1;
+
+ TRACE("Finishing SMU doorbell ! num_interrupt_controllers = %d\n",
+ num_interrupt_controllers);
if (num_interrupt_controllers == 0) {
/*
@@ -332,11 +342,12 @@ static int __devinit finish_node_interru
}
ints = (unsigned int *) get_property(np, "interrupts", &intlen);
+ TRACE("ints=%p, intlen=%d\n", ints, intlen);
if (ints == NULL)
return 0;
intrcells = prom_n_intr_cells(np);
intlen /= intrcells * sizeof(unsigned int);
-
+ TRACE("intrcells=%d, new intlen=%d\n", intrcells, intlen);
np->intrs = prom_alloc(intlen * sizeof(*(np->intrs)), mem_start);
if (!np->intrs)
return -ENOMEM;
@@ -347,6 +358,7 @@ static int __devinit finish_node_interru
intrcount = 0;
for (i = 0; i < intlen; ++i, ints += intrcells) {
n = map_interrupt(&irq, &ic, np, ints, intrcells);
+ TRACE("map, irq=%d, ic=%p, n=%d\n", irq, ic, n);
if (n <= 0)
continue;
@@ -357,6 +369,7 @@ static int __devinit finish_node_interru
np->intrs[intrcount].sense = map_isa_senses[sense];
} else {
virq = virt_irq_create_mapping(irq[0]);
+ TRACE("virq=%d\n", virq);
#ifdef CONFIG_PPC64
if (virq == NO_IRQ) {
printk(KERN_CRIT "Could not allocate interrupt"
@@ -366,6 +379,12 @@ static int __devinit finish_node_interru
#endif
np->intrs[intrcount].line = irq_offset_up(virq);
sense = (n > 1)? (irq[1] & 3): 1;
+
+ /* Apple uses bits in there in a different way, let's
+ * only keep the real sense bit on macs
+ */
+ if (_machine == PLATFORM_POWERMAC)
+ sense &= 0x1;
np->intrs[intrcount].sense = map_mpic_senses[sense];
}
@@ -375,12 +394,13 @@ static int __devinit finish_node_interru
char *name = get_property(ic->parent, "name", NULL);
if (name && !strcmp(name, "u3"))
np->intrs[intrcount].line += 128;
- else if (!(name && !strcmp(name, "mac-io")))
+ else if (!(name && (!strcmp(name, "mac-io") ||
+ !strcmp(name, "u4"))))
/* ignore other cascaded controllers, such as
the k2-sata-root */
break;
}
-#endif
+#endif /* CONFIG_PPC64 */
if (n > 2) {
printk("hmmm, got %d intr cells for %s:", n,
np->full_name);
Index: linux-work/arch/powerpc/platforms/powermac/pci.c
===================================================================
--- linux-work.orig/arch/powerpc/platforms/powermac/pci.c 2005-12-13 17:51:20.000000000 +1100
+++ linux-work/arch/powerpc/platforms/powermac/pci.c 2005-12-13 18:13:12.000000000 +1100
@@ -1,7 +1,7 @@
/*
* Support for PCI bridges found on Power Macintoshes.
*
- * Copyright (C) 2003 Benjamin Herrenschmuidt (benh@kernel.crashing.org)
+ * Copyright (C) 2003-2005 Benjamin Herrenschmuidt (benh@kernel.crashing.org)
* Copyright (C) 1997 Paul Mackerras (paulus@samba.org)
*
* This program is free software; you can redistribute it and/or
@@ -25,7 +25,7 @@
#include <asm/pmac_feature.h>
#include <asm/grackle.h>
#ifdef CONFIG_PPC64
-#include <asm/iommu.h>
+//#include <asm/iommu.h>
#include <asm/ppc-pci.h>
#endif
@@ -44,6 +44,7 @@ static int add_bridge(struct device_node
static int has_uninorth;
#ifdef CONFIG_PPC64
static struct pci_controller *u3_agp;
+static struct pci_controller *u4_pcie;
static struct pci_controller *u3_ht;
#endif /* CONFIG_PPC64 */
@@ -97,11 +98,8 @@ static void __init fixup_bus_range(struc
/* Lookup the "bus-range" property for the hose */
bus_range = (int *) get_property(bridge, "bus-range", &len);
- if (bus_range == NULL || len < 2 * sizeof(int)) {
- printk(KERN_WARNING "Can't get bus-range for %s\n",
- bridge->full_name);
+ if (bus_range == NULL || len < 2 * sizeof(int))
return;
- }
bus_range[1] = fixup_one_level_bus_range(bridge->child, bus_range[1]);
}
@@ -128,14 +126,14 @@ static void __init fixup_bus_range(struc
*/
#define MACRISC_CFA0(devfn, off) \
- ((1 << (unsigned long)PCI_SLOT(dev_fn)) \
- | (((unsigned long)PCI_FUNC(dev_fn)) << 8) \
- | (((unsigned long)(off)) & 0xFCUL))
+ ((1 << (unsigned int)PCI_SLOT(dev_fn)) \
+ | (((unsigned int)PCI_FUNC(dev_fn)) << 8) \
+ | (((unsigned int)(off)) & 0xFCUL))
#define MACRISC_CFA1(bus, devfn, off) \
- ((((unsigned long)(bus)) << 16) \
- |(((unsigned long)(devfn)) << 8) \
- |(((unsigned long)(off)) & 0xFCUL) \
+ ((((unsigned int)(bus)) << 16) \
+ |(((unsigned int)(devfn)) << 8) \
+ |(((unsigned int)(off)) & 0xFCUL) \
|1UL)
static unsigned long macrisc_cfg_access(struct pci_controller* hose,
@@ -168,7 +166,8 @@ static int macrisc_read_config(struct pc
hose = pci_bus_to_host(bus);
if (hose == NULL)
return PCIBIOS_DEVICE_NOT_FOUND;
-
+ if (offset >= 0x100)
+ return PCIBIOS_BAD_REGISTER_NUMBER;
addr = macrisc_cfg_access(hose, bus->number, devfn, offset);
if (!addr)
return PCIBIOS_DEVICE_NOT_FOUND;
@@ -199,7 +198,8 @@ static int macrisc_write_config(struct p
hose = pci_bus_to_host(bus);
if (hose == NULL)
return PCIBIOS_DEVICE_NOT_FOUND;
-
+ if (offset >= 0x100)
+ return PCIBIOS_BAD_REGISTER_NUMBER;
addr = macrisc_cfg_access(hose, bus->number, devfn, offset);
if (!addr)
return PCIBIOS_DEVICE_NOT_FOUND;
@@ -234,12 +234,13 @@ static struct pci_ops macrisc_pci_ops =
/*
* Verify that a specific (bus, dev_fn) exists on chaos
*/
-static int
-chaos_validate_dev(struct pci_bus *bus, int devfn, int offset)
+static int chaos_validate_dev(struct pci_bus *bus, int devfn, int offset)
{
struct device_node *np;
u32 *vendor, *device;
+ if (offset >= 0x100)
+ return PCIBIOS_BAD_REGISTER_NUMBER;
np = pci_busdev_to_OF_node(bus, devfn);
if (np == NULL)
return PCIBIOS_DEVICE_NOT_FOUND;
@@ -341,10 +342,10 @@ static int u3_ht_skip_device(struct pci_
}
#define U3_HT_CFA0(devfn, off) \
- ((((unsigned long)devfn) << 8) | offset)
+ ((((unsigned int)devfn) << 8) | offset)
#define U3_HT_CFA1(bus, devfn, off) \
(U3_HT_CFA0(devfn, off) \
- + (((unsigned long)bus) << 16) \
+ + (((unsigned int)bus) << 16) \
+ 0x01000000UL)
static unsigned long u3_ht_cfg_access(struct pci_controller* hose,
@@ -370,7 +371,8 @@ static int u3_ht_read_config(struct pci_
hose = pci_bus_to_host(bus);
if (hose == NULL)
return PCIBIOS_DEVICE_NOT_FOUND;
-
+ if (offset >= 0x100)
+ return PCIBIOS_BAD_REGISTER_NUMBER;
addr = u3_ht_cfg_access(hose, bus->number, devfn, offset);
if (!addr)
return PCIBIOS_DEVICE_NOT_FOUND;
@@ -419,7 +421,8 @@ static int u3_ht_write_config(struct pci
hose = pci_bus_to_host(bus);
if (hose == NULL)
return PCIBIOS_DEVICE_NOT_FOUND;
-
+ if (offset >= 0x100)
+ return PCIBIOS_BAD_REGISTER_NUMBER;
addr = u3_ht_cfg_access(hose, bus->number, devfn, offset);
if (!addr)
return PCIBIOS_DEVICE_NOT_FOUND;
@@ -459,6 +462,112 @@ static struct pci_ops u3_ht_pci_ops =
u3_ht_read_config,
u3_ht_write_config
};
+
+#define U4_PCIE_CFA0(devfn, off) \
+ ((1 << ((unsigned int)PCI_SLOT(dev_fn))) \
+ | (((unsigned int)PCI_FUNC(dev_fn)) << 8) \
+ | ((((unsigned int)(off)) >> 8) << 28) \
+ | (((unsigned int)(off)) & 0xfcU))
+
+#define U4_PCIE_CFA1(bus, devfn, off) \
+ ((((unsigned int)(bus)) << 16) \
+ |(((unsigned int)(devfn)) << 8) \
+ | ((((unsigned int)(off)) >> 8) << 28) \
+ |(((unsigned int)(off)) & 0xfcU) \
+ |1UL)
+
+static unsigned long u4_pcie_cfg_access(struct pci_controller* hose,
+ u8 bus, u8 dev_fn, int offset)
+{
+ unsigned int caddr;
+
+ if (bus == hose->first_busno) {
+ caddr = U4_PCIE_CFA0(dev_fn, offset);
+ } else
+ caddr = U4_PCIE_CFA1(bus, dev_fn, offset);
+
+ /* Uninorth will return garbage if we don't read back the value ! */
+ do {
+ out_le32(hose->cfg_addr, caddr);
+ } while (in_le32(hose->cfg_addr) != caddr);
+
+ offset &= 0x03;
+ return ((unsigned long)hose->cfg_data) + offset;
+}
+
+static int u4_pcie_read_config(struct pci_bus *bus, unsigned int devfn,
+ int offset, int len, u32 *val)
+{
+ struct pci_controller *hose;
+ unsigned long addr;
+
+ hose = pci_bus_to_host(bus);
+ if (hose == NULL)
+ return PCIBIOS_DEVICE_NOT_FOUND;
+ if (offset >= 0x1000)
+ return PCIBIOS_BAD_REGISTER_NUMBER;
+ addr = u4_pcie_cfg_access(hose, bus->number, devfn, offset);
+ if (!addr)
+ return PCIBIOS_DEVICE_NOT_FOUND;
+ /*
+ * Note: the caller has already checked that offset is
+ * suitably aligned and that len is 1, 2 or 4.
+ */
+ switch (len) {
+ case 1:
+ *val = in_8((u8 *)addr);
+ break;
+ case 2:
+ *val = in_le16((u16 *)addr);
+ break;
+ default:
+ *val = in_le32((u32 *)addr);
+ break;
+ }
+ return PCIBIOS_SUCCESSFUL;
+}
+
+static int u4_pcie_write_config(struct pci_bus *bus, unsigned int devfn,
+ int offset, int len, u32 val)
+{
+ struct pci_controller *hose;
+ unsigned long addr;
+
+ hose = pci_bus_to_host(bus);
+ if (hose == NULL)
+ return PCIBIOS_DEVICE_NOT_FOUND;
+ if (offset >= 0x1000)
+ return PCIBIOS_BAD_REGISTER_NUMBER;
+ addr = u4_pcie_cfg_access(hose, bus->number, devfn, offset);
+ if (!addr)
+ return PCIBIOS_DEVICE_NOT_FOUND;
+ /*
+ * Note: the caller has already checked that offset is
+ * suitably aligned and that len is 1, 2 or 4.
+ */
+ switch (len) {
+ case 1:
+ out_8((u8 *)addr, val);
+ (void) in_8((u8 *)addr);
+ break;
+ case 2:
+ out_le16((u16 *)addr, val);
+ (void) in_le16((u16 *)addr);
+ break;
+ default:
+ out_le32((u32 *)addr, val);
+ (void) in_le32((u32 *)addr);
+ break;
+ }
+ return PCIBIOS_SUCCESSFUL;
+}
+
+static struct pci_ops u4_pcie_pci_ops =
+{
+ u4_pcie_read_config,
+ u4_pcie_write_config
+};
+
#endif /* CONFIG_PPC64 */
#ifdef CONFIG_PPC32
@@ -628,15 +737,36 @@ static void __init setup_u3_agp(struct p
hose->ops = ¯isc_pci_ops;
hose->cfg_addr = ioremap(0xf0000000 + 0x800000, 0x1000);
hose->cfg_data = ioremap(0xf0000000 + 0xc00000, 0x1000);
-
u3_agp = hose;
}
+static void __init setup_u4_pcie(struct pci_controller* hose)
+{
+ /* We currently only implement the "non-atomic" config space, to
+ * be optimised later.
+ */
+ hose->ops = &u4_pcie_pci_ops;
+ hose->cfg_addr = ioremap(0xf0000000 + 0x800000, 0x1000);
+ hose->cfg_data = ioremap(0xf0000000 + 0xc00000, 0x1000);
+
+ /* The bus contains a bridge from root -> device, we need to
+ * make it visible on bus 0 so that we pick the right type
+ * of config cycles. If we didn't, we would have to force all
+ * config cycles to be type 1. So we override the "bus-range"
+ * property here
+ */
+ hose->first_busno = 0x00;
+ hose->last_busno = 0xff;
+ u4_pcie = hose;
+}
+
static void __init setup_u3_ht(struct pci_controller* hose)
{
struct device_node *np = (struct device_node *)hose->arch_data;
+ struct pci_controller *other = NULL;
int i, cur;
+
hose->ops = &u3_ht_pci_ops;
/* We hard code the address because of the different size of
@@ -670,11 +800,20 @@ static void __init setup_u3_ht(struct pc
u3_ht = hose;
- if (u3_agp == NULL) {
- DBG("U3 has no AGP, using full resource range\n");
+ if (u3_agp != NULL)
+ other = u3_agp;
+ else if (u4_pcie != NULL)
+ other = u4_pcie;
+
+ if (other == NULL) {
+ DBG("U3/4 has no AGP/PCIE, using full resource range\n");
return;
}
+ /* Fixup bus range vs. PCIE */
+ if (u4_pcie)
+ hose->last_busno = u4_pcie->first_busno - 1;
+
/* We "remove" the AGP resources from the resources allocated to HT,
* that is we create "holes". However, that code does assumptions
* that so far happen to be true (cross fingers...), typically that
@@ -682,7 +821,7 @@ static void __init setup_u3_ht(struct pc
*/
cur = 0;
for (i=0; i<3; i++) {
- struct resource *res = &u3_agp->mem_resources[i];
+ struct resource *res = &other->mem_resources[i];
if (res->flags != IORESOURCE_MEM)
continue;
/* We don't care about "fine" resources */
@@ -777,9 +916,13 @@ static int __init add_bridge(struct devi
setup_u3_ht(hose);
disp_name = "U3-HT";
primary = 1;
+ } else if (device_is_compatible(dev, "u4-pcie")) {
+ setup_u4_pcie(hose);
+ disp_name = "U4-PCIE";
+ primary = 0;
}
- printk(KERN_INFO "Found %s PCI host bridge. Firmware bus number: %d->%d\n",
- disp_name, hose->first_busno, hose->last_busno);
+ printk(KERN_INFO "Found %s PCI host bridge. Firmware bus number:"
+ " %d->%d\n", disp_name, hose->first_busno, hose->last_busno);
#endif /* CONFIG_PPC64 */
/* 32 bits only bridges */
@@ -900,6 +1043,8 @@ void __init pmac_pci_init(void)
pci_setup_phb_io(u3_ht, 1);
if (u3_agp)
pci_setup_phb_io(u3_agp, 0);
+ if (u4_pcie)
+ pci_setup_phb_io(u4_pcie, 0);
/*
* On ppc64, fixup the IO resources on our host bridges as
@@ -912,7 +1057,8 @@ void __init pmac_pci_init(void)
/* Fixup the PCI<->OF mapping for U3 AGP due to bus renumbering. We
* assume there is no P2P bridge on the AGP bus, which should be a
- * safe assumptions hopefully.
+ * safe assumptions for now. We should do something better in the
+ * future though
*/
if (u3_agp) {
struct device_node *np = u3_agp->arch_data;
@@ -920,7 +1066,6 @@ void __init pmac_pci_init(void)
for (np = np->child; np; np = np->sibling)
PCI_DN(np)->busno = 0xf0;
}
-
/* pmac_check_ht_link(); */
/* Tell pci.c to not use the common resource allocation mechanism */
@@ -1127,7 +1272,8 @@ void pmac_pci_fixup_pciata(struct pci_de
good:
pci_read_config_byte(dev, PCI_CLASS_PROG, &progif);
if ((progif & 5) != 5) {
- printk(KERN_INFO "Forcing PCI IDE into native mode: %s\n", pci_name(dev));
+ printk(KERN_INFO "Forcing PCI IDE into native mode: %s\n",
+ pci_name(dev));
(void) pci_write_config_byte(dev, PCI_CLASS_PROG, progif|5);
if (pci_read_config_byte(dev, PCI_CLASS_PROG, &progif) ||
(progif & 5) != 5)
@@ -1153,7 +1299,8 @@ static void fixup_k2_sata(struct pci_dev
for (i = 0; i < 6; i++) {
dev->resource[i].start = dev->resource[i].end = 0;
dev->resource[i].flags = 0;
- pci_write_config_dword(dev, PCI_BASE_ADDRESS_0 + 4 * i, 0);
+ pci_write_config_dword(dev, PCI_BASE_ADDRESS_0 + 4 * i,
+ 0);
}
} else {
pci_read_config_word(dev, PCI_COMMAND, &cmd);
@@ -1162,7 +1309,8 @@ static void fixup_k2_sata(struct pci_dev
for (i = 0; i < 5; i++) {
dev->resource[i].start = dev->resource[i].end = 0;
dev->resource[i].flags = 0;
- pci_write_config_dword(dev, PCI_BASE_ADDRESS_0 + 4 * i, 0);
+ pci_write_config_dword(dev, PCI_BASE_ADDRESS_0 + 4 * i,
+ 0);
}
}
}
Index: linux-work/arch/powerpc/sysdev/mpic.c
===================================================================
--- linux-work.orig/arch/powerpc/sysdev/mpic.c 2005-12-13 18:02:03.000000000 +1100
+++ linux-work/arch/powerpc/sysdev/mpic.c 2005-12-13 18:13:12.000000000 +1100
@@ -13,6 +13,9 @@
*/
#undef DEBUG
+#undef DEBUG_IPI
+#undef DEBUG_IRQ
+#undef DEBUG_LOW
#include <linux/config.h>
#include <linux/types.h>
@@ -168,35 +171,86 @@ static void __init mpic_test_broken_ipi(
/* Test if an interrupt is sourced from HyperTransport (used on broken U3s)
* to force the edge setting on the MPIC and do the ack workaround.
*/
-static inline int mpic_is_ht_interrupt(struct mpic *mpic, unsigned int source_no)
+static inline int mpic_is_ht_interrupt(struct mpic *mpic, unsigned int source)
{
- if (source_no >= 128 || !mpic->fixups)
+ if (source >= 128 || !mpic->fixups)
return 0;
- return mpic->fixups[source_no].base != NULL;
+ return mpic->fixups[source].base != NULL;
}
-static inline void mpic_apic_end_irq(struct mpic *mpic, unsigned int source_no)
+static inline void mpic_ht_end_irq(struct mpic *mpic, unsigned int source)
{
- struct mpic_irq_fixup *fixup = &mpic->fixups[source_no];
+ struct mpic_irq_fixup *fixup = &mpic->fixups[source];
- spin_lock(&mpic->fixup_lock);
- writeb(0x11 + 2 * fixup->irq, fixup->base + 2);
- writel(fixup->data, fixup->base + 4);
- spin_unlock(&mpic->fixup_lock);
+ if (fixup->applebase) {
+ unsigned int soff = (fixup->index >> 3) & ~3;
+ unsigned int mask = 1U << (fixup->index & 0x1f);
+ writel(mask, fixup->applebase + soff);
+ } else {
+ spin_lock(&mpic->fixup_lock);
+ writeb(0x11 + 2 * fixup->index, fixup->base + 2);
+ writel(fixup->data, fixup->base + 4);
+ spin_unlock(&mpic->fixup_lock);
+ }
}
+static void mpic_startup_ht_interrupt(struct mpic *mpic, unsigned int source,
+ unsigned int irqflags)
+{
+ struct mpic_irq_fixup *fixup = &mpic->fixups[source];
+ unsigned long flags;
+ u32 tmp;
+
+ if (fixup->base == NULL)
+ return;
+
+ DBG("startup_ht_interrupt(%u, %u) index: %d\n",
+ source, irqflags, fixup->index);
+ spin_lock_irqsave(&mpic->fixup_lock, flags);
+ /* Enable and configure */
+ writeb(0x10 + 2 * fixup->index, fixup->base + 2);
+ tmp = readl(fixup->base + 4);
+ tmp &= ~(0x23U);
+ if (irqflags & IRQ_LEVEL)
+ tmp |= 0x22;
+ writel(tmp, fixup->base + 4);
+ spin_unlock_irqrestore(&mpic->fixup_lock, flags);
+}
+
+static void mpic_shutdown_ht_interrupt(struct mpic *mpic, unsigned int source,
+ unsigned int irqflags)
+{
+ struct mpic_irq_fixup *fixup = &mpic->fixups[source];
+ unsigned long flags;
+ u32 tmp;
+
+ if (fixup->base == NULL)
+ return;
+
+ DBG("shutdown_ht_interrupt(%u, %u)\n", source, irqflags);
+
+ /* Disable */
+ spin_lock_irqsave(&mpic->fixup_lock, flags);
+ writeb(0x10 + 2 * fixup->index, fixup->base + 2);
+ tmp = readl(fixup->base + 4);
+ tmp &= ~1U;
+ writel(tmp, fixup->base + 4);
+ spin_unlock_irqrestore(&mpic->fixup_lock, flags);
+}
-static void __init mpic_scan_ioapic(struct mpic *mpic, u8 __iomem *devbase)
+static void __init mpic_scan_ht_pic(struct mpic *mpic, u8 __iomem *devbase,
+ unsigned int devfn, u32 vdid)
{
int i, irq, n;
+ u8 __iomem *base;
u32 tmp;
u8 pos;
- for (pos = readb(devbase + 0x34); pos; pos = readb(devbase + pos + 1)) {
- u8 id = readb(devbase + pos);
-
- if (id == 0x08) {
+ for (pos = readb(devbase + PCI_CAPABILITY_LIST); pos != 0;
+ pos = readb(devbase + pos + PCI_CAP_LIST_NEXT)) {
+ u8 id = readb(devbase + pos + PCI_CAP_LIST_ID);
+ if (id == PCI_CAP_ID_HT_IRQCONF) {
id = readb(devbase + pos + 3);
if (id == 0x80)
break;
@@ -205,33 +259,41 @@ static void __init mpic_scan_ioapic(stru
if (pos == 0)
return;
- printk(KERN_INFO "mpic: - Workarounds @ %p, pos = 0x%02x\n", devbase, pos);
-
- devbase += pos;
-
- writeb(0x01, devbase + 2);
- n = (readl(devbase + 4) >> 16) & 0xff;
+ base = devbase + pos;
+ writeb(0x01, base + 2);
+ n = (readl(base + 4) >> 16) & 0xff;
+
+ printk(KERN_INFO "mpic: - HT:%02x.%x [0x%02x] vendor %04x device %04x"
+ " has %d irqs\n",
+ devfn >> 3, devfn & 0x7, pos, vdid & 0xffff, vdid >> 16, n + 1);
for (i = 0; i <= n; i++) {
- writeb(0x10 + 2 * i, devbase + 2);
- tmp = readl(devbase + 4);
- if ((tmp & 0x21) != 0x20)
- continue;
+ writeb(0x10 + 2 * i, base + 2);
+ tmp = readl(base + 4);
irq = (tmp >> 16) & 0xff;
- mpic->fixups[irq].irq = i;
- mpic->fixups[irq].base = devbase;
- writeb(0x11 + 2 * i, devbase + 2);
- mpic->fixups[irq].data = readl(devbase + 4) | 0x80000000;
+ DBG("HT PIC index 0x%x, irq 0x%x, tmp: %08x\n", i, irq, tmp);
+ /* mask it , will be unmasked later */
+ tmp |= 0x1;
+ writel(tmp, base + 4);
+ mpic->fixups[irq].index = i;
+ mpic->fixups[irq].base = base;
+ /* Apple HT PIC has a non-standard way of doing EOIs */
+ if ((vdid & 0xffff) == 0x106b)
+ mpic->fixups[irq].applebase = devbase + 0x60;
+ else
+ mpic->fixups[irq].applebase = NULL;
+ writeb(0x11 + 2 * i, base + 2);
+ mpic->fixups[irq].data = readl(base + 4) | 0x80000000;
}
}
-static void __init mpic_scan_ioapics(struct mpic *mpic)
+static void __init mpic_scan_ht_pics(struct mpic *mpic)
{
unsigned int devfn;
u8 __iomem *cfgspace;
- printk(KERN_INFO "mpic: Setting up IO-APICs workarounds for U3\n");
+ printk(KERN_INFO "mpic: Setting up HT PICs workarounds for U3/U4\n");
/* Allocate fixups array */
mpic->fixups = alloc_bootmem(128 * sizeof(struct mpic_irq_fixup));
@@ -247,13 +309,14 @@ static void __init mpic_scan_ioapics(str
cfgspace = ioremap(0xf2000000, 0x10000);
BUG_ON(cfgspace == NULL);
- /* Now we scan all slots. We do a very quick scan, we read the header type,
- * vendor ID and device ID only, that's plenty enough
+ /* Now we scan all slots. We do a very quick scan, we read the header
+ * type, vendor ID and device ID only, that's plenty enough
*/
for (devfn = 0; devfn < 0x100; devfn++) {
u8 __iomem *devbase = cfgspace + (devfn << 8);
u8 hdr_type = readb(devbase + PCI_HEADER_TYPE);
u32 l = readl(devbase + PCI_VENDOR_ID);
+ u16 s;
DBG("devfn %x, l: %x\n", devfn, l);
@@ -261,8 +324,12 @@ static void __init mpic_scan_ioapics(str
if (l == 0xffffffff || l == 0x00000000 ||
l == 0x0000ffff || l == 0xffff0000)
goto next;
+ /* Check if is supports capability lists */
+ s = readw(devbase + PCI_STATUS);
+ if (!(s & PCI_STATUS_CAP_LIST))
+ goto next;
- mpic_scan_ioapic(mpic, devbase);
+ mpic_scan_ht_pic(mpic, devbase, devfn, l);
next:
/* next device, if function 0 */
@@ -363,6 +430,31 @@ static void mpic_enable_irq(unsigned int
break;
}
} while(mpic_irq_read(src, MPIC_IRQ_VECTOR_PRI) & MPIC_VECPRI_MASK);
+
+#ifdef CONFIG_MPIC_BROKEN_U3
+ if (mpic->flags & MPIC_BROKEN_U3) {
+ unsigned int src = irq - mpic->irq_offset;
+ if (mpic_is_ht_interrupt(mpic, src) &&
+ (irq_desc[irq].status & IRQ_LEVEL))
+ mpic_ht_end_irq(mpic, src);
+ }
+#endif /* CONFIG_MPIC_BROKEN_U3 */
+}
+
+static unsigned int mpic_startup_irq(unsigned int irq)
+{
+#ifdef CONFIG_MPIC_BROKEN_U3
+ struct mpic *mpic = mpic_from_irq(irq);
+ unsigned int src = irq - mpic->irq_offset;
+
+ if (mpic_is_ht_interrupt(mpic, src))
+ mpic_startup_ht_interrupt(mpic, src, irq_desc[irq].status);
+
+#endif /* CONFIG_MPIC_BROKEN_U3 */
+
+ mpic_enable_irq(irq);
+
+ return 0;
}
static void mpic_disable_irq(unsigned int irq)
@@ -386,12 +478,27 @@ static void mpic_disable_irq(unsigned in
} while(!(mpic_irq_read(src, MPIC_IRQ_VECTOR_PRI) & MPIC_VECPRI_MASK));
}
+static void mpic_shutdown_irq(unsigned int irq)
+{
+#ifdef CONFIG_MPIC_BROKEN_U3
+ struct mpic *mpic = mpic_from_irq(irq);
+ unsigned int src = irq - mpic->irq_offset;
+
+ if (mpic_is_ht_interrupt(mpic, src))
+ mpic_shutdown_ht_interrupt(mpic, src, irq_desc[irq].status);
+
+#endif /* CONFIG_MPIC_BROKEN_U3 */
+
+ mpic_disable_irq(irq);
+}
+
static void mpic_end_irq(unsigned int irq)
{
struct mpic *mpic = mpic_from_irq(irq);
+#ifdef DEBUG_IRQ
DBG("%s: end_irq: %d\n", mpic->name, irq);
-
+#endif
/* We always EOI on end_irq() even for edge interrupts since that
* should only lower the priority, the MPIC should have properly
* latched another edge interrupt coming in anyway
@@ -400,8 +507,9 @@ static void mpic_end_irq(unsigned int ir
#ifdef CONFIG_MPIC_BROKEN_U3
if (mpic->flags & MPIC_BROKEN_U3) {
unsigned int src = irq - mpic->irq_offset;
- if (mpic_is_ht_interrupt(mpic, src))
- mpic_apic_end_irq(mpic, src);
+ if (mpic_is_ht_interrupt(mpic, src) &&
+ (irq_desc[irq].status & IRQ_LEVEL))
+ mpic_ht_end_irq(mpic, src);
}
#endif /* CONFIG_MPIC_BROKEN_U3 */
@@ -482,6 +590,8 @@ struct mpic * __init mpic_alloc(unsigned
mpic->name = name;
mpic->hc_irq.typename = name;
+ mpic->hc_irq.startup = mpic_startup_irq;
+ mpic->hc_irq.shutdown = mpic_shutdown_irq;
mpic->hc_irq.enable = mpic_enable_irq;
mpic->hc_irq.disable = mpic_disable_irq;
mpic->hc_irq.end = mpic_end_irq;
@@ -650,10 +760,10 @@ void __init mpic_init(struct mpic *mpic)
mpic->irq_count = mpic->num_sources;
#ifdef CONFIG_MPIC_BROKEN_U3
- /* Do the ioapic fixups on U3 broken mpic */
+ /* Do the HT PIC fixups on U3 broken mpic */
DBG("MPIC flags: %x\n", mpic->flags);
if ((mpic->flags & MPIC_BROKEN_U3) && (mpic->flags & MPIC_PRIMARY))
- mpic_scan_ioapics(mpic);
+ mpic_scan_ht_pics(mpic);
#endif /* CONFIG_MPIC_BROKEN_U3 */
for (i = 0; i < mpic->num_sources; i++) {
@@ -840,7 +950,9 @@ void mpic_send_ipi(unsigned int ipi_no,
BUG_ON(mpic == NULL);
+#ifdef DEBUG_IPI
DBG("%s: send_ipi(ipi_no: %d)\n", mpic->name, ipi_no);
+#endif
mpic_cpu_write(MPIC_CPU_IPI_DISPATCH_0 + ipi_no * 0x10,
mpic_physmask(cpu_mask & cpus_addr(cpu_online_map)[0]));
@@ -851,19 +963,28 @@ int mpic_get_one_irq(struct mpic *mpic,
u32 irq;
irq = mpic_cpu_read(MPIC_CPU_INTACK) & MPIC_VECPRI_VECTOR_MASK;
+#ifdef DEBUG_LOW
DBG("%s: get_one_irq(): %d\n", mpic->name, irq);
-
+#endif
if (mpic->cascade && irq == mpic->cascade_vec) {
+#ifdef DEBUG_LOW
DBG("%s: cascading ...\n", mpic->name);
+#endif
irq = mpic->cascade(regs, mpic->cascade_data);
mpic_eoi(mpic);
return irq;
}
if (unlikely(irq == MPIC_VEC_SPURRIOUS))
return -1;
- if (irq < MPIC_VEC_IPI_0)
+ if (irq < MPIC_VEC_IPI_0) {
+#ifdef DEBUG_IRQ
+ DBG("%s: irq %d\n", mpic->name, irq + mpic->irq_offset);
+#endif
return irq + mpic->irq_offset;
+ }
+#ifdef DEBUG_IPI
DBG("%s: ipi %d !\n", mpic->name, irq - MPIC_VEC_IPI_0);
+#endif
return irq - MPIC_VEC_IPI_0 + mpic->ipi_offset;
}
Index: linux-work/include/linux/pci_regs.h
===================================================================
--- linux-work.orig/include/linux/pci_regs.h 2005-11-24 17:18:49.000000000 +1100
+++ linux-work/include/linux/pci_regs.h 2005-12-13 18:13:12.000000000 +1100
@@ -196,6 +196,7 @@
#define PCI_CAP_ID_MSI 0x05 /* Message Signalled Interrupts */
#define PCI_CAP_ID_CHSWP 0x06 /* CompactPCI HotSwap */
#define PCI_CAP_ID_PCIX 0x07 /* PCI-X */
+#define PCI_CAP_ID_HT_IRQCONF 0x08 /* HyperTransport IRQ Configuration */
#define PCI_CAP_ID_SHPC 0x0C /* PCI Standard Hot-Plug Controller */
#define PCI_CAP_ID_EXP 0x10 /* PCI Express */
#define PCI_CAP_ID_MSIX 0x11 /* MSI-X */
Index: linux-work/arch/powerpc/kernel/pci_64.c
===================================================================
--- linux-work.orig/arch/powerpc/kernel/pci_64.c 2005-12-13 17:51:20.000000000 +1100
+++ linux-work/arch/powerpc/kernel/pci_64.c 2005-12-13 18:13:12.000000000 +1100
@@ -34,7 +34,7 @@
#ifdef DEBUG
#include <asm/udbg.h>
-#define DBG(fmt...) udbg_printf(fmt)
+#define DBG(fmt...) printk(fmt)
#else
#define DBG(fmt...)
#endif
@@ -323,6 +323,7 @@ static void pci_parse_of_addrs(struct de
addrs = (u32 *) get_property(node, "assigned-addresses", &proplen);
if (!addrs)
return;
+ DBG(" parse addresses (%d bytes) @ %p\n", proplen, addrs);
for (; proplen >= 20; proplen -= 20, addrs += 5) {
flags = pci_parse_of_flags(addrs[0]);
if (!flags)
@@ -332,6 +333,9 @@ static void pci_parse_of_addrs(struct de
if (!size)
continue;
i = addrs[0] & 0xff;
+ DBG(" base: %llx, size: %llx, i: %x\n",
+ (unsigned long long)base, (unsigned long long)size, i);
+
if (PCI_BASE_ADDRESS_0 <= i && i <= PCI_BASE_ADDRESS_5) {
res = &dev->resource[(i - PCI_BASE_ADDRESS_0) >> 2];
} else if (i == dev->rom_base_reg) {
@@ -362,6 +366,8 @@ struct pci_dev *of_create_pci_dev(struct
if (type == NULL)
type = "";
+ DBG(" create device, devfn: %x, type: %s\n", devfn, type);
+
memset(dev, 0, sizeof(struct pci_dev));
dev->bus = bus;
dev->sysdata = node;
@@ -375,12 +381,14 @@ struct pci_dev *of_create_pci_dev(struct
dev->subsystem_vendor = get_int_prop(node, "subsystem-vendor-id", 0);
dev->subsystem_device = get_int_prop(node, "subsystem-id", 0);
- dev->cfg_size = 256; /*pci_cfg_space_size(dev);*/
+ dev->cfg_size = pci_cfg_space_size(dev);
sprintf(pci_name(dev), "%04x:%02x:%02x.%d", pci_domain_nr(bus),
dev->bus->number, PCI_SLOT(devfn), PCI_FUNC(devfn));
dev->class = get_int_prop(node, "class-code", 0);
+ DBG(" class: 0x%x\n", dev->class);
+
dev->current_state = 4; /* unknown power state */
if (!strcmp(type, "pci")) {
@@ -402,6 +410,8 @@ struct pci_dev *of_create_pci_dev(struct
pci_parse_of_addrs(node, dev);
+ DBG(" adding to system ...\n");
+
pci_device_add(dev, bus);
/* XXX pci_scan_msi_device(dev); */
@@ -418,15 +428,21 @@ void __devinit of_scan_bus(struct device
int reglen, devfn;
struct pci_dev *dev;
+ DBG("of_scan_bus(%s) bus no %d... \n", node->full_name, bus->number);
+
while ((child = of_get_next_child(node, child)) != NULL) {
+ DBG(" * %s\n", child->full_name);
reg = (u32 *) get_property(child, "reg", ®len);
if (reg == NULL || reglen < 20)
continue;
devfn = (reg[0] >> 8) & 0xff;
+
/* create a new pci_dev for this device */
dev = of_create_pci_dev(child, bus, devfn);
if (!dev)
continue;
+ DBG("dev header type: %x\n", dev->hdr_type);
+
if (dev->hdr_type == PCI_HEADER_TYPE_BRIDGE ||
dev->hdr_type == PCI_HEADER_TYPE_CARDBUS)
of_scan_pci_bridge(child, dev);
@@ -446,16 +462,18 @@ void __devinit of_scan_pci_bridge(struct
unsigned int flags;
u64 size;
+ DBG("of_scan_pci_bridge(%s)\n", node->full_name);
+
/* parse bus-range property */
busrange = (u32 *) get_property(node, "bus-range", &len);
if (busrange == NULL || len != 8) {
- printk(KERN_ERR "Can't get bus-range for PCI-PCI bridge %s\n",
+ printk(KERN_DEBUG "Can't get bus-range for PCI-PCI bridge %s\n",
node->full_name);
return;
}
ranges = (u32 *) get_property(node, "ranges", &len);
if (ranges == NULL) {
- printk(KERN_ERR "Can't get ranges for PCI-PCI bridge %s\n",
+ printk(KERN_DEBUG "Can't get ranges for PCI-PCI bridge %s\n",
node->full_name);
return;
}
@@ -509,10 +527,13 @@ void __devinit of_scan_pci_bridge(struct
}
sprintf(bus->name, "PCI Bus %04x:%02x", pci_domain_nr(bus),
bus->number);
+ DBG(" bus name: %s\n", bus->name);
mode = PCI_PROBE_NORMAL;
if (ppc_md.pci_probe_mode)
mode = ppc_md.pci_probe_mode(bus);
+ DBG(" probe mode: %d\n", mode);
+
if (mode == PCI_PROBE_DEVTREE)
of_scan_bus(node, bus);
else if (mode == PCI_PROBE_NORMAL)
@@ -528,6 +549,8 @@ void __devinit scan_phb(struct pci_contr
int i, mode;
struct resource *res;
+ DBG("Scanning PHB %s\n", node ? node->full_name : "<NO NAME>");
+
bus = pci_create_bus(NULL, hose->first_busno, hose->ops, node);
if (bus == NULL) {
printk(KERN_ERR "Failed to create bus for PCI domain %04x\n",
@@ -552,8 +575,9 @@ void __devinit scan_phb(struct pci_contr
mode = PCI_PROBE_NORMAL;
#ifdef CONFIG_PPC_MULTIPLATFORM
- if (ppc_md.pci_probe_mode)
+ if (node && ppc_md.pci_probe_mode)
mode = ppc_md.pci_probe_mode(bus);
+ DBG(" probe mode: %d\n", mode);
if (mode == PCI_PROBE_DEVTREE) {
bus->subordinate = hose->last_busno;
of_scan_bus(node, bus);
@@ -842,8 +866,7 @@ pgprot_t pci_phys_mem_access_prot(struct
* Returns a negative error code on failure, zero on success.
*/
int pci_mmap_page_range(struct pci_dev *dev, struct vm_area_struct *vma,
- enum pci_mmap_state mmap_state,
- int write_combine)
+ enum pci_mmap_state mmap_state, int write_combine)
{
unsigned long offset = vma->vm_pgoff << PAGE_SHIFT;
struct resource *rp;
Index: linux-work/arch/powerpc/kernel/udbg.c
===================================================================
--- linux-work.orig/arch/powerpc/kernel/udbg.c 2005-12-06 16:17:43.000000000 +1100
+++ linux-work/arch/powerpc/kernel/udbg.c 2005-12-13 18:13:12.000000000 +1100
@@ -110,10 +110,12 @@ static int early_console_initialized;
void __init disable_early_printk(void)
{
+#if 1
if (!early_console_initialized)
return;
unregister_console(&udbg_console);
early_console_initialized = 0;
+#endif
}
/* called by setup_system */
Index: linux-work/arch/powerpc/platforms/powermac/setup.c
===================================================================
--- linux-work.orig/arch/powerpc/platforms/powermac/setup.c 2005-12-13 17:51:20.000000000 +1100
+++ linux-work/arch/powerpc/platforms/powermac/setup.c 2005-12-13 18:13:12.000000000 +1100
@@ -345,7 +345,7 @@ void __init pmac_setup_arch(void)
#ifdef CONFIG_SMP
/* Check for Core99 */
- if (find_devices("uni-n") || find_devices("u3"))
+ if (find_devices("uni-n") || find_devices("u3") || find_devices("u4"))
smp_ops = &core99_smp_ops;
#ifdef CONFIG_PPC32
else
@@ -733,10 +733,11 @@ static int pmac_pci_probe_mode(struct pc
struct device_node *node = bus->sysdata;
/* We need to use normal PCI probing for the AGP bus,
- since the device for the AGP bridge isn't in the tree. */
- if (bus->self == NULL && device_is_compatible(node, "u3-agp"))
+ * since the device for the AGP bridge isn't in the tree.
+ */
+ if (bus->self == NULL && (device_is_compatible(node, "u3-agp") ||
+ device_is_compatible(node, "u4-pcie")))
return PCI_PROBE_NORMAL;
-
return PCI_PROBE_DEVTREE;
}
#endif
Index: linux-work/arch/powerpc/platforms/powermac/smp.c
===================================================================
--- linux-work.orig/arch/powerpc/platforms/powermac/smp.c 2005-12-13 15:03:21.000000000 +1100
+++ linux-work/arch/powerpc/platforms/powermac/smp.c 2005-12-13 18:13:12.000000000 +1100
@@ -361,7 +361,6 @@ static void __init psurge_dual_sync_tb(i
set_dec(tb_ticks_per_jiffy);
/* XXX fixme */
set_tb(0, 0);
- last_jiffy_stamp(cpu_nr) = 0;
if (cpu_nr > 0) {
mb();
@@ -429,15 +428,62 @@ struct smp_ops_t psurge_smp_ops = {
};
#endif /* CONFIG_PPC32 - actually powersurge support */
+/*
+ * Core 99 and later support
+ */
+
+static void (*pmac_tb_freeze)(int freeze);
+static unsigned long timebase;
+static int tb_req;
+
+static void smp_core99_give_timebase(void)
+{
+ unsigned long flags;
+
+ local_irq_save(flags);
+
+ while(!tb_req)
+ barrier();
+ tb_req = 0;
+ (*pmac_tb_freeze)(1);
+ mb();
+ timebase = get_tb();
+ mb();
+ while (timebase)
+ barrier();
+ mb();
+ (*pmac_tb_freeze)(0);
+ mb();
+
+ local_irq_restore(flags);
+}
+
+
+static void __devinit smp_core99_take_timebase(void)
+{
+ unsigned long flags;
+
+ local_irq_save(flags);
+
+ tb_req = 1;
+ mb();
+ while (!timebase)
+ barrier();
+ mb();
+ set_tb(timebase >> 32, timebase & 0xffffffff);
+ timebase = 0;
+ mb();
+ set_dec(tb_ticks_per_jiffy/2);
+
+ local_irq_restore(flags);
+}
+
#ifdef CONFIG_PPC64
/*
* G5s enable/disable the timebase via an i2c-connected clock chip.
*/
static struct device_node *pmac_tb_clock_chip_host;
static u8 pmac_tb_pulsar_addr;
-static void (*pmac_tb_freeze)(int freeze);
-static DEFINE_SPINLOCK(timebase_lock);
-static unsigned long timebase;
static void smp_core99_cypress_tb_freeze(int freeze)
{
@@ -447,7 +493,8 @@ static void smp_core99_cypress_tb_freeze
/* Strangely, the device-tree says address is 0xd2, but darwin
* accesses 0xd0 ...
*/
- pmac_low_i2c_setmode(pmac_tb_clock_chip_host, pmac_low_i2c_mode_combined);
+ pmac_low_i2c_setmode(pmac_tb_clock_chip_host,
+ pmac_low_i2c_mode_combined);
rc = pmac_low_i2c_xfer(pmac_tb_clock_chip_host,
0xd0 | pmac_low_i2c_read,
0x81, &data, 1);
@@ -475,7 +522,8 @@ static void smp_core99_pulsar_tb_freeze(
u8 data;
int rc;
- pmac_low_i2c_setmode(pmac_tb_clock_chip_host, pmac_low_i2c_mode_combined);
+ pmac_low_i2c_setmode(pmac_tb_clock_chip_host,
+ pmac_low_i2c_mode_combined);
rc = pmac_low_i2c_xfer(pmac_tb_clock_chip_host,
pmac_tb_pulsar_addr | pmac_low_i2c_read,
0x2e, &data, 1);
@@ -496,54 +544,14 @@ static void smp_core99_pulsar_tb_freeze(
}
}
-
-static void smp_core99_give_timebase(void)
-{
- /* Open i2c bus for synchronous access */
- if (pmac_low_i2c_open(pmac_tb_clock_chip_host, 0))
- panic("Can't open i2c for TB sync !\n");
-
- spin_lock(&timebase_lock);
- (*pmac_tb_freeze)(1);
- mb();
- timebase = get_tb();
- spin_unlock(&timebase_lock);
-
- while (timebase)
- barrier();
-
- spin_lock(&timebase_lock);
- (*pmac_tb_freeze)(0);
- spin_unlock(&timebase_lock);
-
- /* Close i2c bus */
- pmac_low_i2c_close(pmac_tb_clock_chip_host);
-}
-
-
-static void __devinit smp_core99_take_timebase(void)
-{
- while (!timebase)
- barrier();
- spin_lock(&timebase_lock);
- set_tb(timebase >> 32, timebase & 0xffffffff);
- timebase = 0;
- spin_unlock(&timebase_lock);
-}
-
-static void __init smp_core99_setup(int ncpus)
+static void __init smp_core99_setup_i2c_hwsync(int ncpus)
{
struct device_node *cc = NULL;
struct device_node *p;
+ const char *name = NULL;
u32 *reg;
int ok;
- /* HW sync only on these platforms */
- if (!machine_is_compatible("PowerMac7,2") &&
- !machine_is_compatible("PowerMac7,3") &&
- !machine_is_compatible("RackMac3,1"))
- return;
-
/* Look for the clock chip */
while ((cc = of_find_node_by_name(cc, "i2c-hwclock")) != NULL) {
p = of_get_parent(cc);
@@ -561,114 +569,64 @@ static void __init smp_core99_setup(int
if (device_is_compatible(cc, "pulsar-legacy-slewing")) {
pmac_tb_freeze = smp_core99_pulsar_tb_freeze;
pmac_tb_pulsar_addr = 0xd2;
- printk(KERN_INFO "Timebase clock is Pulsar chip\n");
+ name = "Pulsar";
} else if (device_is_compatible(cc, "cy28508")) {
pmac_tb_freeze = smp_core99_cypress_tb_freeze;
- printk(KERN_INFO "Timebase clock is Cypress chip\n");
+ name = "Cypress";
}
break;
case 0xd4:
pmac_tb_freeze = smp_core99_pulsar_tb_freeze;
pmac_tb_pulsar_addr = 0xd4;
- printk(KERN_INFO "Timebase clock is Pulsar chip\n");
+ name = "Pulsar";
break;
}
- if (pmac_tb_freeze != NULL) {
- pmac_tb_clock_chip_host = of_get_parent(cc);
- of_node_put(cc);
+ if (pmac_tb_freeze != NULL)
break;
- }
}
- if (pmac_tb_freeze == NULL) {
- smp_ops->give_timebase = smp_generic_give_timebase;
- smp_ops->take_timebase = smp_generic_take_timebase;
+ if (pmac_tb_freeze != NULL) {
+ struct device_node *p = of_get_parent(cc);
+ of_node_put(cc);
+ while(p && strcmp(p->type, "i2c")) {
+ cc = of_get_parent(p);
+ of_node_put(p);
+ p = cc;
+ }
+ if (p == NULL)
+ goto no_i2c_sync;
+ /* Open i2c bus for synchronous access */
+ if (pmac_low_i2c_open(p, 0)) {
+ printk(KERN_ERR "Failed top open i2c bus %s for clock"
+ " sync, fallback to software sync !\n",
+ p->full_name);
+ of_node_put(p);
+ goto no_i2c_sync;
+ }
+ pmac_tb_clock_chip_host = p;
+ printk(KERN_INFO "Processor timebase sync using %s i2c clock\n",
+ name);
+ return;
}
+ no_i2c_sync:
+ pmac_tb_freeze = NULL;
}
-/* nothing to do here, caches are already set up by service processor */
-static inline void __devinit core99_init_caches(int cpu)
-{
-}
+#endif /* CONFIG_PPC64 */
-#else /* CONFIG_PPC64 */
/*
- * SMP G4 powermacs use a GPIO to enable/disable the timebase.
+ * SMP G4 and newer G5 use a GPIO to enable/disable the timebase.
*/
static unsigned int core99_tb_gpio; /* Timebase freeze GPIO */
-static unsigned int pri_tb_hi, pri_tb_lo;
-static unsigned int pri_tb_stamp;
-
-/* not __init, called in sleep/wakeup code */
-void smp_core99_give_timebase(void)
+static void smp_core99_gpio_tb_freeze(int freeze)
{
- unsigned long flags;
- unsigned int t;
-
- /* wait for the secondary to be in take_timebase */
- for (t = 100000; t > 0 && !sec_tb_reset; --t)
- udelay(10);
- if (!sec_tb_reset) {
- printk(KERN_WARNING "Timeout waiting sync on second CPU\n");
- return;
- }
-
- /* freeze the timebase and read it */
- /* disable interrupts so the timebase is disabled for the
- shortest possible time */
- local_irq_save(flags);
- pmac_call_feature(PMAC_FTR_WRITE_GPIO, NULL, core99_tb_gpio, 4);
+ if (freeze)
+ pmac_call_feature(PMAC_FTR_WRITE_GPIO, NULL, core99_tb_gpio, 4);
+ else
+ pmac_call_feature(PMAC_FTR_WRITE_GPIO, NULL, core99_tb_gpio, 0);
pmac_call_feature(PMAC_FTR_READ_GPIO, NULL, core99_tb_gpio, 0);
- mb();
- pri_tb_hi = get_tbu();
- pri_tb_lo = get_tbl();
- pri_tb_stamp = last_jiffy_stamp(smp_processor_id());
- mb();
-
- /* tell the secondary we're ready */
- sec_tb_reset = 2;
- mb();
-
- /* wait for the secondary to have taken it */
- /* note: can't use udelay here, since it needs the timebase running */
- for (t = 10000000; t > 0 && sec_tb_reset; --t)
- barrier();
- if (sec_tb_reset)
- /* XXX BUG_ON here? */
- printk(KERN_WARNING "Timeout waiting sync(2) on second CPU\n");
-
- /* Now, restart the timebase by leaving the GPIO to an open collector */
- pmac_call_feature(PMAC_FTR_WRITE_GPIO, NULL, core99_tb_gpio, 0);
- pmac_call_feature(PMAC_FTR_READ_GPIO, NULL, core99_tb_gpio, 0);
- local_irq_restore(flags);
-}
-
-/* not __init, called in sleep/wakeup code */
-void smp_core99_take_timebase(void)
-{
- unsigned long flags;
-
- /* tell the primary we're here */
- sec_tb_reset = 1;
- mb();
-
- /* wait for the primary to set pri_tb_hi/lo */
- while (sec_tb_reset < 2)
- mb();
-
- /* set our stuff the same as the primary */
- local_irq_save(flags);
- set_dec(1);
- set_tb(pri_tb_hi, pri_tb_lo);
- last_jiffy_stamp(smp_processor_id()) = pri_tb_stamp;
- mb();
-
- /* tell the primary we're done */
- sec_tb_reset = 0;
- mb();
- local_irq_restore(flags);
}
/* L2 and L3 cache settings to pass from CPU0 to CPU1 on G4 cpus */
@@ -677,6 +635,7 @@ volatile static long int core99_l3_cache
static void __devinit core99_init_caches(int cpu)
{
+#ifndef CONFIG_PPC64
if (!cpu_has_feature(CPU_FTR_L2CR))
return;
@@ -702,30 +661,80 @@ static void __devinit core99_init_caches
_set_L3CR(core99_l3_cache);
printk("CPU%d: L3CR set to %lx\n", cpu, core99_l3_cache);
}
+#endif /* !CONFIG_PPC64 */
}
static void __init smp_core99_setup(int ncpus)
{
- struct device_node *cpu;
- u32 *tbprop = NULL;
- int i;
-
- core99_tb_gpio = KL_GPIO_TB_ENABLE; /* default value */
- cpu = of_find_node_by_type(NULL, "cpu");
- if (cpu != NULL) {
- tbprop = (u32 *)get_property(cpu, "timebase-enable", NULL);
- if (tbprop)
- core99_tb_gpio = *tbprop;
- of_node_put(cpu);
- }
-
- /* XXX should get this from reg properties */
- for (i = 1; i < ncpus; ++i)
- smp_hw_index[i] = i;
- powersave_nap = 0;
-}
+#ifdef CONFIG_PPC64
+
+ /* i2c based HW sync on some G5s */
+ if (machine_is_compatible("PowerMac7,2") ||
+ machine_is_compatible("PowerMac7,3") ||
+ machine_is_compatible("RackMac3,1"))
+ smp_core99_setup_i2c_hwsync(ncpus);
+
+ /* GPIO based HW sync on recent G5s */
+ if (pmac_tb_freeze == NULL) {
+ struct device_node *np =
+ of_find_node_by_name(NULL, "timebase-enable");
+ u32 *reg = (u32 *)get_property(np, "reg", NULL);
+
+ if (np && reg && !strcmp(np->type, "gpio")) {
+ core99_tb_gpio = *reg;
+ if (core99_tb_gpio < 0x50)
+ core99_tb_gpio += 0x50;
+ pmac_tb_freeze = smp_core99_gpio_tb_freeze;
+ printk(KERN_INFO "Processor timebase sync using"
+ " GPIO 0x%02x\n", core99_tb_gpio);
+ }
+ }
+
+#else /* CONFIG_PPC64 */
+
+ /* GPIO based HW sync on ppc32 Core99 */
+ if (pmac_tb_freeze == NULL && !machine_is_compatible("MacRISC4")) {
+ struct device_node *cpu;
+ u32 *tbprop = NULL;
+
+ core99_tb_gpio = KL_GPIO_TB_ENABLE; /* default value */
+ cpu = of_find_node_by_type(NULL, "cpu");
+ if (cpu != NULL) {
+ tbprop = (u32 *)get_property(cpu, "timebase-enable",
+ NULL);
+ if (tbprop)
+ core99_tb_gpio = *tbprop;
+ of_node_put(cpu);
+ }
+ pmac_tb_freeze = smp_core99_gpio_tb_freeze;
+ printk(KERN_INFO "Processor timebase sync using"
+ " GPIO 0x%02x\n", core99_tb_gpio);
+ }
+
+#endif /* CONFIG_PPC64 */
+
+ /* No timebase sync, fallback to software */
+ if (pmac_tb_freeze == NULL) {
+ smp_ops->give_timebase = smp_generic_give_timebase;
+ smp_ops->take_timebase = smp_generic_take_timebase;
+ printk(KERN_INFO "Processor timebase sync using software\n");
+ }
+
+#ifndef CONFIG_PPC64
+ {
+ int i;
+
+ /* XXX should get this from reg properties */
+ for (i = 1; i < ncpus; ++i)
+ smp_hw_index[i] = i;
+ }
#endif
+ /* 32 bits SMP can't NAP */
+ if (!machine_is_compatible("MacRISC4"))
+ powersave_nap = 0;
+}
+
static int __init smp_core99_probe(void)
{
struct device_node *cpus;
@@ -803,17 +812,25 @@ static void __devinit smp_core99_setup_c
mpic_setup_this_cpu();
if (cpu_nr == 0) {
-#ifdef CONFIG_POWER4
+#ifdef CONFIG_PPC64
extern void g5_phy_disable_cpu1(void);
+ /* Close i2c bus if it was used for tb sync */
+ if (pmac_tb_clock_chip_host) {
+ pmac_low_i2c_close(pmac_tb_clock_chip_host);
+ pmac_tb_clock_chip_host = NULL;
+ }
+
/* If we didn't start the second CPU, we must take
* it off the bus
*/
if (machine_is_compatible("MacRISC4") &&
num_online_cpus() < 2)
g5_phy_disable_cpu1();
-#endif /* CONFIG_POWER4 */
- if (ppc_md.progress) ppc_md.progress("core99_setup_cpu 0 done", 0x349);
+#endif /* CONFIG_PPC64 */
+
+ if (ppc_md.progress)
+ ppc_md.progress("core99_setup_cpu 0 done", 0x349);
}
}
Index: linux-work/arch/powerpc/sysdev/u3_iommu.c
===================================================================
--- linux-work.orig/arch/powerpc/sysdev/u3_iommu.c 2005-11-24 17:21:41.000000000 +1100
+++ linux-work/arch/powerpc/sysdev/u3_iommu.c 2005-12-13 18:13:12.000000000 +1100
@@ -282,7 +282,7 @@ void iommu_init_early_u3(void)
/* Find the DART in the device-tree */
dn = of_find_compatible_node(NULL, "dart", "u3-dart");
if (dn == NULL)
- return;
+ goto bail;
/* Setup low level TCE operations for the core IOMMU code */
ppc_md.tce_build = dart_build;
@@ -290,20 +290,23 @@ void iommu_init_early_u3(void)
ppc_md.tce_flush = dart_flush;
/* Initialize the DART HW */
- if (dart_init(dn)) {
- /* If init failed, use direct iommu and null setup functions */
- ppc_md.iommu_dev_setup = iommu_dev_setup_null;
- ppc_md.iommu_bus_setup = iommu_bus_setup_null;
-
- /* Setup pci_dma ops */
- pci_direct_iommu_init();
- } else {
+ if (dart_init(dn) == 0) {
ppc_md.iommu_dev_setup = iommu_dev_setup_u3;
ppc_md.iommu_bus_setup = iommu_bus_setup_u3;
/* Setup pci_dma ops */
pci_iommu_init();
+
+ return;
}
+
+ bail:
+ /* If init failed, use direct iommu and null setup functions */
+ ppc_md.iommu_dev_setup = iommu_dev_setup_null;
+ ppc_md.iommu_bus_setup = iommu_bus_setup_null;
+
+ /* Setup pci_dma ops */
+ pci_direct_iommu_init();
}
Index: linux-work/drivers/ide/ppc/pmac.c
===================================================================
--- linux-work.orig/drivers/ide/ppc/pmac.c 2005-12-13 17:51:20.000000000 +1100
+++ linux-work/drivers/ide/ppc/pmac.c 2005-12-13 18:13:12.000000000 +1100
@@ -1686,7 +1686,7 @@ pmac_ide_probe(void)
#else
macio_register_driver(&pmac_ide_macio_driver);
pci_register_driver(&pmac_ide_pci_driver);
-#endif
+#endif
}
#ifdef CONFIG_BLK_DEV_IDEDMA_PMAC
Index: linux-work/drivers/macintosh/smu.c
===================================================================
--- linux-work.orig/drivers/macintosh/smu.c 2005-11-24 17:18:43.000000000 +1100
+++ linux-work/drivers/macintosh/smu.c 2005-12-13 18:13:12.000000000 +1100
@@ -53,7 +53,7 @@
#undef DEBUG_SMU
#ifdef DEBUG_SMU
-#define DPRINTK(fmt, args...) do { udbg_printf(KERN_DEBUG fmt , ##args); } while (0)
+#define DPRINTK(fmt, args...) do { printk(KERN_DEBUG fmt , ##args); } while (0)
#else
#define DPRINTK(fmt, args...) do { } while (0)
#endif
@@ -909,10 +909,13 @@ static struct smu_sdbp_header *smu_creat
struct property *prop;
/* First query the partition info */
+ DPRINTK("SMU: Query partition infos ... (irq=%d)\n", smu->db_irq);
smu_queue_simple(&cmd, SMU_CMD_PARTITION_COMMAND, 2,
smu_done_complete, &comp,
SMU_CMD_PARTITION_LATEST, id);
wait_for_completion(&comp);
+ DPRINTK("SMU: done, status: %d, reply_len: %d\n",
+ cmd.cmd.status, cmd.cmd.reply_len);
/* Partition doesn't exist (or other error) */
if (cmd.cmd.status != 0 || cmd.cmd.reply_len != 6)
@@ -975,6 +978,8 @@ struct smu_sdbp_header *__smu_get_sdb_pa
sprintf(pname, "sdb-partition-%02x", id);
+ DPRINTK("smu_get_sdb_partition(%02x)\n", id);
+
if (interruptible) {
int rc;
rc = down_interruptible(&smu_part_access);
@@ -986,6 +991,7 @@ struct smu_sdbp_header *__smu_get_sdb_pa
part = (struct smu_sdbp_header *)get_property(smu->of_node,
pname, size);
if (part == NULL) {
+ DPRINTK("trying to extract from SMU ...\n");
part = smu_create_sdb_partition(id);
if (part != NULL && size)
*size = part->len << 2;
Index: linux-work/include/asm-powerpc/mpic.h
===================================================================
--- linux-work.orig/include/asm-powerpc/mpic.h 2005-12-13 18:02:03.000000000 +1100
+++ linux-work/include/asm-powerpc/mpic.h 2005-12-13 18:13:12.000000000 +1100
@@ -117,8 +117,9 @@ typedef int (*mpic_cascade_t)(struct pt_
struct mpic_irq_fixup
{
u8 __iomem *base;
+ u8 __iomem *applebase;
u32 data;
- unsigned int irq;
+ unsigned int index;
};
#endif /* CONFIG_MPIC_BROKEN_U3 */
Index: linux-work/include/asm/hardirq.h
===================================================================
--- linux-work.orig/include/asm/hardirq.h 2005-11-24 17:18:48.000000000 +1100
+++ linux-work/include/asm/hardirq.h 2005-12-13 18:13:12.000000000 +1100
@@ -11,13 +11,10 @@
*/
typedef struct {
unsigned int __softirq_pending; /* set_bit is used on this */
- unsigned int __last_jiffy_stamp;
} ____cacheline_aligned irq_cpustat_t;
#include <linux/irq_cpustat.h> /* Standard mappings for irq_cpustat_t above */
-#define last_jiffy_stamp(cpu) __IRQ_STAT((cpu), __last_jiffy_stamp)
-
static inline void ack_bad_irq(int irq)
{
printk(KERN_CRIT "illegal vector %d received!\n", irq);
^ permalink raw reply
* [PATCH] nvidiafb: Fixes for new G5
From: Benjamin Herrenschmidt @ 2005-12-13 7:12 UTC (permalink / raw)
To: Paul Mackerras; +Cc: linuxppc64-dev, linuxppc-dev list
This patch is not for inclusion in anything but powerpc git tree and
only temporarily as it's being dealt with separately for upstream merge
and is already in -mm. It fixes an issue in nvidiafb for the the G5s.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Index: linux-work/drivers/video/nvidia/nv_hw.c
===================================================================
--- linux-work.orig/drivers/video/nvidia/nv_hw.c 2005-11-24 17:18:46.000000000 +1100
+++ linux-work/drivers/video/nvidia/nv_hw.c 2005-12-07 19:14:52.000000000 +1100
@@ -848,7 +848,7 @@ void NVCalcStateExt(struct nvidia_par *p
int width,
int hDisplaySize, int height, int dotClock, int flags)
{
- int pixelDepth, VClk;
+ int pixelDepth, VClk = 0;
/*
* Save mode parameters.
*/
@@ -938,17 +938,26 @@ void NVLoadStateExt(struct nvidia_par *p
if (par->Architecture == NV_ARCH_04) {
NV_WR32(par->PFB, 0x0200, state->config);
- } else if ((par->Chipset & 0xfff0) == 0x0090) {
- for (i = 0; i < 15; i++) {
- NV_WR32(par->PFB, 0x0600 + (i * 0x10), 0);
- NV_WR32(par->PFB, 0x0604 + (i * 0x10), par->FbMapSize - 1);
- }
- } else {
+ } else if ((par->Architecture < NV_ARCH_40) ||
+ (par->Chipset & 0xfff0) == 0x0040) {
for (i = 0; i < 8; i++) {
NV_WR32(par->PFB, 0x0240 + (i * 0x10), 0);
- NV_WR32(par->PFB, 0x0244 + (i * 0x10), par->FbMapSize - 1);
+ NV_WR32(par->PFB, 0x0244 + (i * 0x10),
+ par->FbMapSize - 1);
}
- }
+ } else {
+ int regions = 12;
+
+ if (((par->Chipset & 0xfff0) == 0x0090) ||
+ ((par->Chipset & 0xfff0) == 0x01D0) ||
+ ((par->Chipset & 0xfff0) == 0x0290))
+ regions = 15;
+ for(i = 0; i < regions; i++) {
+ NV_WR32(par->PFB, 0x0600 + (i * 0x10), 0);
+ NV_WR32(par->PFB, 0x0604 + (i * 0x10),
+ par->FbMapSize - 1);
+ }
+ }
if (par->Architecture >= NV_ARCH_40) {
NV_WR32(par->PRAMIN, 0x0000 * 4, 0x80000010);
@@ -1182,11 +1191,17 @@ void NVLoadStateExt(struct nvidia_par *p
NV_WR32(par->PGRAPH, 0x0608, 0xFFFFFFFF);
} else {
if (par->Architecture >= NV_ARCH_40) {
+ u32 tmp;
+
NV_WR32(par->PGRAPH, 0x0084, 0x401287c0);
NV_WR32(par->PGRAPH, 0x008C, 0x60de8051);
NV_WR32(par->PGRAPH, 0x0090, 0x00008000);
NV_WR32(par->PGRAPH, 0x0610, 0x00be3c5f);
+ tmp = NV_RD32(par->REGS, 0x1540) & 0xff;
+ for(i = 0; tmp && !(tmp & 1); tmp >>= 1, i++);
+ NV_WR32(par->PGRAPH, 0x5000, i);
+
if ((par->Chipset & 0xfff0) == 0x0040) {
NV_WR32(par->PGRAPH, 0x09b0,
0x83280fff);
@@ -1211,6 +1226,7 @@ void NVLoadStateExt(struct nvidia_par *p
0xffff7fff);
break;
case 0x00C0:
+ case 0x0120:
NV_WR32(par->PGRAPH, 0x0828,
0x007596ff);
NV_WR32(par->PGRAPH, 0x082C,
@@ -1245,6 +1261,7 @@ void NVLoadStateExt(struct nvidia_par *p
0x00100000);
break;
case 0x0090:
+ case 0x0290:
NV_WR32(par->PRAMDAC, 0x0608,
NV_RD32(par->PRAMDAC, 0x0608) |
0x00100000);
@@ -1310,14 +1327,44 @@ void NVLoadStateExt(struct nvidia_par *p
}
}
- if ((par->Chipset & 0xfff0) == 0x0090) {
- for (i = 0; i < 60; i++)
- NV_WR32(par->PGRAPH, 0x0D00 + i,
- NV_RD32(par->PFB, 0x0600 + i));
+ if ((par->Architecture < NV_ARCH_40) ||
+ ((par->Chipset & 0xfff0) == 0x0040)) {
+ for (i = 0; i < 32; i++) {
+ NV_WR32(par->PGRAPH, 0x0900 + i*4,
+ NV_RD32(par->PFB, 0x0240 +i*4));
+ NV_WR32(par->PGRAPH, 0x6900 + i*4,
+ NV_RD32(par->PFB, 0x0240 +i*4));
+ }
} else {
- for (i = 0; i < 32; i++)
- NV_WR32(par->PGRAPH, 0x0900 + i,
- NV_RD32(par->PFB, 0x0240 + i));
+ if (((par->Chipset & 0xfff0) == 0x0090) ||
+ ((par->Chipset & 0xfff0) == 0x01D0) ||
+ ((par->Chipset & 0xfff0) == 0x0290)) {
+ for (i = 0; i < 60; i++) {
+ NV_WR32(par->PGRAPH,
+ 0x0D00 + i*4,
+ NV_RD32(par->PFB,
+ 0x0600 + i*4));
+ NV_WR32(par->PGRAPH,
+ 0x6900 + i*4,
+ NV_RD32(par->PFB,
+ 0x0600 + i*4));
+ }
+ } else {
+ for (i = 0; i < 48; i++) {
+ NV_WR32(par->PGRAPH,
+ 0x0900 + i*4,
+ NV_RD32(par->PFB,
+ 0x0600 + i*4));
+ if(((par->Chipset & 0xfff0)
+ != 0x0160) &&
+ ((par->Chipset & 0xfff0)
+ != 0x0220))
+ NV_WR32(par->PGRAPH,
+ 0x6900 + i*4,
+ NV_RD32(par->PFB,
+ 0x0600 + i*4));
+ }
+ }
}
if (par->Architecture >= NV_ARCH_40) {
@@ -1338,7 +1385,9 @@ void NVLoadStateExt(struct nvidia_par *p
NV_WR32(par->PGRAPH, 0x0868,
par->FbMapSize - 1);
} else {
- if((par->Chipset & 0xfff0) == 0x0090) {
+ if ((par->Chipset & 0xfff0) == 0x0090 ||
+ (par->Chipset & 0xfff0) == 0x01D0 ||
+ (par->Chipset & 0xfff0) == 0x0290) {
NV_WR32(par->PGRAPH, 0x0DF0,
NV_RD32(par->PFB, 0x0200));
NV_WR32(par->PGRAPH, 0x0DF4,
Index: linux-work/drivers/video/nvidia/nv_setup.c
===================================================================
--- linux-work.orig/drivers/video/nvidia/nv_setup.c 2005-11-24 17:18:46.000000000 +1100
+++ linux-work/drivers/video/nvidia/nv_setup.c 2005-12-07 19:09:01.000000000 +1100
@@ -285,7 +285,6 @@ static void nv10GetConfig(struct nvidia_
par->CrystalFreqKHz = 27000;
}
- par->CursorStart = (par->RamAmountKBytes - 96) * 1024;
par->CURSOR = NULL; /* can't set this here */
par->MinVClockFreqKHz = 12000;
par->MaxVClockFreqKHz = par->twoStagePLL ? 400000 : 350000;
@@ -382,6 +381,8 @@ void NVCommonSetup(struct fb_info *info)
case 0x0146:
case 0x0147:
case 0x0148:
+ case 0x0098:
+ case 0x0099:
mobile = 1;
break;
default:
Index: linux-work/drivers/video/nvidia/nvidia.c
===================================================================
--- linux-work.orig/drivers/video/nvidia/nvidia.c 2005-11-24 17:18:46.000000000 +1100
+++ linux-work/drivers/video/nvidia/nvidia.c 2005-12-07 18:52:36.000000000 +1100
@@ -1485,6 +1485,8 @@ static u32 __devinit nvidia_get_arch(str
case 0x0210:
case 0x0220:
case 0x0230:
+ case 0x0290:
+ case 0x0390:
arch = NV_ARCH_40;
break;
case 0x0020: /* TNT, TNT2 */
@@ -1581,10 +1583,15 @@ static int __devinit nvidiafb_probe(stru
if (par->FbMapSize > 64 * 1024 * 1024)
par->FbMapSize = 64 * 1024 * 1024;
- par->FbUsableSize = par->FbMapSize - (128 * 1024);
+ if(par->Architecture >= NV_ARCH_40)
+ par->FbUsableSize = par->FbMapSize - (560 * 1024);
+ else
+ par->FbUsableSize = par->FbMapSize - (128 * 1024);
par->ScratchBufferSize = (par->Architecture < NV_ARCH_10) ? 8 * 1024 :
16 * 1024;
par->ScratchBufferStart = par->FbUsableSize - par->ScratchBufferSize;
+ par->CursorStart = par->FbUsableSize + (32 * 1024);
+
info->screen_base = ioremap(nvidiafb_fix.smem_start, par->FbMapSize);
info->screen_size = par->FbUsableSize;
nvidiafb_fix.smem_len = par->RamAmountKBytes * 1024;
^ permalink raw reply
* [PATCH] PCI: Export pci_cfg_space_size
From: Benjamin Herrenschmidt @ 2005-12-13 7:09 UTC (permalink / raw)
To: Greg KH, Paul Mackerras; +Cc: linuxppc64-dev, linux-pci, linuxppc-dev list
The powerpc PCI code sets up the PCI tree without doing config space
accesses in most cases, from the firmware tree. However, it still wants
to call pci_cfg_space_size() under some conditions, thus it needs to
be made non-static (though I don't see a point to export it to modules).
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
Greg: I don't remember if I already sent you that patch or not, if not,
then heh, here it is :) The powerpc patch that requires this one will be
going next to the powerpc git tree and -mm, with a 2.6.16 target. Let me
know if there is any objection.
Paul: You need that for the new G5 support patch that I'm sending next.
Index: linux-work/drivers/pci/probe.c
===================================================================
--- linux-work.orig/drivers/pci/probe.c 2005-11-24 17:18:45.000000000 +1100
+++ linux-work/drivers/pci/probe.c 2005-12-08 10:15:41.000000000 +1100
@@ -678,7 +678,7 @@ static void pci_release_dev(struct devic
* reading the dword at 0x100 which must either be 0 or a valid extended
* capability header.
*/
-static int pci_cfg_space_size(struct pci_dev *dev)
+int pci_cfg_space_size(struct pci_dev *dev)
{
int pos;
u32 status;
Index: linux-work/include/linux/pci.h
===================================================================
--- linux-work.orig/include/linux/pci.h 2005-11-30 10:43:15.000000000 +1100
+++ linux-work/include/linux/pci.h 2005-12-08 10:16:19.000000000 +1100
@@ -515,6 +515,7 @@ int pci_scan_bridge(struct pci_bus *bus,
void pci_walk_bus(struct pci_bus *top, void (*cb)(struct pci_dev *, void *),
void *userdata);
+int pci_cfg_space_size(struct pci_dev *dev);
/* kmem_cache style wrapper around pci_alloc_consistent() */
^ permalink raw reply
* [PATCH] powerpc: Update MPIC workarounds
From: Benjamin Herrenschmidt @ 2005-12-13 7:04 UTC (permalink / raw)
To: Paul Mackerras; +Cc: linuxppc64-dev, linuxppc-dev list
From: Segher Boessenkool <segher@kernel.crashing.org>
Cleanup the MPIC IO-APIC workarounds, make them a bit more generic,
smaller and faster.
Signed-off-by: Segher Boessenkool <segher@kernel.crashing.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Index: linux-work/arch/powerpc/sysdev/mpic.c
===================================================================
--- linux-work.orig/arch/powerpc/sysdev/mpic.c 2005-12-06 16:17:43.000000000 +1100
+++ linux-work/arch/powerpc/sysdev/mpic.c 2005-12-07 13:30:45.000000000 +1100
@@ -175,57 +175,57 @@ static inline int mpic_is_ht_interrupt(s
return mpic->fixups[source_no].base != NULL;
}
+
static inline void mpic_apic_end_irq(struct mpic *mpic, unsigned int source_no)
{
struct mpic_irq_fixup *fixup = &mpic->fixups[source_no];
- u32 tmp;
spin_lock(&mpic->fixup_lock);
- writeb(0x11 + 2 * fixup->irq, fixup->base);
- tmp = readl(fixup->base + 2);
- writel(tmp | 0x80000000ul, fixup->base + 2);
- /* config writes shouldn't be posted but let's be safe ... */
- (void)readl(fixup->base + 2);
+ writeb(0x11 + 2 * fixup->irq, fixup->base + 2);
+ writel(fixup->data, fixup->base + 4);
spin_unlock(&mpic->fixup_lock);
}
-static void __init mpic_amd8111_read_irq(struct mpic *mpic, u8 __iomem *devbase)
+static void __init mpic_scan_ioapic(struct mpic *mpic, u8 __iomem *devbase)
{
- int i, irq;
+ int i, irq, n;
u32 tmp;
+ u8 pos;
- printk(KERN_INFO "mpic: - Workarounds on AMD 8111 @ %p\n", devbase);
+ for (pos = readb(devbase + 0x34); pos; pos = readb(devbase + pos + 1)) {
+ u8 id = readb(devbase + pos);
- for (i=0; i < 24; i++) {
- writeb(0x10 + 2*i, devbase + 0xf2);
- tmp = readl(devbase + 0xf4);
- if ((tmp & 0x1) || !(tmp & 0x20))
- continue;
- irq = (tmp >> 16) & 0xff;
- mpic->fixups[irq].irq = i;
- mpic->fixups[irq].base = devbase + 0xf2;
+ if (id == 0x08) {
+ id = readb(devbase + pos + 3);
+ if (id == 0x80)
+ break;
+ }
}
-}
-
-static void __init mpic_amd8131_read_irq(struct mpic *mpic, u8 __iomem *devbase)
-{
- int i, irq;
- u32 tmp;
+ if (pos == 0)
+ return;
+
+ printk(KERN_INFO "mpic: - Workarounds @ %p, pos = 0x%02x\n", devbase, pos);
- printk(KERN_INFO "mpic: - Workarounds on AMD 8131 @ %p\n", devbase);
+ devbase += pos;
- for (i=0; i < 4; i++) {
- writeb(0x10 + 2*i, devbase + 0xba);
- tmp = readl(devbase + 0xbc);
- if ((tmp & 0x1) || !(tmp & 0x20))
+ writeb(0x01, devbase + 2);
+ n = (readl(devbase + 4) >> 16) & 0xff;
+
+ for (i = 0; i <= n; i++) {
+ writeb(0x10 + 2 * i, devbase + 2);
+ tmp = readl(devbase + 4);
+ if ((tmp & 0x21) != 0x20)
continue;
irq = (tmp >> 16) & 0xff;
mpic->fixups[irq].irq = i;
- mpic->fixups[irq].base = devbase + 0xba;
+ mpic->fixups[irq].base = devbase;
+ writeb(0x11 + 2 * i, devbase + 2);
+ mpic->fixups[irq].data = readl(devbase + 4) | 0x80000000;
}
}
+
static void __init mpic_scan_ioapics(struct mpic *mpic)
{
unsigned int devfn;
@@ -241,21 +241,19 @@ static void __init mpic_scan_ioapics(str
/* Init spinlock */
spin_lock_init(&mpic->fixup_lock);
- /* Map u3 config space. We assume all IO-APICs are on the primary bus
- * and slot will never be above "0xf" so we only need to map 32k
+ /* Map U3 config space. We assume all IO-APICs are on the primary bus
+ * so we only need to map 64kB.
*/
- cfgspace = (unsigned char __iomem *)ioremap(0xf2000000, 0x8000);
+ cfgspace = ioremap(0xf2000000, 0x10000);
BUG_ON(cfgspace == NULL);
/* Now we scan all slots. We do a very quick scan, we read the header type,
* vendor ID and device ID only, that's plenty enough
*/
- for (devfn = 0; devfn < PCI_DEVFN(0x10,0); devfn ++) {
+ for (devfn = 0; devfn < 0x100; devfn++) {
u8 __iomem *devbase = cfgspace + (devfn << 8);
u8 hdr_type = readb(devbase + PCI_HEADER_TYPE);
u32 l = readl(devbase + PCI_VENDOR_ID);
- u16 vendor_id, device_id;
- int multifunc = 0;
DBG("devfn %x, l: %x\n", devfn, l);
@@ -264,21 +262,11 @@ static void __init mpic_scan_ioapics(str
l == 0x0000ffff || l == 0xffff0000)
goto next;
- /* Check if it's a multifunction device (only really used
- * to function 0 though
- */
- multifunc = !!(hdr_type & 0x80);
- vendor_id = l & 0xffff;
- device_id = (l >> 16) & 0xffff;
-
- /* If a known device, go to fixup setup code */
- if (vendor_id == PCI_VENDOR_ID_AMD && device_id == 0x7460)
- mpic_amd8111_read_irq(mpic, devbase);
- if (vendor_id == PCI_VENDOR_ID_AMD && device_id == 0x7450)
- mpic_amd8131_read_irq(mpic, devbase);
+ mpic_scan_ioapic(mpic, devbase);
+
next:
/* next device, if function 0 */
- if ((PCI_FUNC(devfn) == 0) && !multifunc)
+ if (PCI_FUNC(devfn) == 0 && (hdr_type & 0x80) == 0)
devfn += 7;
}
}
Index: linux-work/include/asm-powerpc/mpic.h
===================================================================
--- linux-work.orig/include/asm-powerpc/mpic.h 2005-11-24 17:18:48.000000000 +1100
+++ linux-work/include/asm-powerpc/mpic.h 2005-12-07 10:32:51.000000000 +1100
@@ -117,7 +117,8 @@ typedef int (*mpic_cascade_t)(struct pt_
struct mpic_irq_fixup
{
u8 __iomem *base;
- unsigned int irq;
+ u32 data;
+ unsigned int irq;
};
#endif /* CONFIG_MPIC_BROKEN_U3 */
^ permalink raw reply
* [PATCH] powerpc: Remove device_node addrs/n_addr
From: Benjamin Herrenschmidt @ 2005-12-13 7:01 UTC (permalink / raw)
To: Paul Mackerras; +Cc: linuxppc64-dev, linuxppc-dev list
The pre-parsed addrs/n_addrs fields in struct device_node are finally
gone. Remove the dodgy heuristics that did that parsing at boot and
remove the fields themselves since we now have a good replacement with
the new OF parsing code. This patch also fixes a bunch of drivers to use
the new code instead, so that at least pmac32, pseries, iseries and g5
defconfigs build.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Index: linux-work/arch/powerpc/kernel/prom.c
===================================================================
--- linux-work.orig/arch/powerpc/kernel/prom.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/arch/powerpc/kernel/prom.c 2005-12-13 17:51:20.000000000 +1100
@@ -57,21 +57,6 @@
#define DBG(fmt...)
#endif
-struct pci_reg_property {
- struct pci_address addr;
- u32 size_hi;
- u32 size_lo;
-};
-
-struct isa_reg_property {
- u32 space;
- u32 address;
- u32 size;
-};
-
-
-typedef int interpret_func(struct device_node *, unsigned long *,
- int, int, int);
static int __initdata dt_root_addr_cells;
static int __initdata dt_root_size_cells;
@@ -410,237 +395,19 @@ static int __devinit finish_node_interru
return 0;
}
-static int __devinit interpret_pci_props(struct device_node *np,
- unsigned long *mem_start,
- int naddrc, int nsizec,
- int measure_only)
-{
- struct address_range *adr;
- struct pci_reg_property *pci_addrs;
- int i, l, n_addrs;
-
- pci_addrs = (struct pci_reg_property *)
- get_property(np, "assigned-addresses", &l);
- if (!pci_addrs)
- return 0;
-
- n_addrs = l / sizeof(*pci_addrs);
-
- adr = prom_alloc(n_addrs * sizeof(*adr), mem_start);
- if (!adr)
- return -ENOMEM;
-
- if (measure_only)
- return 0;
-
- np->addrs = adr;
- np->n_addrs = n_addrs;
-
- for (i = 0; i < n_addrs; i++) {
- adr[i].space = pci_addrs[i].addr.a_hi;
- adr[i].address = pci_addrs[i].addr.a_lo |
- ((u64)pci_addrs[i].addr.a_mid << 32);
- adr[i].size = pci_addrs[i].size_lo;
- }
-
- return 0;
-}
-
-static int __init interpret_dbdma_props(struct device_node *np,
- unsigned long *mem_start,
- int naddrc, int nsizec,
- int measure_only)
-{
- struct reg_property32 *rp;
- struct address_range *adr;
- unsigned long base_address;
- int i, l;
- struct device_node *db;
-
- base_address = 0;
- if (!measure_only) {
- for (db = np->parent; db != NULL; db = db->parent) {
- if (!strcmp(db->type, "dbdma") && db->n_addrs != 0) {
- base_address = db->addrs[0].address;
- break;
- }
- }
- }
-
- rp = (struct reg_property32 *) get_property(np, "reg", &l);
- if (rp != 0 && l >= sizeof(struct reg_property32)) {
- i = 0;
- adr = (struct address_range *) (*mem_start);
- while ((l -= sizeof(struct reg_property32)) >= 0) {
- if (!measure_only) {
- adr[i].space = 2;
- adr[i].address = rp[i].address + base_address;
- adr[i].size = rp[i].size;
- }
- ++i;
- }
- np->addrs = adr;
- np->n_addrs = i;
- (*mem_start) += i * sizeof(struct address_range);
- }
-
- return 0;
-}
-
-static int __init interpret_macio_props(struct device_node *np,
- unsigned long *mem_start,
- int naddrc, int nsizec,
- int measure_only)
-{
- struct reg_property32 *rp;
- struct address_range *adr;
- unsigned long base_address;
- int i, l;
- struct device_node *db;
-
- base_address = 0;
- if (!measure_only) {
- for (db = np->parent; db != NULL; db = db->parent) {
- if (!strcmp(db->type, "mac-io") && db->n_addrs != 0) {
- base_address = db->addrs[0].address;
- break;
- }
- }
- }
-
- rp = (struct reg_property32 *) get_property(np, "reg", &l);
- if (rp != 0 && l >= sizeof(struct reg_property32)) {
- i = 0;
- adr = (struct address_range *) (*mem_start);
- while ((l -= sizeof(struct reg_property32)) >= 0) {
- if (!measure_only) {
- adr[i].space = 2;
- adr[i].address = rp[i].address + base_address;
- adr[i].size = rp[i].size;
- }
- ++i;
- }
- np->addrs = adr;
- np->n_addrs = i;
- (*mem_start) += i * sizeof(struct address_range);
- }
-
- return 0;
-}
-
-static int __init interpret_isa_props(struct device_node *np,
- unsigned long *mem_start,
- int naddrc, int nsizec,
- int measure_only)
-{
- struct isa_reg_property *rp;
- struct address_range *adr;
- int i, l;
-
- rp = (struct isa_reg_property *) get_property(np, "reg", &l);
- if (rp != 0 && l >= sizeof(struct isa_reg_property)) {
- i = 0;
- adr = (struct address_range *) (*mem_start);
- while ((l -= sizeof(struct isa_reg_property)) >= 0) {
- if (!measure_only) {
- adr[i].space = rp[i].space;
- adr[i].address = rp[i].address;
- adr[i].size = rp[i].size;
- }
- ++i;
- }
- np->addrs = adr;
- np->n_addrs = i;
- (*mem_start) += i * sizeof(struct address_range);
- }
-
- return 0;
-}
-
-static int __init interpret_root_props(struct device_node *np,
- unsigned long *mem_start,
- int naddrc, int nsizec,
- int measure_only)
-{
- struct address_range *adr;
- int i, l;
- unsigned int *rp;
- int rpsize = (naddrc + nsizec) * sizeof(unsigned int);
-
- rp = (unsigned int *) get_property(np, "linux,usable-memory", &l);
- if (rp == NULL)
- rp = (unsigned int *) get_property(np, "reg", &l);
-
- if (rp != 0 && l >= rpsize) {
- i = 0;
- adr = (struct address_range *) (*mem_start);
- while ((l -= rpsize) >= 0) {
- if (!measure_only) {
- adr[i].space = 0;
- adr[i].address = rp[naddrc - 1];
- adr[i].size = rp[naddrc + nsizec - 1];
- }
- ++i;
- rp += naddrc + nsizec;
- }
- np->addrs = adr;
- np->n_addrs = i;
- (*mem_start) += i * sizeof(struct address_range);
- }
-
- return 0;
-}
-
static int __devinit finish_node(struct device_node *np,
unsigned long *mem_start,
- interpret_func *ifunc,
- int naddrc, int nsizec,
int measure_only)
{
struct device_node *child;
- int *ip, rc = 0;
-
- /* get the device addresses and interrupts */
- if (ifunc != NULL)
- rc = ifunc(np, mem_start, naddrc, nsizec, measure_only);
- if (rc)
- goto out;
+ int rc = 0;
rc = finish_node_interrupts(np, mem_start, measure_only);
if (rc)
goto out;
- /* Look for #address-cells and #size-cells properties. */
- ip = (int *) get_property(np, "#address-cells", NULL);
- if (ip != NULL)
- naddrc = *ip;
- ip = (int *) get_property(np, "#size-cells", NULL);
- if (ip != NULL)
- nsizec = *ip;
-
- if (!strcmp(np->name, "device-tree") || np->parent == NULL)
- ifunc = interpret_root_props;
- else if (np->type == 0)
- ifunc = NULL;
- else if (!strcmp(np->type, "pci") || !strcmp(np->type, "vci"))
- ifunc = interpret_pci_props;
- else if (!strcmp(np->type, "dbdma"))
- ifunc = interpret_dbdma_props;
- else if (!strcmp(np->type, "mac-io") || ifunc == interpret_macio_props)
- ifunc = interpret_macio_props;
- else if (!strcmp(np->type, "isa"))
- ifunc = interpret_isa_props;
- else if (!strcmp(np->name, "uni-n") || !strcmp(np->name, "u3"))
- ifunc = interpret_root_props;
- else if (!((ifunc == interpret_dbdma_props
- || ifunc == interpret_macio_props)
- && (!strcmp(np->type, "escc")
- || !strcmp(np->type, "media-bay"))))
- ifunc = NULL;
-
for (child = np->child; child != NULL; child = child->sibling) {
- rc = finish_node(child, mem_start, ifunc,
- naddrc, nsizec, measure_only);
+ rc = finish_node(child, mem_start, measure_only);
if (rc)
goto out;
}
@@ -702,10 +469,10 @@ void __init finish_device_tree(void)
* reason and then remove those additional 16 bytes
*/
size = 16;
- finish_node(allnodes, &size, NULL, 0, 0, 1);
+ finish_node(allnodes, &size, 1);
size -= 16;
end = start = (unsigned long) __va(lmb_alloc(size, 128));
- finish_node(allnodes, &end, NULL, 0, 0, 0);
+ finish_node(allnodes, &end, 0);
BUG_ON(end != start + size);
DBG(" <- finish_device_tree\n");
@@ -1822,7 +1589,6 @@ static void of_node_release(struct kref
prop = next;
}
kfree(node->intrs);
- kfree(node->addrs);
kfree(node->full_name);
kfree(node->data);
kfree(node);
@@ -1904,9 +1670,7 @@ void of_detach_node(const struct device_
* This should probably be split up into smaller chunks.
*/
-static int of_finish_dynamic_node(struct device_node *node,
- unsigned long *unused1, int unused2,
- int unused3, int unused4)
+static int of_finish_dynamic_node(struct device_node *node)
{
struct device_node *parent = of_get_parent(node);
int err = 0;
@@ -1927,7 +1691,8 @@ static int of_finish_dynamic_node(struct
return -ENODEV;
/* fix up new node's linux_phandle field */
- if ((ibm_phandle = (unsigned int *)get_property(node, "ibm,phandle", NULL)))
+ if ((ibm_phandle = (unsigned int *)get_property(node,
+ "ibm,phandle", NULL)))
node->linux_phandle = *ibm_phandle;
out:
@@ -1942,7 +1707,9 @@ static int prom_reconfig_notifier(struct
switch (action) {
case PSERIES_RECONFIG_ADD:
- err = finish_node(node, NULL, of_finish_dynamic_node, 0, 0, 0);
+ err = of_finish_dynamic_node(node);
+ if (!err)
+ finish_node(node, NULL, 0);
if (err < 0) {
printk(KERN_ERR "finish_node returned %d\n", err);
err = NOTIFY_BAD;
@@ -2016,175 +1783,4 @@ int prom_add_property(struct device_node
return 0;
}
-/* I quickly hacked that one, check against spec ! */
-static inline unsigned long
-bus_space_to_resource_flags(unsigned int bus_space)
-{
- u8 space = (bus_space >> 24) & 0xf;
- if (space == 0)
- space = 0x02;
- if (space == 0x02)
- return IORESOURCE_MEM;
- else if (space == 0x01)
- return IORESOURCE_IO;
- else {
- printk(KERN_WARNING "prom.c: bus_space_to_resource_flags(), space: %x\n",
- bus_space);
- return 0;
- }
-}
-
-#ifdef CONFIG_PCI
-static struct resource *find_parent_pci_resource(struct pci_dev* pdev,
- struct address_range *range)
-{
- unsigned long mask;
- int i;
-
- /* Check this one */
- mask = bus_space_to_resource_flags(range->space);
- for (i=0; i<DEVICE_COUNT_RESOURCE; i++) {
- if ((pdev->resource[i].flags & mask) == mask &&
- pdev->resource[i].start <= range->address &&
- pdev->resource[i].end > range->address) {
- if ((range->address + range->size - 1) > pdev->resource[i].end) {
- /* Add better message */
- printk(KERN_WARNING "PCI/OF resource overlap !\n");
- return NULL;
- }
- break;
- }
- }
- if (i == DEVICE_COUNT_RESOURCE)
- return NULL;
- return &pdev->resource[i];
-}
-
-/*
- * Request an OF device resource. Currently handles child of PCI devices,
- * or other nodes attached to the root node. Ultimately, put some
- * link to resources in the OF node.
- */
-struct resource *request_OF_resource(struct device_node* node, int index,
- const char* name_postfix)
-{
- struct pci_dev* pcidev;
- u8 pci_bus, pci_devfn;
- unsigned long iomask;
- struct device_node* nd;
- struct resource* parent;
- struct resource *res = NULL;
- int nlen, plen;
-
- if (index >= node->n_addrs)
- goto fail;
-
- /* Sanity check on bus space */
- iomask = bus_space_to_resource_flags(node->addrs[index].space);
- if (iomask & IORESOURCE_MEM)
- parent = &iomem_resource;
- else if (iomask & IORESOURCE_IO)
- parent = &ioport_resource;
- else
- goto fail;
-
- /* Find a PCI parent if any */
- nd = node;
- pcidev = NULL;
- while (nd) {
- if (!pci_device_from_OF_node(nd, &pci_bus, &pci_devfn))
- pcidev = pci_find_slot(pci_bus, pci_devfn);
- if (pcidev) break;
- nd = nd->parent;
- }
- if (pcidev)
- parent = find_parent_pci_resource(pcidev, &node->addrs[index]);
- if (!parent) {
- printk(KERN_WARNING "request_OF_resource(%s), parent not found\n",
- node->name);
- goto fail;
- }
-
- res = __request_region(parent, node->addrs[index].address,
- node->addrs[index].size, NULL);
- if (!res)
- goto fail;
- nlen = strlen(node->name);
- plen = name_postfix ? strlen(name_postfix) : 0;
- res->name = (const char *)kmalloc(nlen+plen+1, GFP_KERNEL);
- if (res->name) {
- strcpy((char *)res->name, node->name);
- if (plen)
- strcpy((char *)res->name+nlen, name_postfix);
- }
- return res;
-fail:
- return NULL;
-}
-EXPORT_SYMBOL(request_OF_resource);
-int release_OF_resource(struct device_node *node, int index)
-{
- struct pci_dev* pcidev;
- u8 pci_bus, pci_devfn;
- unsigned long iomask, start, end;
- struct device_node* nd;
- struct resource* parent;
- struct resource *res = NULL;
-
- if (index >= node->n_addrs)
- return -EINVAL;
-
- /* Sanity check on bus space */
- iomask = bus_space_to_resource_flags(node->addrs[index].space);
- if (iomask & IORESOURCE_MEM)
- parent = &iomem_resource;
- else if (iomask & IORESOURCE_IO)
- parent = &ioport_resource;
- else
- return -EINVAL;
-
- /* Find a PCI parent if any */
- nd = node;
- pcidev = NULL;
- while(nd) {
- if (!pci_device_from_OF_node(nd, &pci_bus, &pci_devfn))
- pcidev = pci_find_slot(pci_bus, pci_devfn);
- if (pcidev) break;
- nd = nd->parent;
- }
- if (pcidev)
- parent = find_parent_pci_resource(pcidev, &node->addrs[index]);
- if (!parent) {
- printk(KERN_WARNING "release_OF_resource(%s), parent not found\n",
- node->name);
- return -ENODEV;
- }
-
- /* Find us in the parent and its childs */
- res = parent->child;
- start = node->addrs[index].address;
- end = start + node->addrs[index].size - 1;
- while (res) {
- if (res->start == start && res->end == end &&
- (res->flags & IORESOURCE_BUSY))
- break;
- if (res->start <= start && res->end >= end)
- res = res->child;
- else
- res = res->sibling;
- }
- if (!res)
- return -ENODEV;
-
- if (res->name) {
- kfree(res->name);
- res->name = NULL;
- }
- release_resource(res);
- kfree(res);
-
- return 0;
-}
-EXPORT_SYMBOL(release_OF_resource);
-#endif /* CONFIG_PCI */
Index: linux-work/include/asm-powerpc/prom.h
===================================================================
--- linux-work.orig/include/asm-powerpc/prom.h 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/include/asm-powerpc/prom.h 2005-12-13 17:51:20.000000000 +1100
@@ -65,49 +65,11 @@ struct boot_param_header
typedef u32 phandle;
typedef u32 ihandle;
-struct address_range {
- unsigned long space;
- unsigned long address;
- unsigned long size;
-};
-
struct interrupt_info {
int line;
int sense; /* +ve/-ve logic, edge or level, etc. */
};
-struct pci_address {
- u32 a_hi;
- u32 a_mid;
- u32 a_lo;
-};
-
-struct isa_address {
- u32 a_hi;
- u32 a_lo;
-};
-
-struct isa_range {
- struct isa_address isa_addr;
- struct pci_address pci_addr;
- unsigned int size;
-};
-
-struct reg_property {
- unsigned long address;
- unsigned long size;
-};
-
-struct reg_property32 {
- unsigned int address;
- unsigned int size;
-};
-
-struct reg_property64 {
- u64 address;
- u64 size;
-};
-
struct property {
char *name;
int length;
@@ -120,8 +82,6 @@ struct device_node {
char *type;
phandle node;
phandle linux_phandle;
- int n_addrs;
- struct address_range *addrs;
int n_intrs;
struct interrupt_info *intrs;
char *full_name;
Index: linux-work/arch/powerpc/platforms/powermac/feature.c
===================================================================
--- linux-work.orig/arch/powerpc/platforms/powermac/feature.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/arch/powerpc/platforms/powermac/feature.c 2005-12-13 17:51:20.000000000 +1100
@@ -1445,20 +1445,55 @@ static long g5_i2s_enable(struct device_
/* Very crude implementation for now */
struct macio_chip *macio = &macio_chips[0];
unsigned long flags;
+ int cell;
+ u32 fcrs[3][3] = {
+ { 0,
+ K2_FCR1_I2S0_CELL_ENABLE |
+ K2_FCR1_I2S0_CLK_ENABLE_BIT | K2_FCR1_I2S0_ENABLE,
+ KL3_I2S0_CLK18_ENABLE
+ },
+ { KL0_SCC_A_INTF_ENABLE,
+ K2_FCR1_I2S1_CELL_ENABLE |
+ K2_FCR1_I2S1_CLK_ENABLE_BIT | K2_FCR1_I2S1_ENABLE,
+ KL3_I2S1_CLK18_ENABLE
+ },
+ { KL0_SCC_B_INTF_ENABLE,
+ SH_FCR1_I2S2_CELL_ENABLE |
+ SH_FCR1_I2S2_CLK_ENABLE_BIT | SH_FCR1_I2S2_ENABLE,
+ SH_FCR3_I2S2_CLK18_ENABLE
+ },
+ };
- if (value == 0)
- return 0; /* don't disable yet */
+ if (macio->type != macio_keylargo2 /* && macio->type != macio_shasta*/)
+ return -ENODEV;
+ if (strncmp(node->name, "i2s-", 4))
+ return -ENODEV;
+ cell = node->name[4] - 'a';
+ switch(cell) {
+ case 0:
+ case 1:
+ break;
+#if 0
+ case 2:
+ if (macio->type == macio_shasta)
+ break;
+#endif
+ default:
+ return -ENODEV;
+ }
LOCK(flags);
- MACIO_BIS(KEYLARGO_FCR3, KL3_CLK45_ENABLE | KL3_CLK49_ENABLE |
- KL3_I2S0_CLK18_ENABLE);
- udelay(10);
- MACIO_BIS(KEYLARGO_FCR1, K2_FCR1_I2S0_CELL_ENABLE |
- K2_FCR1_I2S0_CLK_ENABLE_BIT | K2_FCR1_I2S0_ENABLE);
+ if (value) {
+ MACIO_BIC(KEYLARGO_FCR0, fcrs[cell][0]);
+ MACIO_BIS(KEYLARGO_FCR1, fcrs[cell][1]);
+ MACIO_BIS(KEYLARGO_FCR3, fcrs[cell][2]);
+ } else {
+ MACIO_BIC(KEYLARGO_FCR3, fcrs[cell][2]);
+ MACIO_BIC(KEYLARGO_FCR1, fcrs[cell][1]);
+ MACIO_BIS(KEYLARGO_FCR0, fcrs[cell][0]);
+ }
udelay(10);
- MACIO_BIC(KEYLARGO_FCR1, K2_FCR1_I2S0_RESET);
UNLOCK(flags);
- udelay(10);
return 0;
}
@@ -2960,26 +2995,6 @@ pmac_feature_init(void)
set_initial_features();
}
-int __init pmac_feature_late_init(void)
-{
-#if 0
- struct device_node *np;
-
- /* Request some resources late */
- if (uninorth_node)
- request_OF_resource(uninorth_node, 0, NULL);
- np = find_devices("hammerhead");
- if (np)
- request_OF_resource(np, 0, NULL);
- np = find_devices("interrupt-controller");
- if (np)
- request_OF_resource(np, 0, NULL);
-#endif
- return 0;
-}
-
-device_initcall(pmac_feature_late_init);
-
#if 0
static void dump_HT_speeds(char *name, u32 cfg, u32 frq)
{
Index: linux-work/arch/powerpc/platforms/powermac/pic.c
===================================================================
--- linux-work.orig/arch/powerpc/platforms/powermac/pic.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/arch/powerpc/platforms/powermac/pic.c 2005-12-13 17:51:20.000000000 +1100
@@ -5,8 +5,8 @@
* in a separate file
*
* Copyright (C) 1997 Paul Mackerras (paulus@samba.org)
- *
- * Maintained by Benjamin Herrenschmidt (benh@kernel.crashing.org)
+ * Copyright (C) 2005 Benjamin Herrenschmidt (benh@kernel.crashing.org)
+ * IBM, Corp.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
@@ -54,12 +54,7 @@ struct pmac_irq_hw {
};
/* Default addresses */
-static volatile struct pmac_irq_hw *pmac_irq_hw[4] = {
- (struct pmac_irq_hw *) 0xf3000020,
- (struct pmac_irq_hw *) 0xf3000010,
- (struct pmac_irq_hw *) 0xf4000020,
- (struct pmac_irq_hw *) 0xf4000010,
-};
+static volatile struct pmac_irq_hw __iomem *pmac_irq_hw[4];
#define GC_LEVEL_MASK 0x3ff00000
#define OHARE_LEVEL_MASK 0x1ff00000
@@ -82,8 +77,7 @@ static unsigned long ppc_lost_interrupts
* since it can lose interrupts (see pmac_set_irq_mask).
* -- Cort
*/
-void
-__set_lost(unsigned long irq_nr, int nokick)
+void __set_lost(unsigned long irq_nr, int nokick)
{
if (!test_and_set_bit(irq_nr, ppc_lost_interrupts)) {
atomic_inc(&ppc_n_lost_interrupts);
@@ -92,8 +86,7 @@ __set_lost(unsigned long irq_nr, int nok
}
}
-static void
-pmac_mask_and_ack_irq(unsigned int irq_nr)
+static void pmac_mask_and_ack_irq(unsigned int irq_nr)
{
unsigned long bit = 1UL << (irq_nr & 0x1f);
int i = irq_nr >> 5;
@@ -224,8 +217,7 @@ static irqreturn_t gatwick_action(int cp
return IRQ_NONE;
}
-int
-pmac_get_irq(struct pt_regs *regs)
+static int pmac_get_irq(struct pt_regs *regs)
{
int irq;
unsigned long bits = 0;
@@ -256,34 +248,40 @@ pmac_get_irq(struct pt_regs *regs)
/* This routine will fix some missing interrupt values in the device tree
* on the gatwick mac-io controller used by some PowerBooks
+ *
+ * Walking of OF nodes could use a bit more fixing up here, but it's not
+ * very important as this is all boot time code on static portions of the
+ * device-tree.
+ *
+ * However, the modifications done to "intrs" will have to be removed and
+ * replaced with proper updates of the "interrupts" properties or
+ * AAPL,interrupts, yet to be decided, once the dynamic parsing is there.
*/
-static void __init
-pmac_fix_gatwick_interrupts(struct device_node *gw, int irq_base)
+static void __init pmac_fix_gatwick_interrupts(struct device_node *gw,
+ int irq_base)
{
struct device_node *node;
int count;
memset(gatwick_int_pool, 0, sizeof(gatwick_int_pool));
- node = gw->child;
count = 0;
- while(node)
- {
+ for (node = NULL; (node = of_get_next_child(gw, node)) != NULL;) {
/* Fix SCC */
- if (strcasecmp(node->name, "escc") == 0)
- if (node->child) {
- if (node->child->n_intrs < 3) {
- node->child->intrs = &gatwick_int_pool[count];
- count += 3;
- }
- node->child->n_intrs = 3;
- node->child->intrs[0].line = 15+irq_base;
- node->child->intrs[1].line = 4+irq_base;
- node->child->intrs[2].line = 5+irq_base;
- printk(KERN_INFO "irq: fixed SCC on second controller (%d,%d,%d)\n",
- node->child->intrs[0].line,
- node->child->intrs[1].line,
- node->child->intrs[2].line);
+ if ((strcasecmp(node->name, "escc") == 0) && node->child) {
+ if (node->child->n_intrs < 3) {
+ node->child->intrs = &gatwick_int_pool[count];
+ count += 3;
}
+ node->child->n_intrs = 3;
+ node->child->intrs[0].line = 15+irq_base;
+ node->child->intrs[1].line = 4+irq_base;
+ node->child->intrs[2].line = 5+irq_base;
+ printk(KERN_INFO "irq: fixed SCC on gatwick"
+ " (%d,%d,%d)\n",
+ node->child->intrs[0].line,
+ node->child->intrs[1].line,
+ node->child->intrs[2].line);
+ }
/* Fix media-bay & left SWIM */
if (strcasecmp(node->name, "media-bay") == 0) {
struct device_node* ya_node;
@@ -292,12 +290,11 @@ pmac_fix_gatwick_interrupts(struct devic
node->intrs = &gatwick_int_pool[count++];
node->n_intrs = 1;
node->intrs[0].line = 29+irq_base;
- printk(KERN_INFO "irq: fixed media-bay on second controller (%d)\n",
- node->intrs[0].line);
+ printk(KERN_INFO "irq: fixed media-bay on gatwick"
+ " (%d)\n", node->intrs[0].line);
ya_node = node->child;
- while(ya_node)
- {
+ while(ya_node) {
if (strcasecmp(ya_node->name, "floppy") == 0) {
if (ya_node->n_intrs < 2) {
ya_node->intrs = &gatwick_int_pool[count];
@@ -323,7 +320,6 @@ pmac_fix_gatwick_interrupts(struct devic
ya_node = ya_node->sibling;
}
}
- node = node->sibling;
}
if (count > 10) {
printk("WARNING !! Gatwick interrupt pool overflow\n");
@@ -338,45 +334,41 @@ pmac_fix_gatwick_interrupts(struct devic
* controller. If we find this second ohare, set it up and fix the
* interrupt value in the device tree for the ethernet chip.
*/
-static int __init enable_second_ohare(void)
+static void __init enable_second_ohare(struct device_node *np)
{
unsigned char bus, devfn;
unsigned short cmd;
- unsigned long addr;
- struct device_node *irqctrler = find_devices("pci106b,7");
struct device_node *ether;
- if (irqctrler == NULL || irqctrler->n_addrs <= 0)
- return -1;
- addr = (unsigned long) ioremap(irqctrler->addrs[0].address, 0x40);
- pmac_irq_hw[1] = (volatile struct pmac_irq_hw *)(addr + 0x20);
- max_irqs = 64;
- if (pci_device_from_OF_node(irqctrler, &bus, &devfn) == 0) {
- struct pci_controller* hose = pci_find_hose_for_OF_device(irqctrler);
- if (!hose)
- printk(KERN_ERR "Can't find PCI hose for OHare2 !\n");
- else {
- early_read_config_word(hose, bus, devfn, PCI_COMMAND, &cmd);
- cmd |= PCI_COMMAND_MEMORY | PCI_COMMAND_MASTER;
- cmd &= ~PCI_COMMAND_IO;
- early_write_config_word(hose, bus, devfn, PCI_COMMAND, cmd);
+ /* This code doesn't strictly belong here, it could be part of
+ * either the PCI initialisation or the feature code. It's kept
+ * here for historical reasons.
+ */
+ if (pci_device_from_OF_node(np, &bus, &devfn) == 0) {
+ struct pci_controller* hose =
+ pci_find_hose_for_OF_device(np);
+ if (!hose) {
+ printk(KERN_ERR "Can't find PCI hose for OHare2 !\n");
+ return;
}
+ early_read_config_word(hose, bus, devfn, PCI_COMMAND, &cmd);
+ cmd |= PCI_COMMAND_MEMORY | PCI_COMMAND_MASTER;
+ cmd &= ~PCI_COMMAND_IO;
+ early_write_config_word(hose, bus, devfn, PCI_COMMAND, cmd);
}
/* Fix interrupt for the modem/ethernet combo controller. The number
- in the device tree (27) is bogus (correct for the ethernet-only
- board but not the combo ethernet/modem board).
- The real interrupt is 28 on the second controller -> 28+32 = 60.
- */
- ether = find_devices("pci1011,14");
+ * in the device tree (27) is bogus (correct for the ethernet-only
+ * board but not the combo ethernet/modem board).
+ * The real interrupt is 28 on the second controller -> 28+32 = 60.
+ */
+ ether = of_find_node_by_name(NULL, "pci1011,14");
if (ether && ether->n_intrs > 0) {
ether->intrs[0].line = 60;
printk(KERN_INFO "irq: Fixed ethernet IRQ to %d\n",
ether->intrs[0].line);
}
-
- /* Return the interrupt number of the cascade */
- return irqctrler->intrs[0].line;
+ of_node_put(ether);
}
#ifdef CONFIG_XMON
@@ -394,189 +386,233 @@ static struct irqaction gatwick_cascade_
.mask = CPU_MASK_NONE,
.name = "cascade",
};
-#endif /* CONFIG_PPC32 */
-static int pmac_u3_cascade(struct pt_regs *regs, void *data)
+static void __init pmac_pic_probe_oldstyle(void)
{
- return mpic_get_one_irq((struct mpic *)data, regs);
-}
-
-void __init pmac_pic_init(void)
-{
- struct device_node *irqctrler = NULL;
- struct device_node *irqctrler2 = NULL;
- struct device_node *np;
-#ifdef CONFIG_PPC32
int i;
- unsigned long addr;
int irq_cascade = -1;
-#endif
- struct mpic *mpic1, *mpic2;
+ struct device_node *master = NULL;
+ struct device_node *slave = NULL;
+ u8 __iomem *addr;
+ struct resource r;
- /* We first try to detect Apple's new Core99 chipset, since mac-io
- * is quite different on those machines and contains an IBM MPIC2.
- */
- np = find_type_devices("open-pic");
- while (np) {
- if (np->parent && !strcmp(np->parent->name, "u3"))
- irqctrler2 = np;
- else
- irqctrler = np;
- np = np->next;
- }
- if (irqctrler != NULL && irqctrler->n_addrs > 0) {
- unsigned char senses[128];
-
- printk(KERN_INFO "PowerMac using OpenPIC irq controller at 0x%08x\n",
- (unsigned int)irqctrler->addrs[0].address);
- pmac_call_feature(PMAC_FTR_ENABLE_MPIC, irqctrler, 0, 0);
-
- prom_get_irq_senses(senses, 0, 128);
- mpic1 = mpic_alloc(irqctrler->addrs[0].address,
- MPIC_PRIMARY | MPIC_WANTS_RESET,
- 0, 0, 128, 252, senses, 128, " OpenPIC ");
- BUG_ON(mpic1 == NULL);
- mpic_init(mpic1);
-
- if (irqctrler2 != NULL && irqctrler2->n_intrs > 0 &&
- irqctrler2->n_addrs > 0) {
- printk(KERN_INFO "Slave OpenPIC at 0x%08x hooked on IRQ %d\n",
- (u32)irqctrler2->addrs[0].address,
- irqctrler2->intrs[0].line);
-
- pmac_call_feature(PMAC_FTR_ENABLE_MPIC, irqctrler2, 0, 0);
- prom_get_irq_senses(senses, 128, 128 + 124);
-
- /* We don't need to set MPIC_BROKEN_U3 here since we don't have
- * hypertransport interrupts routed to it
- */
- mpic2 = mpic_alloc(irqctrler2->addrs[0].address,
- MPIC_BIG_ENDIAN | MPIC_WANTS_RESET,
- 0, 128, 124, 0, senses, 124,
- " U3-MPIC ");
- BUG_ON(mpic2 == NULL);
- mpic_init(mpic2);
- mpic_setup_cascade(irqctrler2->intrs[0].line,
- pmac_u3_cascade, mpic2);
- }
-#if defined(CONFIG_XMON) && defined(CONFIG_PPC32)
- {
- struct device_node* pswitch;
- int nmi_irq;
-
- pswitch = find_devices("programmer-switch");
- if (pswitch && pswitch->n_intrs) {
- nmi_irq = pswitch->intrs[0].line;
- mpic_irq_set_priority(nmi_irq, 9);
- setup_irq(nmi_irq, &xmon_action);
- }
- }
-#endif /* defined(CONFIG_XMON) && defined(CONFIG_PPC32) */
- return;
- }
- irqctrler = NULL;
+ /* Set our get_irq function */
+ ppc_md.get_irq = pmac_get_irq;
-#ifdef CONFIG_PPC32
- /* Get the level/edge settings, assume if it's not
- * a Grand Central nor an OHare, then it's an Heathrow
- * (or Paddington).
+ /*
+ * Find the interrupt controller type & node
*/
- ppc_md.get_irq = pmac_get_irq;
- if (find_devices("gc"))
+
+ if ((master = of_find_node_by_name(NULL, "gc")) != NULL) {
+ max_irqs = max_real_irqs = 32;
level_mask[0] = GC_LEVEL_MASK;
- else if (find_devices("ohare")) {
+ } else if ((master = of_find_node_by_name(NULL, "ohare")) != NULL) {
+ max_irqs = max_real_irqs = 32;
level_mask[0] = OHARE_LEVEL_MASK;
+
/* We might have a second cascaded ohare */
- level_mask[1] = OHARE_LEVEL_MASK;
- } else {
+ slave = of_find_node_by_name(NULL, "pci106b,7");
+ if (slave) {
+ max_irqs = 64;
+ level_mask[1] = OHARE_LEVEL_MASK;
+ enable_second_ohare(slave);
+ }
+ } else if ((master = of_find_node_by_name(NULL, "mac-io")) != NULL) {
+ max_irqs = max_real_irqs = 64;
level_mask[0] = HEATHROW_LEVEL_MASK;
level_mask[1] = 0;
+
/* We might have a second cascaded heathrow */
- level_mask[2] = HEATHROW_LEVEL_MASK;
- level_mask[3] = 0;
- }
+ slave = of_find_node_by_name(master, "mac-io");
- /*
- * G3 powermacs and 1999 G3 PowerBooks have 64 interrupts,
- * 1998 G3 Series PowerBooks have 128,
- * other powermacs have 32.
- * The combo ethernet/modem card for the Powerstar powerbooks
- * (2400/3400/3500, ohare based) has a second ohare chip
- * effectively making a total of 64.
- */
- max_irqs = max_real_irqs = 32;
- irqctrler = find_devices("mac-io");
- if (irqctrler)
- {
- max_real_irqs = 64;
- if (irqctrler->next)
+ /* Check ordering of master & slave */
+ if (device_is_compatible(master, "gatwick")) {
+ struct device_node *tmp;
+ BUG_ON(slave == NULL);
+ tmp = master;
+ master = slave;
+ slave = tmp;
+ }
+
+ /* We found a slave */
+ if (slave) {
max_irqs = 128;
- else
- max_irqs = 64;
+ level_mask[2] = HEATHROW_LEVEL_MASK;
+ level_mask[3] = 0;
+ pmac_fix_gatwick_interrupts(slave, max_real_irqs);
+ }
}
+ BUG_ON(master == NULL);
+
+ /* Set the handler for the main PIC */
for ( i = 0; i < max_real_irqs ; i++ )
irq_desc[i].handler = &pmac_pic;
- /* get addresses of first controller */
- if (irqctrler) {
- if (irqctrler->n_addrs > 0) {
- addr = (unsigned long)
- ioremap(irqctrler->addrs[0].address, 0x40);
- for (i = 0; i < 2; ++i)
- pmac_irq_hw[i] = (volatile struct pmac_irq_hw*)
- (addr + (2 - i) * 0x10);
- }
+ /* Get addresses of first controller if we have a node for it */
+ BUG_ON(of_address_to_resource(master, 0, &r));
- /* get addresses of second controller */
- irqctrler = irqctrler->next;
- if (irqctrler && irqctrler->n_addrs > 0) {
- addr = (unsigned long)
- ioremap(irqctrler->addrs[0].address, 0x40);
- for (i = 2; i < 4; ++i)
- pmac_irq_hw[i] = (volatile struct pmac_irq_hw*)
- (addr + (4 - i) * 0x10);
- irq_cascade = irqctrler->intrs[0].line;
- if (device_is_compatible(irqctrler, "gatwick"))
- pmac_fix_gatwick_interrupts(irqctrler, max_real_irqs);
- }
- } else {
- /* older powermacs have a GC (grand central) or ohare at
- f3000000, with interrupt control registers at f3000020. */
- addr = (unsigned long) ioremap(0xf3000000, 0x40);
- pmac_irq_hw[0] = (volatile struct pmac_irq_hw *) (addr + 0x20);
+ /* Map interrupts of primary controller */
+ addr = (u8 __iomem *) ioremap(r.start, 0x40);
+ i = 0;
+ pmac_irq_hw[i++] = (volatile struct pmac_irq_hw __iomem *)
+ (addr + 0x20);
+ if (max_real_irqs > 32)
+ pmac_irq_hw[i++] = (volatile struct pmac_irq_hw __iomem *)
+ (addr + 0x10);
+ of_node_put(master);
+
+ printk(KERN_INFO "irq: Found primary Apple PIC %s for %d irqs\n",
+ master->full_name, max_real_irqs);
+
+ /* Map interrupts of cascaded controller */
+ if (slave && !of_address_to_resource(slave, 0, &r)) {
+ addr = (u8 __iomem *)ioremap(r.start, 0x40);
+ pmac_irq_hw[i++] = (volatile struct pmac_irq_hw __iomem *)
+ (addr + 0x20);
+ if (max_irqs > 64)
+ pmac_irq_hw[i++] =
+ (volatile struct pmac_irq_hw __iomem *)
+ (addr + 0x10);
+ irq_cascade = slave->intrs[0].line;
+
+ printk(KERN_INFO "irq: Found slave Apple PIC %s for %d irqs"
+ " cascade: %d\n", slave->full_name,
+ max_irqs - max_real_irqs, irq_cascade);
}
-
- /* PowerBooks 3400 and 3500 can have a second controller in a second
- ohare chip, on the combo ethernet/modem card */
- if (machine_is_compatible("AAPL,3400/2400")
- || machine_is_compatible("AAPL,3500"))
- irq_cascade = enable_second_ohare();
+ of_node_put(slave);
/* disable all interrupts in all controllers */
for (i = 0; i * 32 < max_irqs; ++i)
out_le32(&pmac_irq_hw[i]->enable, 0);
+
/* mark level interrupts */
for (i = 0; i < max_irqs; i++)
if (level_mask[i >> 5] & (1UL << (i & 0x1f)))
irq_desc[i].status = IRQ_LEVEL;
- /* get interrupt line of secondary interrupt controller */
- if (irq_cascade >= 0) {
- printk(KERN_INFO "irq: secondary controller on irq %d\n",
- (int)irq_cascade);
+ /* Setup handlers for secondary controller and hook cascade irq*/
+ if (slave) {
for ( i = max_real_irqs ; i < max_irqs ; i++ )
irq_desc[i].handler = &gatwick_pic;
setup_irq(irq_cascade, &gatwick_cascade_action);
}
- printk("System has %d possible interrupts\n", max_irqs);
- if (max_irqs != max_real_irqs)
- printk(KERN_DEBUG "%d interrupts on main controller\n",
- max_real_irqs);
-
+ printk(KERN_INFO "irq: System has %d possible interrupts\n", max_irqs);
#ifdef CONFIG_XMON
setup_irq(20, &xmon_action);
-#endif /* CONFIG_XMON */
-#endif /* CONFIG_PPC32 */
+#endif
+}
+#endif /* CONFIG_PPC32 */
+
+static int pmac_u3_cascade(struct pt_regs *regs, void *data)
+{
+ return mpic_get_one_irq((struct mpic *)data, regs);
+}
+
+static void __init pmac_pic_setup_mpic_nmi(struct mpic *mpic)
+{
+#if defined(CONFIG_XMON) && defined(CONFIG_PPC32)
+ struct device_node* pswitch;
+ int nmi_irq;
+
+ pswitch = of_find_node_by_name(NULL, "programmer-switch");
+ if (pswitch && pswitch->n_intrs) {
+ nmi_irq = pswitch->intrs[0].line;
+ mpic_irq_set_priority(nmi_irq, 9);
+ setup_irq(nmi_irq, &xmon_action);
+ }
+ of_node_put(pswitch);
+#endif /* defined(CONFIG_XMON) && defined(CONFIG_PPC32) */
+}
+
+static int __init pmac_pic_probe_mpic(void)
+{
+ struct mpic *mpic1, *mpic2;
+ struct device_node *np, *master = NULL, *slave = NULL;
+ unsigned char senses[128];
+ struct resource r;
+
+ /* We can have up to 2 MPICs cascaded */
+ for (np = NULL; (np = of_find_node_by_type(np, "open-pic"))
+ != NULL;) {
+ if (master == NULL &&
+ get_property(np, "interrupt-parent", NULL) != NULL)
+ master = of_node_get(np);
+ else if (slave == NULL)
+ slave = of_node_get(np);
+ if (master && slave)
+ break;
+ }
+
+ /* Check for bogus setups */
+ if (master == NULL && slave != NULL) {
+ master = slave;
+ slave = NULL;
+ }
+
+ /* Not found, default to good old pmac pic */
+ if (master == NULL)
+ return -ENODEV;
+
+ /* Set master handler */
+ ppc_md.get_irq = mpic_get_irq;
+
+ /* Setup master */
+ BUG_ON(of_address_to_resource(master, 0, &r));
+ pmac_call_feature(PMAC_FTR_ENABLE_MPIC, master, 0, 0);
+ prom_get_irq_senses(senses, 0, 128);
+ mpic1 = mpic_alloc(r.start, MPIC_PRIMARY | MPIC_WANTS_RESET,
+ 0, 0, 128, 252, senses, 128, " OpenPIC ");
+ BUG_ON(mpic1 == NULL);
+ mpic_init(mpic1);
+
+ /* Install NMI if any */
+ pmac_pic_setup_mpic_nmi(mpic1);
+
+ of_node_put(master);
+
+ /* No slave, let's go out */
+ if (slave == NULL || slave->n_intrs < 1)
+ return 0;
+
+ /* Setup slave, failures are non-fatal */
+ if (of_address_to_resource(slave, 0, &r)) {
+ printk(KERN_ERR "Can't get address of MPIC %s\n",
+ slave->full_name);
+ return 0;
+ }
+ pmac_call_feature(PMAC_FTR_ENABLE_MPIC, slave, 0, 0);
+ prom_get_irq_senses(senses, 128, 128 + 124);
+
+ /* We don't need to set MPIC_BROKEN_U3 here since we don't have
+ * hypertransport interrupts routed to it, at least not on currently
+ * supported machines, that may change.
+ */
+ mpic2 = mpic_alloc(r.start, MPIC_BIG_ENDIAN | MPIC_WANTS_RESET,
+ 0, 128, 124, 0, senses, 124, " U3-MPIC ");
+ if (mpic2 == NULL) {
+ printk(KERN_ERR "Can't create slave MPIC %s\n",
+ slave->full_name);
+ return 0;
+ }
+ mpic_init(mpic2);
+ mpic_setup_cascade(slave->intrs[0].line, pmac_u3_cascade, mpic2);
+
+ of_node_put(slave);
+ return 0;
+}
+
+
+void __init pmac_pic_init(void)
+{
+ /* We first try to detect Apple's new Core99 chipset, since mac-io
+ * is quite different on those machines and contains an IBM MPIC2.
+ */
+ if (pmac_pic_probe_mpic() == 0)
+ return;
+
+#ifdef CONFIG_PPC32
+ pmac_pic_probe_oldstyle();
+#endif
}
#if defined(CONFIG_PM) && defined(CONFIG_PPC32)
Index: linux-work/drivers/macintosh/via-cuda.c
===================================================================
--- linux-work.orig/drivers/macintosh/via-cuda.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/drivers/macintosh/via-cuda.c 2005-12-13 17:51:20.000000000 +1100
@@ -193,10 +193,6 @@ static int __init via_cuda_start(void)
if (via == NULL)
return -ENODEV;
-#ifdef CONFIG_PPC
- request_OF_resource(vias, 0, NULL);
-#endif
-
if (request_irq(CUDA_IRQ, cuda_interrupt, 0, "ADB", cuda_interrupt)) {
printk(KERN_ERR "cuda_init: can't get irq %d\n", CUDA_IRQ);
return -EAGAIN;
Index: linux-work/drivers/macintosh/via-pmu.c
===================================================================
--- linux-work.orig/drivers/macintosh/via-pmu.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/drivers/macintosh/via-pmu.c 2005-12-13 17:51:20.000000000 +1100
@@ -298,7 +298,7 @@ static struct backlight_controller pmu_b
int __init find_via_pmu(void)
{
- phys_addr_t taddr;
+ u64 taddr;
u32 *reg;
if (via != 0)
@@ -337,7 +337,7 @@ int __init find_via_pmu(void)
else if (device_is_compatible(vias->parent, "Keylargo")
|| device_is_compatible(vias->parent, "K2-Keylargo")) {
struct device_node *gpiop;
- phys_addr_t gaddr = 0;
+ u64 gaddr = OF_BAD_ADDR;
pmu_kind = PMU_KEYLARGO_BASED;
pmu_has_adb = (find_type_devices("adb") != NULL);
@@ -352,7 +352,7 @@ int __init find_via_pmu(void)
reg = (u32 *)get_property(gpiop, "reg", NULL);
if (reg)
gaddr = of_translate_address(gpiop, reg);
- if (gaddr != 0)
+ if (gaddr != OF_BAD_ADDR)
gpio_reg = ioremap(gaddr, 0x10);
}
if (gpio_reg == NULL)
@@ -479,9 +479,6 @@ static int __init via_pmu_dev_init(void)
if (vias == NULL)
return -ENODEV;
-#ifndef CONFIG_PPC64
- request_OF_resource(vias, 0, NULL);
-#endif
#ifdef CONFIG_PMAC_BACKLIGHT
/* Enable backlight */
register_backlight_controller(&pmu_backlight_controller, NULL, "pmu");
Index: linux-work/arch/powerpc/platforms/powermac/nvram.c
===================================================================
--- linux-work.orig/arch/powerpc/platforms/powermac/nvram.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/arch/powerpc/platforms/powermac/nvram.c 2005-12-13 17:51:20.000000000 +1100
@@ -514,7 +514,7 @@ static void core99_nvram_sync(void)
#endif
}
-static int __init core99_nvram_setup(struct device_node *dp)
+static int __init core99_nvram_setup(struct device_node *dp, unsigned long addr)
{
int i;
u32 gen_bank0, gen_bank1;
@@ -528,7 +528,7 @@ static int __init core99_nvram_setup(str
printk(KERN_ERR "nvram: can't allocate ram image\n");
return -ENOMEM;
}
- nvram_data = ioremap(dp->addrs[0].address, NVRAM_SIZE*2);
+ nvram_data = ioremap(addr, NVRAM_SIZE*2);
nvram_naddrs = 1; /* Make sure we get the correct case */
DBG("nvram: Checking bank 0...\n");
@@ -570,34 +570,48 @@ static int __init core99_nvram_setup(str
int __init pmac_nvram_init(void)
{
struct device_node *dp;
+ struct resource r1, r2;
+ unsigned int s1 = 0, s2 = 0;
int err = 0;
nvram_naddrs = 0;
- dp = find_devices("nvram");
+ dp = of_find_node_by_name(NULL, "nvram");
if (dp == NULL) {
printk(KERN_ERR "Can't find NVRAM device\n");
return -ENODEV;
}
- nvram_naddrs = dp->n_addrs;
+
+ /* Try to obtain an address */
+ if (of_address_to_resource(dp, 0, &r1) == 0) {
+ nvram_naddrs = 1;
+ s1 = (r1.end - r1.start) + 1;
+ if (of_address_to_resource(dp, 1, &r2) == 0) {
+ nvram_naddrs = 2;
+ s2 = (r2.end - r2.start) + 1;
+ }
+ }
+
is_core_99 = device_is_compatible(dp, "nvram,flash");
- if (is_core_99)
- err = core99_nvram_setup(dp);
+ if (is_core_99) {
+ err = core99_nvram_setup(dp, r1.start);
+ goto bail;
+ }
+
#ifdef CONFIG_PPC32
- else if (_machine == _MACH_chrp && nvram_naddrs == 1) {
- nvram_data = ioremap(dp->addrs[0].address + isa_mem_base,
- dp->addrs[0].size);
+ if (_machine == _MACH_chrp && nvram_naddrs == 1) {
+ nvram_data = ioremap(r1.start, s1);
nvram_mult = 1;
ppc_md.nvram_read_val = direct_nvram_read_byte;
ppc_md.nvram_write_val = direct_nvram_write_byte;
} else if (nvram_naddrs == 1) {
- nvram_data = ioremap(dp->addrs[0].address, dp->addrs[0].size);
- nvram_mult = (dp->addrs[0].size + NVRAM_SIZE - 1) / NVRAM_SIZE;
+ nvram_data = ioremap(r1.start, s1);
+ nvram_mult = (s1 + NVRAM_SIZE - 1) / NVRAM_SIZE;
ppc_md.nvram_read_val = direct_nvram_read_byte;
ppc_md.nvram_write_val = direct_nvram_write_byte;
} else if (nvram_naddrs == 2) {
- nvram_addr = ioremap(dp->addrs[0].address, dp->addrs[0].size);
- nvram_data = ioremap(dp->addrs[1].address, dp->addrs[1].size);
+ nvram_addr = ioremap(r1.start, s1);
+ nvram_data = ioremap(r2.start, s2);
ppc_md.nvram_read_val = indirect_nvram_read_byte;
ppc_md.nvram_write_val = indirect_nvram_write_byte;
} else if (nvram_naddrs == 0 && sys_ctrler == SYS_CTRLER_PMU) {
@@ -606,13 +620,15 @@ int __init pmac_nvram_init(void)
ppc_md.nvram_read_val = pmu_nvram_read_byte;
ppc_md.nvram_write_val = pmu_nvram_write_byte;
#endif /* CONFIG_ADB_PMU */
- }
-#endif
- else {
+ } else {
printk(KERN_ERR "Incompatible type of NVRAM\n");
- return -ENXIO;
+ err = -ENXIO;
}
- lookup_partitions();
+#endif /* CONFIG_PPC32 */
+bail:
+ of_node_put(dp);
+ if (err == 0)
+ lookup_partitions();
return err;
}
Index: linux-work/arch/powerpc/platforms/powermac/pci.c
===================================================================
--- linux-work.orig/arch/powerpc/platforms/powermac/pci.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/arch/powerpc/platforms/powermac/pci.c 2005-12-13 17:51:20.000000000 +1100
@@ -285,15 +285,13 @@ static struct pci_ops chaos_pci_ops =
};
static void __init setup_chaos(struct pci_controller *hose,
- struct reg_property *addr)
+ struct resource *addr)
{
/* assume a `chaos' bridge */
hose->ops = &chaos_pci_ops;
- hose->cfg_addr = ioremap(addr->address + 0x800000, 0x1000);
- hose->cfg_data = ioremap(addr->address + 0xc00000, 0x1000);
+ hose->cfg_addr = ioremap(addr->start + 0x800000, 0x1000);
+ hose->cfg_data = ioremap(addr->start + 0xc00000, 0x1000);
}
-#else
-#define setup_chaos(hose, addr)
#endif /* CONFIG_PPC32 */
#ifdef CONFIG_PPC64
@@ -356,9 +354,11 @@ static unsigned long u3_ht_cfg_access(st
/* For now, we don't self probe U3 HT bridge */
if (PCI_SLOT(devfn) == 0)
return 0;
- return ((unsigned long)hose->cfg_data) + U3_HT_CFA0(devfn, offset);
+ return ((unsigned long)hose->cfg_data) +
+ U3_HT_CFA0(devfn, offset);
} else
- return ((unsigned long)hose->cfg_data) + U3_HT_CFA1(bus, devfn, offset);
+ return ((unsigned long)hose->cfg_data) +
+ U3_HT_CFA1(bus, devfn, offset);
}
static int u3_ht_read_config(struct pci_bus *bus, unsigned int devfn,
@@ -532,7 +532,8 @@ static void __init init_p2pbridge(void)
}
if (early_read_config_word(hose, bus, devfn,
PCI_BRIDGE_CONTROL, &val) < 0) {
- printk(KERN_ERR "init_p2pbridge: couldn't read bridge control\n");
+ printk(KERN_ERR "init_p2pbridge: couldn't read bridge"
+ " control\n");
return;
}
val &= ~PCI_BRIDGE_CTL_MASTER_ABORT;
@@ -576,36 +577,38 @@ static void __init fixup_nec_usb2(void)
continue;
early_read_config_dword(hose, bus, devfn, 0xe4, &data);
if (data & 1UL) {
- printk("Found NEC PD720100A USB2 chip with disabled EHCI, fixing up...\n");
+ printk("Found NEC PD720100A USB2 chip with disabled"
+ " EHCI, fixing up...\n");
data &= ~1UL;
early_write_config_dword(hose, bus, devfn, 0xe4, data);
- early_write_config_byte(hose, bus, devfn | 2, PCI_INTERRUPT_LINE,
+ early_write_config_byte(hose, bus,
+ devfn | 2, PCI_INTERRUPT_LINE,
nec->intrs[0].line);
}
}
}
static void __init setup_bandit(struct pci_controller *hose,
- struct reg_property *addr)
+ struct resource *addr)
{
hose->ops = ¯isc_pci_ops;
- hose->cfg_addr = ioremap(addr->address + 0x800000, 0x1000);
- hose->cfg_data = ioremap(addr->address + 0xc00000, 0x1000);
+ hose->cfg_addr = ioremap(addr->start + 0x800000, 0x1000);
+ hose->cfg_data = ioremap(addr->start + 0xc00000, 0x1000);
init_bandit(hose);
}
static int __init setup_uninorth(struct pci_controller *hose,
- struct reg_property *addr)
+ struct resource *addr)
{
pci_assign_all_buses = 1;
has_uninorth = 1;
hose->ops = ¯isc_pci_ops;
- hose->cfg_addr = ioremap(addr->address + 0x800000, 0x1000);
- hose->cfg_data = ioremap(addr->address + 0xc00000, 0x1000);
+ hose->cfg_addr = ioremap(addr->start + 0x800000, 0x1000);
+ hose->cfg_data = ioremap(addr->start + 0xc00000, 0x1000);
/* We "know" that the bridge at f2000000 has the PCI slots. */
- return addr->address == 0xf2000000;
+ return addr->start == 0xf2000000;
}
-#endif
+#endif /* CONFIG_PPC32 */
#ifdef CONFIG_PPC64
static void __init setup_u3_agp(struct pci_controller* hose)
@@ -722,7 +725,7 @@ static void __init setup_u3_ht(struct pc
hose->mem_resources[cur-1].end = res->start - 1;
}
}
-#endif
+#endif /* CONFIG_PPC64 */
/*
* We assume that if we have a G3 powermac, we have one bridge called
@@ -733,24 +736,17 @@ static int __init add_bridge(struct devi
{
int len;
struct pci_controller *hose;
-#ifdef CONFIG_PPC32
- struct reg_property *addr;
-#endif
+ struct resource rsrc;
char *disp_name;
int *bus_range;
- int primary = 1;
+ int primary = 1, has_address = 0;
DBG("Adding PCI host bridge %s\n", dev->full_name);
-#ifdef CONFIG_PPC32
- /* XXX fix this */
- addr = (struct reg_property *) get_property(dev, "reg", &len);
- if (addr == NULL || len < sizeof(*addr)) {
- printk(KERN_WARNING "Can't use %s: no address\n",
- dev->full_name);
- return -ENODEV;
- }
-#endif
+ /* Fetch host bridge registers address */
+ has_address = (of_address_to_resource(dev, 0, &rsrc) == 0);
+
+ /* Get bus range if any */
bus_range = (int *) get_property(dev, "bus-range", &len);
if (bus_range == NULL || len < 2 * sizeof(int)) {
printk(KERN_WARNING "Can't get bus-range for %s, assume"
@@ -770,6 +766,8 @@ static int __init add_bridge(struct devi
hose->last_busno = bus_range ? bus_range[1] : 0xff;
disp_name = NULL;
+
+ /* 64 bits only bridges */
#ifdef CONFIG_PPC64
if (device_is_compatible(dev, "u3-agp")) {
setup_u3_agp(hose);
@@ -782,25 +780,30 @@ static int __init add_bridge(struct devi
}
printk(KERN_INFO "Found %s PCI host bridge. Firmware bus number: %d->%d\n",
disp_name, hose->first_busno, hose->last_busno);
-#else
+#endif /* CONFIG_PPC64 */
+
+ /* 32 bits only bridges */
+#ifdef CONFIG_PPC32
if (device_is_compatible(dev, "uni-north")) {
- primary = setup_uninorth(hose, addr);
+ primary = setup_uninorth(hose, &rsrc);
disp_name = "UniNorth";
} else if (strcmp(dev->name, "pci") == 0) {
/* XXX assume this is a mpc106 (grackle) */
setup_grackle(hose);
disp_name = "Grackle (MPC106)";
} else if (strcmp(dev->name, "bandit") == 0) {
- setup_bandit(hose, addr);
+ setup_bandit(hose, &rsrc);
disp_name = "Bandit";
} else if (strcmp(dev->name, "chaos") == 0) {
- setup_chaos(hose, addr);
+ setup_chaos(hose, &rsrc);
disp_name = "Chaos";
primary = 0;
}
- printk(KERN_INFO "Found %s PCI host bridge at 0x%08lx. Firmware bus number: %d->%d\n",
- disp_name, addr->address, hose->first_busno, hose->last_busno);
-#endif
+ printk(KERN_INFO "Found %s PCI host bridge at 0x%08lx. "
+ "Firmware bus number: %d->%d\n",
+ disp_name, rsrc.start, hose->first_busno, hose->last_busno);
+#endif /* CONFIG_PPC32 */
+
DBG(" ->Hose at 0x%p, cfg_addr=0x%p,cfg_data=0x%p\n",
hose, hose->cfg_addr, hose->cfg_data);
@@ -814,8 +817,7 @@ static int __init add_bridge(struct devi
return 0;
}
-static void __init
-pcibios_fixup_OF_interrupts(void)
+static void __init pcibios_fixup_OF_interrupts(void)
{
struct pci_dev* dev = NULL;
@@ -835,8 +837,7 @@ pcibios_fixup_OF_interrupts(void)
}
}
-void __init
-pmac_pcibios_fixup(void)
+void __init pmac_pcibios_fixup(void)
{
/* Fixup interrupts according to OF tree */
pcibios_fixup_OF_interrupts();
Index: linux-work/arch/powerpc/platforms/powermac/pmac.h
===================================================================
--- linux-work.orig/arch/powerpc/platforms/powermac/pmac.h 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/arch/powerpc/platforms/powermac/pmac.h 2005-12-13 17:51:20.000000000 +1100
@@ -42,10 +42,6 @@ extern void pmac_ide_init_hwif_ports(hw_
unsigned long data_port, unsigned long ctrl_port, int *irq);
extern int pmac_nvram_init(void);
-
-extern struct hw_interrupt_type pmac_pic;
-
-void pmac_pic_init(void);
-int pmac_get_irq(struct pt_regs *regs);
+extern void pmac_pic_init(void);
#endif /* __PMAC_H__ */
Index: linux-work/arch/powerpc/platforms/powermac/setup.c
===================================================================
--- linux-work.orig/arch/powerpc/platforms/powermac/setup.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/arch/powerpc/platforms/powermac/setup.c 2005-12-13 17:51:20.000000000 +1100
@@ -75,7 +75,6 @@
#include <asm/iommu.h>
#include <asm/smu.h>
#include <asm/pmc.h>
-#include <asm/mpic.h>
#include <asm/lmb.h>
#include <asm/udbg.h>
@@ -751,7 +750,7 @@ struct machdep_calls __initdata pmac_md
.init_early = pmac_init_early,
.show_cpuinfo = pmac_show_cpuinfo,
.init_IRQ = pmac_pic_init,
- .get_irq = mpic_get_irq, /* changed later */
+ .get_irq = NULL, /* changed later */
.pcibios_fixup = pmac_pcibios_fixup,
.restart = pmac_restart,
.power_off = pmac_power_off,
Index: linux-work/arch/powerpc/platforms/powermac/time.c
===================================================================
--- linux-work.orig/arch/powerpc/platforms/powermac/time.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/arch/powerpc/platforms/powermac/time.c 2005-12-13 17:51:20.000000000 +1100
@@ -258,15 +258,20 @@ int __init via_calibrate_decr(void)
volatile unsigned char __iomem *via;
int count = VIA_TIMER_FREQ_6 / 100;
unsigned int dstart, dend;
+ struct resource rsrc;
- vias = find_devices("via-cuda");
+ vias = of_find_node_by_name(NULL, "via-cuda");
if (vias == 0)
- vias = find_devices("via-pmu");
+ vias = of_find_node_by_name(NULL, "via-pmu");
if (vias == 0)
- vias = find_devices("via");
- if (vias == 0 || vias->n_addrs == 0)
+ vias = of_find_node_by_name(NULL, "via");
+ if (vias == 0 || of_address_to_resource(vias, 0, &rsrc))
return 0;
- via = ioremap(vias->addrs[0].address, vias->addrs[0].size);
+ via = ioremap(rsrc.start, rsrc.end - rsrc.start + 1);
+ if (via == NULL) {
+ printk(KERN_ERR "Failed to map VIA for timer calibration !\n");
+ return 0;
+ }
/* set timer 1 for continuous interrupts */
out_8(&via[ACR], (via[ACR] & ~T1MODE) | T1MODE_CONT);
Index: linux-work/drivers/macintosh/macio_asic.c
===================================================================
--- linux-work.orig/drivers/macintosh/macio_asic.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/drivers/macintosh/macio_asic.c 2005-12-13 17:51:20.000000000 +1100
@@ -256,42 +256,42 @@ static int macio_resource_quirks(struct
{
if (res->flags & IORESOURCE_MEM) {
/* Grand Central has too large resource 0 on some machines */
- if (index == 0 && !strcmp(np->name, "gc")) {
- np->addrs[0].size = 0x20000;
+ if (index == 0 && !strcmp(np->name, "gc"))
res->end = res->start + 0x1ffff;
- }
+
/* Airport has bogus resource 2 */
if (index >= 2 && !strcmp(np->name, "radio"))
return 1;
+
+#ifndef CONFIG_PPC64
/* DBDMAs may have bogus sizes */
- if ((res->start & 0x0001f000) == 0x00008000) {
- np->addrs[index].size = 0x100;
+ if ((res->start & 0x0001f000) == 0x00008000)
res->end = res->start + 0xff;
- }
+#endif /* CONFIG_PPC64 */
+
/* ESCC parent eats child resources. We could have added a
* level of hierarchy, but I don't really feel the need
* for it
*/
if (!strcmp(np->name, "escc"))
return 1;
+
/* ESCC has bogus resources >= 3 */
if (index >= 3 && !(strcmp(np->name, "ch-a") &&
strcmp(np->name, "ch-b")))
return 1;
+
/* Media bay has too many resources, keep only first one */
if (index > 0 && !strcmp(np->name, "media-bay"))
return 1;
+
/* Some older IDE resources have bogus sizes */
if (!(strcmp(np->name, "IDE") && strcmp(np->name, "ATA") &&
strcmp(np->type, "ide") && strcmp(np->type, "ata"))) {
- if (index == 0 && np->addrs[0].size > 0x1000) {
- np->addrs[0].size = 0x1000;
+ if (index == 0 && (res->end - res->start) > 0xfff)
res->end = res->start + 0xfff;
- }
- if (index == 1 && np->addrs[1].size > 0x100) {
- np->addrs[1].size = 0x100;
+ if (index == 1 && (res->end - res->start) > 0xff)
res->end = res->start + 0xff;
- }
}
}
return 0;
@@ -349,7 +349,7 @@ static void macio_setup_resources(struct
/* Currently, we consider failure as harmless, this may
* change in the future, once I've found all the device
* tree bugs in older machines & worked around them
-l */
+ */
if (insert_resource(parent_res, res)) {
printk(KERN_WARNING "Can't request resource "
"%d for MacIO device %s\n",
Index: linux-work/drivers/macintosh/mediabay.c
===================================================================
--- linux-work.orig/drivers/macintosh/mediabay.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/drivers/macintosh/mediabay.c 2005-12-13 17:51:20.000000000 +1100
@@ -647,6 +647,7 @@ static int __devinit media_bay_attach(st
struct media_bay_info* bay;
u32 __iomem *regbase;
struct device_node *ofnode;
+ unsigned long base;
int i;
ofnode = mdev->ofdev.node;
@@ -656,10 +657,11 @@ static int __devinit media_bay_attach(st
if (macio_request_resources(mdev, "media-bay"))
return -EBUSY;
/* Media bay registers are located at the beginning of the
- * mac-io chip, we get the parent address for now (hrm...)
+ * mac-io chip, for now, we trick and align down the first
+ * resource passed in
*/
- regbase = (u32 __iomem *)
- ioremap(ofnode->parent->addrs[0].address, 0x100);
+ base = macio_resource_start(mdev, 0) & 0xffff0000u;
+ regbase = (u32 __iomem *)ioremap(base, 0x100);
if (regbase == NULL) {
macio_release_resources(mdev);
return -ENOMEM;
Index: linux-work/drivers/ide/ppc/pmac.c
===================================================================
--- linux-work.orig/drivers/ide/ppc/pmac.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/drivers/ide/ppc/pmac.c 2005-12-13 17:51:20.000000000 +1100
@@ -1271,7 +1271,7 @@ static int
pmac_ide_setup_device(pmac_ide_hwif_t *pmif, ide_hwif_t *hwif)
{
struct device_node *np = pmif->node;
- int *bidp, i;
+ int *bidp;
pmif->cable_80 = 0;
pmif->broken_dma = pmif->broken_dma_warn = 0;
@@ -1430,7 +1430,7 @@ pmac_ide_macio_attach(struct macio_dev *
pmif = &pmac_ide[i];
hwif = &ide_hwifs[i];
- if (mdev->ofdev.node->n_addrs == 0) {
+ if (macio_resource_count(mdev) == 0) {
printk(KERN_WARNING "ide%d: no address for %s\n",
i, mdev->ofdev.node->full_name);
return -ENXIO;
Index: linux-work/drivers/scsi/mac53c94.c
===================================================================
--- linux-work.orig/drivers/scsi/mac53c94.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/drivers/scsi/mac53c94.c 2005-12-13 17:51:20.000000000 +1100
@@ -432,11 +432,12 @@ static int mac53c94_probe(struct macio_d
struct Scsi_Host *host;
void *dma_cmd_space;
unsigned char *clkprop;
- int proplen;
+ int proplen, rc = -ENODEV;
if (macio_resource_count(mdev) != 2 || macio_irq_count(mdev) != 2) {
- printk(KERN_ERR "mac53c94: expected 2 addrs and intrs (got %d/%d)\n",
- node->n_addrs, node->n_intrs);
+ printk(KERN_ERR "mac53c94: expected 2 addrs and intrs"
+ " (got %d/%d)\n",
+ macio_resource_count(mdev), macio_irq_count(mdev));
return -ENODEV;
}
@@ -448,6 +449,7 @@ static int mac53c94_probe(struct macio_d
host = scsi_host_alloc(&mac53c94_template, sizeof(struct fsc_state));
if (host == NULL) {
printk(KERN_ERR "mac53c94: couldn't register host");
+ rc = -ENOMEM;
goto out_release;
}
@@ -486,6 +488,7 @@ static int mac53c94_probe(struct macio_d
if (dma_cmd_space == 0) {
printk(KERN_ERR "mac53c94: couldn't allocate dma "
"command space for %s\n", node->full_name);
+ rc = -ENOMEM;
goto out_free;
}
state->dma_cmds = (struct dbdma_cmd *)DBDMA_ALIGN(dma_cmd_space);
@@ -495,18 +498,21 @@ static int mac53c94_probe(struct macio_d
mac53c94_init(state);
- if (request_irq(state->intr, do_mac53c94_interrupt, 0, "53C94", state)) {
+ if (request_irq(state->intr, do_mac53c94_interrupt, 0, "53C94",state)) {
printk(KERN_ERR "mac53C94: can't get irq %d for %s\n",
state->intr, node->full_name);
goto out_free_dma;
}
- /* XXX FIXME: handle failure */
- scsi_add_host(host, &mdev->ofdev.dev);
- scsi_scan_host(host);
+ rc = scsi_add_host(host, &mdev->ofdev.dev);
+ if (rc != 0)
+ goto out_release_irq;
+ scsi_scan_host(host);
return 0;
+ out_release_irq:
+ free_irq(state->intr, state);
out_free_dma:
kfree(state->dma_cmd_space);
out_free:
@@ -518,7 +524,7 @@ static int mac53c94_probe(struct macio_d
out_release:
macio_release_resources(mdev);
- return -ENODEV;
+ return rc;
}
static int mac53c94_remove(struct macio_dev *mdev)
Index: linux-work/drivers/scsi/mesh.c
===================================================================
--- linux-work.orig/drivers/scsi/mesh.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/drivers/scsi/mesh.c 2005-12-13 17:51:20.000000000 +1100
@@ -1869,7 +1869,8 @@ static int mesh_probe(struct macio_dev *
if (macio_resource_count(mdev) != 2 || macio_irq_count(mdev) != 2) {
printk(KERN_ERR "mesh: expected 2 addrs and 2 intrs"
- " (got %d,%d)\n", mesh->n_addrs, mesh->n_intrs);
+ " (got %d,%d)\n", macio_resource_count(mdev),
+ macio_irq_count(mdev));
return -ENODEV;
}
Index: linux-work/drivers/video/offb.c
===================================================================
--- linux-work.orig/drivers/video/offb.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/drivers/video/offb.c 2005-12-13 17:51:20.000000000 +1100
@@ -223,6 +223,7 @@ static int offb_blank(int blank, struct
int __init offb_init(void)
{
struct device_node *dp = NULL, *boot_disp = NULL;
+
#if defined(CONFIG_BOOTX_TEXT) && defined(CONFIG_PPC32)
struct device_node *macos_display = NULL;
#endif
@@ -234,60 +235,54 @@ int __init offb_init(void)
if (boot_infos != 0) {
unsigned long addr =
(unsigned long) boot_infos->dispDeviceBase;
+ u32 *addrp;
+ u64 daddr, dsize;
+ unsigned int flags;
+
/* find the device node corresponding to the macos display */
while ((dp = of_find_node_by_type(dp, "display"))) {
int i;
- /*
- * Grrr... It looks like the MacOS ATI driver
- * munges the assigned-addresses property (but
- * the AAPL,address value is OK).
- */
- if (strncmp(dp->name, "ATY,", 4) == 0
- && dp->n_addrs == 1) {
- unsigned int *ap =
- (unsigned int *) get_property(dp,
- "AAPL,address",
- NULL);
- if (ap != NULL) {
- dp->addrs[0].address = *ap;
- dp->addrs[0].size = 0x01000000;
- }
- }
/*
- * The LTPro on the Lombard powerbook has no addresses
- * on the display nodes, they are on their parent.
+ * Look for an AAPL,address property first.
*/
- if (dp->n_addrs == 0
- && device_is_compatible(dp, "ATY,264LTPro")) {
- int na;
- unsigned int *ap = (unsigned int *)
- get_property(dp, "AAPL,address", &na);
- if (ap != 0)
- for (na /= sizeof(unsigned int);
- na > 0; --na, ++ap)
- if (*ap <= addr
- && addr <
- *ap + 0x1000000)
- goto foundit;
+ unsigned int na;
+ unsigned int *ap =
+ (unsigned int *)get_property(dp, "AAPL,address",
+ &na);
+ if (ap != 0) {
+ for (na /= sizeof(unsigned int); na > 0;
+ --na, ++ap)
+ if (*ap <= addr &&
+ addr < *ap + 0x1000000) {
+ macos_display = dp;
+ goto foundit;
+ }
}
/*
* See if the display address is in one of the address
* ranges for this display.
*/
- for (i = 0; i < dp->n_addrs; ++i) {
- if (dp->addrs[i].address <= addr
- && addr <
- dp->addrs[i].address +
- dp->addrs[i].size)
+ i = 0;
+ for (;;) {
+ addrp = of_get_address(dp, i++, &dsize, &flags);
+ if (addrp == NULL)
break;
+ if (!(flags & IORESOURCE_MEM))
+ continue;
+ daddr = of_translate_address(dp, addrp);
+ if (daddr == OF_BAD_ADDR)
+ continue;
+ if (daddr <= addr && addr < (daddr + dsize)) {
+ macos_display = dp;
+ goto foundit;
+ }
}
- if (i < dp->n_addrs) {
- foundit:
+ foundit:
+ if (macos_display) {
printk(KERN_INFO "MacOS display is %s\n",
dp->full_name);
- macos_display = dp;
break;
}
}
@@ -326,8 +321,10 @@ static void __init offb_init_nodriver(st
int *pp, i;
unsigned int len;
int width = 640, height = 480, depth = 8, pitch;
- unsigned int rsize, *up;
- unsigned long address = 0;
+ unsigned int flags, rsize, *up;
+ u64 address = OF_BAD_ADDR;
+ u32 *addrp;
+ u64 asize;
if ((pp = (int *) get_property(dp, "depth", &len)) != NULL
&& len == sizeof(int))
@@ -363,7 +360,7 @@ static void __init offb_init_nodriver(st
break;
}
if (pdev) {
- for (i = 0; i < 6 && address == 0; i++) {
+ for (i = 0; i < 6 && address == OF_BAD_ADDR; i++) {
if ((pci_resource_flags(pdev, i) &
IORESOURCE_MEM) &&
(pci_resource_len(pdev, i) >= rsize))
@@ -374,27 +371,33 @@ static void __init offb_init_nodriver(st
}
#endif /* CONFIG_PCI */
- if (address == 0 &&
- (up = (unsigned *) get_property(dp, "address", &len)) != NULL &&
- len == sizeof(unsigned))
- address = (u_long) * up;
- if (address == 0) {
- for (i = 0; i < dp->n_addrs; ++i)
- if (dp->addrs[i].size >=
- pitch * height * depth / 8)
- break;
- if (i >= dp->n_addrs) {
+ /* This one is dodgy, we may drop it ... */
+ if (address == OF_BAD_ADDR &&
+ (up = (unsigned *) get_property(dp, "address", &len)) != NULL &&
+ len == sizeof(unsigned int))
+ address = (u64) * up;
+
+ if (address == OF_BAD_ADDR) {
+ for (i = 0; (addrp = of_get_address(dp, i, &asize, &flags))
+ != NULL; i++) {
+ if (!(flags & IORESOURCE_MEM))
+ continue;
+ if (asize >= pitch * height * depth / 8)
+ break;
+ }
+ if (addrp == NULL) {
printk(KERN_ERR
"no framebuffer address found for %s\n",
dp->full_name);
return;
}
-
- address = (u_long) dp->addrs[i].address;
-
-#ifdef CONFIG_PPC64
- address += ((struct pci_dn *)dp->data)->phb->pci_mem_offset;
-#endif
+ address = of_translate_address(dp, addrp);
+ if (address == OF_BAD_ADDR) {
+ printk(KERN_ERR
+ "can't translate framebuffer address for %s\n",
+ dp->full_name);
+ return;
+ }
/* kludge for valkyrie */
if (strcmp(dp->name, "valkyrie") == 0)
@@ -459,7 +462,9 @@ static void __init offb_init_fb(const ch
par->cmap_type = cmap_unknown;
if (depth == 8) {
- /* XXX kludge for ati */
+
+ /* Palette hacks disabled for now */
+#if 0
if (dp && !strncmp(name, "ATY,Rage128", 11)) {
unsigned long regbase = dp->addrs[2].address;
par->cmap_adr = ioremap(regbase, 0x1FFF);
@@ -490,6 +495,7 @@ static void __init offb_init_fb(const ch
par->cmap_adr = ioremap(regbase + 0x6000, 0x1000);
par->cmap_type = cmap_gxt2000;
}
+#endif
fix->visual = par->cmap_adr ? FB_VISUAL_PSEUDOCOLOR
: FB_VISUAL_STATIC_PSEUDOCOLOR;
} else
Index: linux-work/arch/powerpc/kernel/btext.c
===================================================================
--- linux-work.orig/arch/powerpc/kernel/btext.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/arch/powerpc/kernel/btext.c 2005-12-13 17:51:20.000000000 +1100
@@ -211,8 +211,6 @@ int __init btext_find_display(int allow_
struct device_node *np = NULL;
int rc = -ENODEV;
- printk("trying to initialize btext ...\n");
-
name = (char *)get_property(of_chosen, "linux,stdout-path", NULL);
if (name != NULL) {
np = of_find_node_by_path(name);
Index: linux-work/drivers/serial/pmac_zilog.c
===================================================================
--- linux-work.orig/drivers/serial/pmac_zilog.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/drivers/serial/pmac_zilog.c 2005-12-13 17:51:20.000000000 +1100
@@ -1431,11 +1431,14 @@ static int __init pmz_init_port(struct u
char name[1];
} *slots;
int len;
+ struct resource r_ports, r_rxdma, r_txdma;
/*
* Request & map chip registers
*/
- uap->port.mapbase = np->addrs[0].address;
+ if (of_address_to_resource(np, 0, &r_ports))
+ return -ENODEV;
+ uap->port.mapbase = r_ports.start;
uap->port.membase = ioremap(uap->port.mapbase, 0x1000);
uap->control_reg = uap->port.membase;
@@ -1445,16 +1448,20 @@ static int __init pmz_init_port(struct u
* Request & map DBDMA registers
*/
#ifdef HAS_DBDMA
- if (np->n_addrs >= 3 && np->n_intrs >= 3)
+ if (of_address_to_resource(np, 1, &r_txdma) == 0 &&
+ of_address_to_resource(np, 2, &r_rxdma) == 0)
uap->flags |= PMACZILOG_FLAG_HAS_DMA;
+#else
+ memset(&r_txdma, 0, sizeof(struct resource));
+ memset(&r_rxdma, 0, sizeof(struct resource));
#endif
if (ZS_HAS_DMA(uap)) {
- uap->tx_dma_regs = ioremap(np->addrs[np->n_addrs - 2].address, 0x1000);
+ uap->tx_dma_regs = ioremap(r_txdma.start, 0x100);
if (uap->tx_dma_regs == NULL) {
uap->flags &= ~PMACZILOG_FLAG_HAS_DMA;
goto no_dma;
}
- uap->rx_dma_regs = ioremap(np->addrs[np->n_addrs - 1].address, 0x1000);
+ uap->rx_dma_regs = ioremap(r_rxdma.start, 0x100);
if (uap->rx_dma_regs == NULL) {
iounmap(uap->tx_dma_regs);
uap->tx_dma_regs = NULL;
Index: linux-work/arch/powerpc/kernel/pci_64.c
===================================================================
--- linux-work.orig/arch/powerpc/kernel/pci_64.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/arch/powerpc/kernel/pci_64.c 2005-12-13 17:51:20.000000000 +1100
@@ -896,6 +896,25 @@ static void __devinit pci_process_ISA_OF
unsigned long phb_io_base_phys,
void __iomem * phb_io_base_virt)
{
+ /* Remove these asap */
+
+ struct pci_address {
+ u32 a_hi;
+ u32 a_mid;
+ u32 a_lo;
+ };
+
+ struct isa_address {
+ u32 a_hi;
+ u32 a_lo;
+ };
+
+ struct isa_range {
+ struct isa_address isa_addr;
+ struct pci_address pci_addr;
+ unsigned int size;
+ };
+
struct isa_range *range;
unsigned long pci_addr;
unsigned int isa_addr;
@@ -1330,8 +1349,9 @@ unsigned int pci_address_to_pio(phys_add
list_for_each_entry_safe(hose, tmp, &hose_list, list_node) {
if (address >= hose->io_base_phys &&
address < (hose->io_base_phys + hose->pci_io_size))
- return (unsigned int)hose->io_base_virt +
- (address - hose->io_base_phys);
+ return (unsigned int)
+ ((unsigned long)hose->io_base_virt +
+ (address - hose->io_base_phys));
}
return (unsigned int)-1;
}
Index: linux-work/sound/ppc/pmac.c
===================================================================
--- linux-work.orig/sound/ppc/pmac.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/sound/ppc/pmac.c 2005-12-13 17:51:20.000000000 +1100
@@ -817,21 +817,17 @@ static int snd_pmac_free(pmac_t *chip)
iounmap(chip->playback.dma);
if (chip->capture.dma)
iounmap(chip->capture.dma);
-#ifndef CONFIG_PPC64
+
if (chip->node) {
int i;
-
for (i = 0; i < 3; i++) {
- if (chip->of_requested & (1 << i)) {
- if (chip->is_k2)
- release_OF_resource(chip->node->parent,
- i);
- else
- release_OF_resource(chip->node, i);
- }
+ if (chip->requested & (1 << i))
+ release_mem_region(chip->rsrc[i].start,
+ chip->rsrc[i].end -
+ chip->rsrc[i].start + 1);
}
}
-#endif /* CONFIG_PPC64 */
+
if (chip->pdev)
pci_dev_put(chip->pdev);
kfree(chip);
@@ -1005,6 +1001,11 @@ static int __init snd_pmac_detect(pmac_t
chip->can_byte_swap = 0; /* FIXME: check this */
chip->control_mask = MASK_IEPC | 0x11;/* disable IEE */
break;
+ default:
+ printk(KERN_ERR "snd: Unknown layout ID 0x%x\n",
+ layout_id);
+ return -ENODEV;
+
}
}
prop = (unsigned int *)get_property(sound, "device-id", NULL);
@@ -1186,46 +1187,69 @@ int __init snd_pmac_new(snd_card_t *card
}
np = chip->node;
+ chip->requested = 0;
if (chip->is_k2) {
- if (np->parent->n_addrs < 2 || np->n_intrs < 3) {
+ static char *rnames[] = {
+ "Sound Control", "Sound DMA" };
+ if (np->n_intrs < 3) {
err = -ENODEV;
goto __error;
}
- for (i = 0; i < 2; i++) {
-#ifndef CONFIG_PPC64
- static char *name[2] = { "- Control", "- DMA" };
- if (! request_OF_resource(np->parent, i, name[i])) {
- snd_printk(KERN_ERR "pmac: can't request resource %d!\n", i);
+ for (i = 0; i < 2; i ++) {
+ if (of_address_to_resource(np->parent, i,
+ &chip->rsrc[i])) {
+ printk(KERN_ERR "snd: can't translate rsrc "
+ " %d (%s)\n", i, rnames[i]);
+ err = -ENODEV;
+ goto __error;
+ }
+ if (request_mem_region(chip->rsrc[i].start,
+ chip->rsrc[i].end -
+ chip->rsrc[i].start + 1,
+ rnames[i]) == NULL) {
+ printk(KERN_ERR "snd: can't request rsrc "
+ " %d (%s: 0x%08lx:%08lx)\n",
+ i, rnames[i], chip->rsrc[i].start,
+ chip->rsrc[i].end);
err = -ENODEV;
goto __error;
}
- chip->of_requested |= (1 << i);
-#endif /* CONFIG_PPC64 */
- ctrl_addr = np->parent->addrs[0].address;
- txdma_addr = np->parent->addrs[1].address;
- rxdma_addr = txdma_addr + 0x100;
+ chip->requested |= (1 << i);
}
-
+ ctrl_addr = chip->rsrc[0].start;
+ txdma_addr = chip->rsrc[1].start;
+ rxdma_addr = txdma_addr + 0x100;
} else {
- if (np->n_addrs < 3 || np->n_intrs < 3) {
+ static char *rnames[] = {
+ "Sound Control", "Sound Tx DMA", "Sound Rx DMA" };
+ if (np->n_intrs < 3) {
err = -ENODEV;
goto __error;
}
-
- for (i = 0; i < 3; i++) {
-#ifndef CONFIG_PPC64
- static char *name[3] = { "- Control", "- Tx DMA", "- Rx DMA" };
- if (! request_OF_resource(np, i, name[i])) {
- snd_printk(KERN_ERR "pmac: can't request resource %d!\n", i);
+ for (i = 0; i < 3; i ++) {
+ if (of_address_to_resource(np->parent, i,
+ &chip->rsrc[i])) {
+ printk(KERN_ERR "snd: can't translate rsrc "
+ " %d (%s)\n", i, rnames[i]);
+ err = -ENODEV;
+ goto __error;
+ }
+ if (request_mem_region(chip->rsrc[i].start,
+ chip->rsrc[i].end -
+ chip->rsrc[i].start + 1,
+ rnames[i]) == NULL) {
+ printk(KERN_ERR "snd: can't request rsrc "
+ " %d (%s: 0x%08lx:%08lx)\n",
+ i, rnames[i], chip->rsrc[i].start,
+ chip->rsrc[i].end);
err = -ENODEV;
goto __error;
}
- chip->of_requested |= (1 << i);
-#endif /* CONFIG_PPC64 */
- ctrl_addr = np->addrs[0].address;
- txdma_addr = np->addrs[1].address;
- rxdma_addr = np->addrs[2].address;
+ chip->requested |= (1 << i);
}
+ ctrl_addr = chip->rsrc[0].start;
+ txdma_addr = chip->rsrc[1].start;
+ rxdma_addr = chip->rsrc[2].start;
}
chip->awacs = ioremap(ctrl_addr, 0x1000);
@@ -1277,9 +1301,11 @@ int __init snd_pmac_new(snd_card_t *card
} else if (chip->is_pbook_G3) {
struct device_node* mio;
for (mio = chip->node->parent; mio; mio = mio->parent) {
- if (strcmp(mio->name, "mac-io") == 0
- && mio->n_addrs > 0) {
- chip->macio_base = ioremap(mio->addrs[0].address, 0x40);
+ if (strcmp(mio->name, "mac-io") == 0) {
+ struct resource r;
+ if (of_address_to_resource(mio, 0, &r) == 0)
+ chip->macio_base =
+ ioremap(r.start, 0x40);
break;
}
}
Index: linux-work/sound/ppc/pmac.h
===================================================================
--- linux-work.orig/sound/ppc/pmac.h 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/sound/ppc/pmac.h 2005-12-13 17:51:20.000000000 +1100
@@ -122,7 +122,8 @@ struct snd_pmac {
unsigned int initialized : 1;
unsigned int feature_is_set : 1;
- unsigned int of_requested;
+ unsigned int requested;
+ struct resource rsrc[3];
int num_freqs;
int *freq_table;
Index: linux-work/include/asm-powerpc/keylargo.h
===================================================================
--- linux-work.orig/include/asm-powerpc/keylargo.h 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/include/asm-powerpc/keylargo.h 2005-12-13 17:51:20.000000000 +1100
@@ -232,10 +232,12 @@
#define K2_FCR1_I2S0_RESET 0x00000800
#define K2_FCR1_I2S0_CLK_ENABLE_BIT 0x00001000
#define K2_FCR1_I2S0_ENABLE 0x00002000
-
#define K2_FCR1_PCI1_CLK_ENABLE 0x00004000
#define K2_FCR1_FW_CLK_ENABLE 0x00008000
#define K2_FCR1_FW_RESET_N 0x00010000
+#define K2_FCR1_I2S1_CELL_ENABLE 0x00020000
+#define K2_FCR1_I2S1_CLK_ENABLE_BIT 0x00080000
+#define K2_FCR1_I2S1_ENABLE 0x00100000
#define K2_FCR1_GMAC_CLK_ENABLE 0x00400000
#define K2_FCR1_GMAC_POWER_DOWN 0x00800000
#define K2_FCR1_GMAC_RESET_N 0x01000000
@@ -246,3 +248,9 @@
#define K2_FCR1_UATA_RESET_N 0x40000000
#define K2_FCR1_UATA_CHOOSE_CLK66 0x80000000
+/* Shasta definitions */
+#define SH_FCR1_I2S2_CELL_ENABLE 0x00000010
+#define SH_FCR1_I2S2_CLK_ENABLE_BIT 0x00000040
+#define SH_FCR1_I2S2_ENABLE 0x00000080
+#define SH_FCR3_I2S2_CLK18_ENABLE 0x00008000
+
Index: linux-work/drivers/video/platinumfb.c
===================================================================
--- linux-work.orig/drivers/video/platinumfb.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/drivers/video/platinumfb.c 2005-12-13 17:51:20.000000000 +1100
@@ -69,6 +69,8 @@ struct fb_info_platinum {
unsigned long total_vram;
int clktype;
int dactype;
+
+ struct resource rsrc_fb, rsrc_reg;
};
/*
@@ -97,9 +99,6 @@ static int platinum_var_to_par(struct fb
* Interface used by the world
*/
-int platinumfb_init(void);
-int platinumfb_setup(char*);
-
static struct fb_ops platinumfb_ops = {
.owner = THIS_MODULE,
.fb_check_var = platinumfb_check_var,
@@ -485,7 +484,7 @@ static int platinum_var_to_par(struct fb
/*
* Parse user speficied options (`video=platinumfb:')
*/
-int __init platinumfb_setup(char *options)
+static int __init platinumfb_setup(char *options)
{
char *this_opt;
@@ -526,19 +525,15 @@ int __init platinumfb_setup(char *option
#define invalidate_cache(addr)
#endif
-static int __devinit platinumfb_probe(struct of_device* odev, const struct of_device_id *match)
+static int __devinit platinumfb_probe(struct of_device* odev,
+ const struct of_device_id *match)
{
struct device_node *dp = odev->node;
struct fb_info *info;
struct fb_info_platinum *pinfo;
- unsigned long addr, size;
volatile __u8 *fbuffer;
- int i, bank0, bank1, bank2, bank3, rc;
+ int bank0, bank1, bank2, bank3, rc;
- if (dp->n_addrs != 2) {
- printk(KERN_ERR "expecting 2 address for platinum (got %d)", dp->n_addrs);
- return -ENXIO;
- }
printk(KERN_INFO "platinumfb: Found Apple Platinum video hardware\n");
info = framebuffer_alloc(sizeof(*pinfo), &odev->dev);
@@ -546,27 +541,40 @@ static int __devinit platinumfb_probe(st
return -ENOMEM;
pinfo = info->par;
- /* Map in frame buffer and registers */
- for (i = 0; i < dp->n_addrs; ++i) {
- addr = dp->addrs[i].address;
- size = dp->addrs[i].size;
- /* Let's assume we can request either all or nothing */
- if (!request_mem_region(addr, size, "platinumfb")) {
- framebuffer_release(info);
- return -ENXIO;
- }
- if (size >= 0x400000) {
- /* frame buffer - map only 4MB */
- pinfo->frame_buffer_phys = addr;
- pinfo->frame_buffer = __ioremap(addr, 0x400000, _PAGE_WRITETHRU);
- pinfo->base_frame_buffer = pinfo->frame_buffer;
- } else {
- /* registers */
- pinfo->platinum_regs_phys = addr;
- pinfo->platinum_regs = ioremap(addr, size);
- }
+ if (of_address_to_resource(dp, 0, &pinfo->rsrc_reg) ||
+ of_address_to_resource(dp, 1, &pinfo->rsrc_fb)) {
+ printk(KERN_ERR "platinumfb: Can't get resources\n");
+ framebuffer_release(info);
+ return -ENXIO;
+ }
+ if (!request_mem_region(pinfo->rsrc_reg.start,
+ pinfo->rsrc_reg.start -
+ pinfo->rsrc_reg.end + 1,
+ "platinumfb registers")) {
+ framebuffer_release(info);
+ return -ENXIO;
+ }
+ if (!request_mem_region(pinfo->rsrc_fb.start,
+ pinfo->rsrc_fb.start
+ - pinfo->rsrc_fb.end + 1,
+ "platinumfb framebuffer")) {
+ release_mem_region(pinfo->rsrc_reg.start,
+ pinfo->rsrc_reg.end -
+ pinfo->rsrc_reg.start + 1);
+ framebuffer_release(info);
+ return -ENXIO;
}
+ /* frame buffer - map only 4MB */
+ pinfo->frame_buffer_phys = pinfo->rsrc_fb.start;
+ pinfo->frame_buffer = __ioremap(pinfo->rsrc_fb.start, 0x400000,
+ _PAGE_WRITETHRU);
+ pinfo->base_frame_buffer = pinfo->frame_buffer;
+
+ /* registers */
+ pinfo->platinum_regs_phys = pinfo->rsrc_reg.start;
+ pinfo->platinum_regs = ioremap(pinfo->rsrc_reg.start, 0x1000);
+
pinfo->cmap_regs_phys = 0xf301b000; /* XXX not in prom? */
request_mem_region(pinfo->cmap_regs_phys, 0x1000, "platinumfb cmap");
pinfo->cmap_regs = ioremap(pinfo->cmap_regs_phys, 0x1000);
@@ -628,18 +636,16 @@ static int __devexit platinumfb_remove(s
{
struct fb_info *info = dev_get_drvdata(&odev->dev);
struct fb_info_platinum *pinfo = info->par;
- struct device_node *dp = odev->node;
- unsigned long addr, size;
- int i;
unregister_framebuffer (info);
/* Unmap frame buffer and registers */
- for (i = 0; i < dp->n_addrs; ++i) {
- addr = dp->addrs[i].address;
- size = dp->addrs[i].size;
- release_mem_region(addr, size);
- }
+ release_mem_region(pinfo->rsrc_fb.start,
+ pinfo->rsrc_fb.end -
+ pinfo->rsrc_fb.start + 1);
+ release_mem_region(pinfo->rsrc_reg.start,
+ pinfo->rsrc_reg.end -
+ pinfo->rsrc_reg.start + 1);
iounmap(pinfo->frame_buffer);
iounmap(pinfo->platinum_regs);
release_mem_region(pinfo->cmap_regs_phys, 0x1000);
@@ -666,7 +672,7 @@ static struct of_platform_driver platinu
.remove = platinumfb_remove,
};
-int __init platinumfb_init(void)
+static int __init platinumfb_init(void)
{
#ifndef MODULE
char *option = NULL;
@@ -680,7 +686,7 @@ int __init platinumfb_init(void)
return 0;
}
-void __exit platinumfb_exit(void)
+static void __exit platinumfb_exit(void)
{
of_unregister_driver(&platinum_driver);
}
Index: linux-work/drivers/video/controlfb.c
===================================================================
--- linux-work.orig/drivers/video/controlfb.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/drivers/video/controlfb.c 2005-12-13 17:51:20.000000000 +1100
@@ -133,12 +133,6 @@ static int controlfb_mmap(struct fb_info
static int controlfb_set_par (struct fb_info *info);
static int controlfb_check_var (struct fb_var_screeninfo *var, struct fb_info *info);
-/*
- * inititialization
- */
-int control_init(void);
-void control_setup(char *);
-
/******************** Prototypes for internal functions **********************/
static void set_control_clock(unsigned char *params);
@@ -550,9 +544,46 @@ static void control_set_hardware(struct
/*
- * Called from fbmem.c for probing & initializing
+ * Parse user speficied options (`video=controlfb:')
*/
-int __init control_init(void)
+static void __init control_setup(char *options)
+{
+ char *this_opt;
+
+ if (!options || !*options)
+ return;
+
+ while ((this_opt = strsep(&options, ",")) != NULL) {
+ if (!strncmp(this_opt, "vmode:", 6)) {
+ int vmode = simple_strtoul(this_opt+6, NULL, 0);
+ if (vmode > 0 && vmode <= VMODE_MAX &&
+ control_mac_modes[vmode - 1].m[1] >= 0)
+ default_vmode = vmode;
+ } else if (!strncmp(this_opt, "cmode:", 6)) {
+ int depth = simple_strtoul(this_opt+6, NULL, 0);
+ switch (depth) {
+ case CMODE_8:
+ case CMODE_16:
+ case CMODE_32:
+ default_cmode = depth;
+ break;
+ case 8:
+ default_cmode = CMODE_8;
+ break;
+ case 15:
+ case 16:
+ default_cmode = CMODE_16;
+ break;
+ case 24:
+ case 32:
+ default_cmode = CMODE_32;
+ break;
+ }
+ }
+ }
+}
+
+static int __init control_init(void)
{
struct device_node *dp;
char *option = NULL;
@@ -651,15 +682,16 @@ static void __init find_vram_size(struct
static int __init control_of_init(struct device_node *dp)
{
struct fb_info_control *p;
- unsigned long addr;
- int i;
+ struct resource fb_res, reg_res;
if (control_fb) {
printk(KERN_ERR "controlfb: only one control is supported\n");
return -ENXIO;
}
- if(dp->n_addrs != 2) {
- printk(KERN_ERR "expecting 2 address for control (got %d)", dp->n_addrs);
+
+ if (of_pci_address_to_resource(dp, 2, &fb_res) ||
+ of_pci_address_to_resource(dp, 1, ®_res)) {
+ printk(KERN_ERR "can't get 2 addresses for control\n");
return -ENXIO;
}
p = kmalloc(sizeof(*p), GFP_KERNEL);
@@ -669,18 +701,12 @@ static int __init control_of_init(struct
memset(p, 0, sizeof(*p));
/* Map in frame buffer and registers */
- for (i = 0; i < dp->n_addrs; ++i) {
- addr = dp->addrs[i].address;
- if (dp->addrs[i].size >= 0x800000) {
- p->fb_orig_base = addr;
- p->fb_orig_size = dp->addrs[i].size;
- /* use the big-endian aperture (??) */
- p->frame_buffer_phys = addr + 0x800000;
- } else {
- p->control_regs_phys = addr;
- p->control_regs_size = dp->addrs[i].size;
- }
- }
+ p->fb_orig_base = fb_res.start;
+ p->fb_orig_size = fb_res.end - fb_res.start + 1;
+ /* use the big-endian aperture (??) */
+ p->frame_buffer_phys = fb_res.start + 0x800000;
+ p->control_regs_phys = reg_res.start;
+ p->control_regs_size = reg_res.end - reg_res.start + 1;
if (!p->fb_orig_base ||
!request_mem_region(p->fb_orig_base,p->fb_orig_size,"controlfb")) {
@@ -1059,43 +1085,3 @@ static void control_cleanup(void)
}
-/*
- * Parse user speficied options (`video=controlfb:')
- */
-void __init control_setup(char *options)
-{
- char *this_opt;
-
- if (!options || !*options)
- return;
-
- while ((this_opt = strsep(&options, ",")) != NULL) {
- if (!strncmp(this_opt, "vmode:", 6)) {
- int vmode = simple_strtoul(this_opt+6, NULL, 0);
- if (vmode > 0 && vmode <= VMODE_MAX &&
- control_mac_modes[vmode - 1].m[1] >= 0)
- default_vmode = vmode;
- } else if (!strncmp(this_opt, "cmode:", 6)) {
- int depth = simple_strtoul(this_opt+6, NULL, 0);
- switch (depth) {
- case CMODE_8:
- case CMODE_16:
- case CMODE_32:
- default_cmode = depth;
- break;
- case 8:
- default_cmode = CMODE_8;
- break;
- case 15:
- case 16:
- default_cmode = CMODE_16;
- break;
- case 24:
- case 32:
- default_cmode = CMODE_32;
- break;
- }
- }
- }
-}
-
Index: linux-work/sound/oss/dmasound/dmasound_awacs.c
===================================================================
--- linux-work.orig/sound/oss/dmasound/dmasound_awacs.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/sound/oss/dmasound/dmasound_awacs.c 2005-12-13 17:51:20.000000000 +1100
@@ -125,6 +125,7 @@ static int awacs_rate_index;
static int awacs_subframe;
static struct device_node* awacs_node;
static struct device_node* i2s_node;
+static struct resource awacs_rsrc[3];
static char awacs_name[64];
static int awacs_revision;
@@ -667,9 +668,12 @@ static void PMacIrqCleanup(void)
iounmap(awacs_txdma);
iounmap(awacs_rxdma);
- release_OF_resource(awacs_node, 0);
- release_OF_resource(awacs_node, 1);
- release_OF_resource(awacs_node, 2);
+ release_mem_region(awacs_rsrc[0].start,
+ awacs_rsrc[0].end - awacs_rsrc[0].start + 1);
+ release_mem_region(awacs_rsrc[1].start,
+ awacs_rsrc[1].end - awacs_rsrc[1].start + 1);
+ release_mem_region(awacs_rsrc[2].start,
+ awacs_rsrc[2].end - awacs_rsrc[2].start + 1);
kfree(awacs_tx_cmd_space);
kfree(awacs_rx_cmd_space);
@@ -2863,46 +2867,58 @@ printk("dmasound_pmac: couldn't find a C
* other info if necessary (early AWACS we want to read chip ids)
*/
- if (io->n_addrs < 3 || io->n_intrs < 3) {
+ if (of_get_address(io, 2, NULL, NULL) == NULL || io->n_intrs < 3) {
/* OK - maybe we need to use the 'awacs' node (on earlier
* machines).
- */
+ */
if (awacs_node) {
io = awacs_node ;
- if (io->n_addrs < 3 || io->n_intrs < 3) {
- printk("dmasound_pmac: can't use %s"
- " (%d addrs, %d intrs)\n",
- io->full_name, io->n_addrs, io->n_intrs);
+ if (of_get_address(io, 2, NULL, NULL) == NULL ||
+ io->n_intrs < 3) {
+ printk("dmasound_pmac: can't use %s\n",
+ io->full_name);
return -ENODEV;
}
- } else {
- printk("dmasound_pmac: can't use %s (%d addrs, %d intrs)\n",
- io->full_name, io->n_addrs, io->n_intrs);
- }
+ } else
+ printk("dmasound_pmac: can't use %s\n", io->full_name);
}
- if (!request_OF_resource(io, 0, NULL)) {
+ if (of_address_to_resource(io, 0, &awacs_rsrc[0]) ||
+ request_mem_region(awacs_rsrc[0].start,
+ awacs_rsrc[0].end - awacs_rsrc[0].start + 1,
+ " (IO)") == NULL) {
printk(KERN_ERR "dmasound: can't request IO resource !\n");
return -ENODEV;
}
- if (!request_OF_resource(io, 1, " (tx dma)")) {
- release_OF_resource(io, 0);
- printk(KERN_ERR "dmasound: can't request TX DMA resource !\n");
+ if (of_address_to_resource(io, 1, &awacs_rsrc[1]) ||
+ request_mem_region(awacs_rsrc[1].start,
+ awacs_rsrc[1].end - awacs_rsrc[1].start + 1,
+ " (tx dma)") == NULL) {
+ release_mem_region(awacs_rsrc[0].start,
+ awacs_rsrc[0].end - awacs_rsrc[0].start + 1);
+ printk(KERN_ERR "dmasound: can't request Tx DMA resource !\n");
return -ENODEV;
}
-
- if (!request_OF_resource(io, 2, " (rx dma)")) {
- release_OF_resource(io, 0);
- release_OF_resource(io, 1);
- printk(KERN_ERR "dmasound: can't request RX DMA resource !\n");
+ if (of_address_to_resource(io, 2, &awacs_rsrc[2]) ||
+ request_mem_region(awacs_rsrc[2].start,
+ awacs_rsrc[2].end - awacs_rsrc[2].start + 1,
+ " (rx dma)") == NULL) {
+ release_mem_region(awacs_rsrc[0].start,
+ awacs_rsrc[0].end - awacs_rsrc[0].start + 1);
+ release_mem_region(awacs_rsrc[1].start,
+ awacs_rsrc[1].end - awacs_rsrc[1].start + 1);
+ printk(KERN_ERR "dmasound: can't request Rx DMA resource !\n");
return -ENODEV;
}
awacs_beep_dev = input_allocate_device();
if (!awacs_beep_dev) {
- release_OF_resource(io, 0);
- release_OF_resource(io, 1);
- release_OF_resource(io, 2);
+ release_mem_region(awacs_rsrc[0].start,
+ awacs_rsrc[0].end - awacs_rsrc[0].start + 1);
+ release_mem_region(awacs_rsrc[1].start,
+ awacs_rsrc[1].end - awacs_rsrc[1].start + 1);
+ release_mem_region(awacs_rsrc[2].start,
+ awacs_rsrc[2].end - awacs_rsrc[2].start + 1);
printk(KERN_ERR "dmasound: can't allocate input device !\n");
return -ENOMEM;
}
@@ -2916,11 +2932,11 @@ printk("dmasound_pmac: couldn't find a C
/* all OF versions I've seen use this value */
if (i2s_node)
- i2s = ioremap(io->addrs[0].address, 0x1000);
+ i2s = ioremap(awacs_rsrc[0].start, 0x1000);
else
- awacs = ioremap(io->addrs[0].address, 0x1000);
- awacs_txdma = ioremap(io->addrs[1].address, 0x100);
- awacs_rxdma = ioremap(io->addrs[2].address, 0x100);
+ awacs = ioremap(awacs_rsrc[0].start, 0x1000);
+ awacs_txdma = ioremap(awacs_rsrc[1].start, 0x100);
+ awacs_rxdma = ioremap(awacs_rsrc[2].start, 0x100);
/* first of all make sure that the chip is powered up....*/
pmac_call_feature(PMAC_FTR_SOUND_CHIP_ENABLE, io, 0, 1);
@@ -3083,9 +3099,10 @@ printk("dmasound_pmac: Awacs/Screamer Co
struct device_node* mio;
macio_base = NULL;
for (mio = io->parent; mio; mio = mio->parent) {
- if (strcmp(mio->name, "mac-io") == 0
- && mio->n_addrs > 0) {
- macio_base = ioremap(mio->addrs[0].address, 0x40);
+ if (strcmp(mio->name, "mac-io") == 0) {
+ struct resource r;
+ if (of_address_to_resource(mio, 0, &r) == 0)
+ macio_base = ioremap(r.start, 0x40);
break;
}
}
Index: linux-work/drivers/video/valkyriefb.c
===================================================================
--- linux-work.orig/drivers/video/valkyriefb.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/drivers/video/valkyriefb.c 2005-12-13 17:51:20.000000000 +1100
@@ -342,19 +342,19 @@ int __init valkyriefb_init(void)
#else /* ppc (!CONFIG_MAC) */
{
struct device_node *dp;
+ struct resource r;
- dp = find_devices("valkyrie");
+ dp = of_find_node_by_name(NULL, "valkyrie");
if (dp == 0)
return 0;
- if (dp->n_addrs != 1) {
- printk(KERN_ERR "expecting 1 address for valkyrie (got %d)\n",
- dp->n_addrs);
+ if (of_address_to_resource(dp, 0, &r)) {
+ printk(KERN_ERR "can't find address for valkyrie\n");
return 0;
}
- frame_buffer_phys = dp->addrs[0].address;
- cmap_regs_phys = dp->addrs[0].address+0x304000;
+ frame_buffer_phys = r.start;
+ cmap_regs_phys = r.start + 0x304000;
flags = _PAGE_WRITETHRU;
}
#endif /* ppc (!CONFIG_MAC) */
Index: linux-work/drivers/block/swim3.c
===================================================================
--- linux-work.orig/drivers/block/swim3.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/drivers/block/swim3.c 2005-12-13 17:51:20.000000000 +1100
@@ -1083,23 +1083,33 @@ static int swim3_add_device(struct devic
{
struct device_node *mediabay;
struct floppy_state *fs = &floppy_states[floppy_count];
+ struct resource res_reg, res_dma;
- if (swim->n_addrs < 2)
- {
- printk(KERN_INFO "swim3: expecting 2 addrs (n_addrs:%d, n_intrs:%d)\n",
- swim->n_addrs, swim->n_intrs);
+ if (of_address_to_resource(swim, 0, &res_reg) ||
+ of_address_to_resource(swim, 1, &res_dma)) {
+ printk(KERN_ERR "swim3: Can't get addresses\n");
return -EINVAL;
}
-
- if (swim->n_intrs < 2)
- {
- printk(KERN_INFO "swim3: expecting 2 intrs (n_addrs:%d, n_intrs:%d)\n",
- swim->n_addrs, swim->n_intrs);
+ if (request_mem_region(res_reg.start, res_reg.end - res_reg.start + 1,
+ " (reg)") == NULL) {
+ printk(KERN_ERR "swim3: Can't request register space\n");
+ return -EINVAL;
+ }
+ if (request_mem_region(res_dma.start, res_dma.end - res_dma.start + 1,
+ " (dma)") == NULL) {
+ release_mem_region(res_reg.start,
+ res_reg.end - res_reg.start + 1);
+ printk(KERN_ERR "swim3: Can't request DMA space\n");
return -EINVAL;
}
- if (!request_OF_resource(swim, 0, NULL)) {
- printk(KERN_INFO "swim3: can't request IO resource !\n");
+ if (swim->n_intrs < 2) {
+ printk(KERN_INFO "swim3: expecting 2 intrs (n_intrs:%d)\n",
+ swim->n_intrs);
+ release_mem_region(res_reg.start,
+ res_reg.end - res_reg.start + 1);
+ release_mem_region(res_dma.start,
+ res_dma.end - res_dma.start + 1);
return -EINVAL;
}
@@ -1110,10 +1120,8 @@ static int swim3_add_device(struct devic
memset(fs, 0, sizeof(*fs));
spin_lock_init(&fs->lock);
fs->state = idle;
- fs->swim3 = (struct swim3 __iomem *)
- ioremap(swim->addrs[0].address, 0x200);
- fs->dma = (struct dbdma_regs __iomem *)
- ioremap(swim->addrs[1].address, 0x200);
+ fs->swim3 = (struct swim3 __iomem *)ioremap(res_reg.start, 0x200);
+ fs->dma = (struct dbdma_regs __iomem *)ioremap(res_dma.start, 0x200);
fs->swim3_intr = swim->intrs[0].line;
fs->dma_intr = swim->intrs[1].line;
fs->cur_cyl = -1;
Index: linux-work/arch/powerpc/kernel/prom_init.c
===================================================================
--- linux-work.orig/arch/powerpc/kernel/prom_init.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/arch/powerpc/kernel/prom_init.c 2005-12-13 17:51:20.000000000 +1100
@@ -558,7 +558,8 @@ unsigned long prom_memparse(const char *
static void __init early_cmdline_parse(void)
{
struct prom_t *_prom = &RELOC(prom);
- char *opt, *p;
+ const char *opt;
+ char *p;
int l = 0;
RELOC(prom_cmd_line[0]) = 0;
Index: linux-work/arch/powerpc/kernel/rtas_pci.c
===================================================================
--- linux-work.orig/arch/powerpc/kernel/rtas_pci.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/arch/powerpc/kernel/rtas_pci.c 2005-12-13 17:51:20.000000000 +1100
@@ -188,39 +188,19 @@ int is_python(struct device_node *dev)
return 0;
}
-static int get_phb_reg_prop(struct device_node *dev,
- unsigned int addr_size_words,
- struct reg_property64 *reg)
+static void python_countermeasures(struct device_node *dev)
{
- unsigned int *ui_ptr = NULL, len;
-
- /* Found a PHB, now figure out where his registers are mapped. */
- ui_ptr = (unsigned int *)get_property(dev, "reg", &len);
- if (ui_ptr == NULL)
- return 1;
-
- if (addr_size_words == 1) {
- reg->address = ((struct reg_property32 *)ui_ptr)->address;
- reg->size = ((struct reg_property32 *)ui_ptr)->size;
- } else {
- *reg = *((struct reg_property64 *)ui_ptr);
- }
-
- return 0;
-}
-
-static void python_countermeasures(struct device_node *dev,
- unsigned int addr_size_words)
-{
- struct reg_property64 reg_struct;
+ struct resource registers;
void __iomem *chip_regs;
volatile u32 val;
- if (get_phb_reg_prop(dev, addr_size_words, ®_struct))
+ if (of_address_to_resource(dev, 0, ®isters)) {
+ printk(KERN_ERR "Can't get address for Python workarounds !\n");
return;
+ }
/* Python's register file is 1 MB in size. */
- chip_regs = ioremap(reg_struct.address & ~(0xfffffUL), 0x100000);
+ chip_regs = ioremap(registers.start & ~(0xfffffUL), 0x100000);
/*
* Firmware doesn't always clear this bit which is critical
@@ -301,11 +281,10 @@ static int phb_set_bus_ranges(struct dev
}
static int __devinit setup_phb(struct device_node *dev,
- struct pci_controller *phb,
- unsigned int addr_size_words)
+ struct pci_controller *phb)
{
if (is_python(dev))
- python_countermeasures(dev, addr_size_words);
+ python_countermeasures(dev);
if (phb_set_bus_ranges(dev, phb))
return 1;
@@ -320,8 +299,8 @@ unsigned long __init find_and_init_phbs(
{
struct device_node *node;
struct pci_controller *phb;
- unsigned int root_size_cells = 0;
unsigned int index;
+ unsigned int root_size_cells = 0;
unsigned int *opprop = NULL;
struct device_node *root = of_find_node_by_path("/");
@@ -343,10 +322,11 @@ unsigned long __init find_and_init_phbs(
phb = pcibios_alloc_controller(node);
if (!phb)
continue;
- setup_phb(node, phb, root_size_cells);
+ setup_phb(node, phb);
pci_process_bridge_OF_ranges(phb, node, 0);
pci_setup_phb_io(phb, index == 0);
#ifdef CONFIG_PPC_PSERIES
+ /* XXX This code need serious fixing ... --BenH */
if (ppc64_interrupt_controller == IC_OPEN_PIC && pSeries_mpic) {
int addr = root_size_cells * (index + 2) - 1;
mpic_assign_isu(pSeries_mpic, index, opprop[addr]);
@@ -381,22 +361,17 @@ unsigned long __init find_and_init_phbs(
struct pci_controller * __devinit init_phb_dynamic(struct device_node *dn)
{
- struct device_node *root = of_find_node_by_path("/");
- unsigned int root_size_cells = 0;
struct pci_controller *phb;
int primary;
- root_size_cells = prom_n_size_cells(root);
-
primary = list_empty(&hose_list);
phb = pcibios_alloc_controller(dn);
if (!phb)
return NULL;
- setup_phb(dn, phb, root_size_cells);
+ setup_phb(dn, phb);
pci_process_bridge_OF_ranges(phb, dn, primary);
pci_setup_phb_io_dynamic(phb, primary);
- of_node_put(root);
pci_devs_phb_init_dynamic(phb);
scan_phb(phb);
Index: linux-work/arch/powerpc/mm/numa.c
===================================================================
--- linux-work.orig/arch/powerpc/mm/numa.c 2005-12-13 17:49:22.000000000 +1100
+++ linux-work/arch/powerpc/mm/numa.c 2005-12-13 17:51:20.000000000 +1100
@@ -432,7 +432,8 @@ static int __init parse_numa_properties(
if (!memcell_buf || len <= 0)
continue;
- ranges = memory->n_addrs;
+ /* ranges in cell */
+ ranges = (len >> 2) / (n_mem_addr_cells + n_mem_size_cells);
new_range:
/* these are order-sensitive, and modify the buffer pointer */
start = read_n_cells(n_mem_addr_cells, &memcell_buf);
@@ -746,7 +747,8 @@ int hot_add_scn_to_nid(unsigned long scn
if (!memcell_buf || len <= 0)
continue;
- ranges = memory->n_addrs; /* ranges in cell */
+ /* ranges in cell */
+ ranges = (len >> 2) / (n_mem_addr_cells + n_mem_size_cells);
ha_new_range:
start = read_n_cells(n_mem_addr_cells, &memcell_buf);
size = read_n_cells(n_mem_size_cells, &memcell_buf);
^ permalink raw reply
* [PATCH] powerpc: Fix platinumfb for some modes
From: Benjamin Herrenschmidt @ 2005-12-13 6:48 UTC (permalink / raw)
To: Paul Mackerras; +Cc: linuxppc64-dev, linuxppc-dev list
The platinumfb driver used only on some powermacs has an issue with some
video modes & limited VRAM. This patch fixes it.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Index: linux-work/drivers/video/platinumfb.h
===================================================================
--- linux-work.orig/drivers/video/platinumfb.h 2005-09-22 14:07:19.000000000 +1000
+++ linux-work/drivers/video/platinumfb.h 2005-11-25 17:17:17.000000000 +1100
@@ -158,7 +158,9 @@
/* 832x624, 75Hz (13) */
static struct platinum_regvals platinum_reg_init_13 = {
0x70,
- { 864, 1680, 3360 }, /* MacOS does 1680 instead of 1696 to fit 16bpp in 1MB */
+ { 864, 1680, 3344 }, /* MacOS does 1680 instead of 1696 to fit 16bpp in 1MB,
+ * and we use 3344 instead of 3360 to fit in 2Mb
+ */
{ 0xff0, 4, 0, 0, 0, 0, 0x299, 0,
0, 0x21e, 0x120, 0x10, 0x23f, 0x1f, 0x25, 0x37,
0x8a, 0x22a, 0x23e, 0x536, 0x534, 4, 9, 0x52,
Index: linux-work/drivers/video/platinumfb.c
===================================================================
--- linux-work.orig/drivers/video/platinumfb.c 2005-11-10 08:20:23.000000000 +1100
+++ linux-work/drivers/video/platinumfb.c 2005-11-25 17:17:17.000000000 +1100
@@ -138,13 +138,15 @@
init = platinum_reg_init[pinfo->vmode-1];
- if (pinfo->vmode == 13 && pinfo->cmode > 0)
- offset = 0x10;
+ if ((pinfo->vmode == VMODE_832_624_75) && (pinfo->cmode > CMODE_8))
+ offset = 0x10;
+
info->screen_base = pinfo->frame_buffer + init->fb_offset + offset;
info->fix.smem_start = (pinfo->frame_buffer_phys) + init->fb_offset + offset;
info->fix.visual = (pinfo->cmode == CMODE_8) ?
FB_VISUAL_PSEUDOCOLOR : FB_VISUAL_DIRECTCOLOR;
- info->fix.line_length = vmode_attrs[pinfo->vmode-1].hres * (1<<pinfo->cmode) + offset;
+ info->fix.line_length = vmode_attrs[pinfo->vmode-1].hres * (1<<pinfo->cmode)
+ + offset;
printk("line_length: %x\n", info->fix.line_length);
return 0;
}
@@ -221,7 +223,9 @@
static inline int platinum_vram_reqd(int video_mode, int color_mode)
{
return vmode_attrs[video_mode-1].vres *
- (vmode_attrs[video_mode-1].hres * (1<<color_mode) + 0x20) +0x1000;
+ (vmode_attrs[video_mode-1].hres * (1<<color_mode) +
+ ((video_mode == VMODE_832_624_75) &&
+ (color_mode > CMODE_8)) ? 0x10 : 0x20) + 0x1000;
}
#define STORE_D2(a, d) { \
^ 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