Changes to gt64260 interrupt support code

Changes to gt64260 interrupt support code

[mod+linuxppc-dev]

[-- Attachment #1: Type: text/plain, Size: 665 bytes --]



Please introduce these changes into the linuxppc
sources in order that the existing gt64260 interrupt
support code be usable on platforms other than the
Galileo Eval Board.

The current gt64260 interrupt support code forces
use of particular Linux irqs (starting from 0) for
particular GT64260 interrupt sources.  These changes
allow use of that support code in other platforms and
(of particular interest in my situation) in combination
with other interrupt controllers (like the elegant
and flexible 8259) that also want to map *their*
interrupt sources to the Linux irqs starting at 0...


Regards,
          Michael O'Donnell
        Mission Critical Linux



[-- Attachment #2: gt64260irqPatch --]
[-- Type: text/plain , Size: 3906 bytes --]

begin 664 gt64260irqPatch
M/3T]/3T@:6YC;'5D92]A<VTM<'!C+V=T-C0R-C`N:"`Q+C8@=G,@961I=&5D
M(#T]/3T]"BTM+2`Q+C8O:6YC;'5D92]A<VTM<'!C+V=T-C0R-C`N:`E-;VX@
M1&5C(#$W(#$W.C,U.C$W(#(P,#$**RLK(&5D:71E9"]I;F-L=61E+V%S;2UP
M<&,O9W0V-#(V,"YH"51H=2!$96,@,C<@,3<Z-#`Z,3(@,C`P,0I`0"`M,S`L
M-B`K,S`L-R!`0`H@"B`*(&5X=&5R;B!U,S(@("`@(&=T-C0R-C!?8F%S93L*
M*V5X=&5R;B!U,S(@("`@(&=T-C0R-C!?9FER<W1)4E$[("`@("`O*B!792!H
M86YD;&4@=&AE(&YE>'0@.38@25)1<R!F<F]M(&AE<F4@*B\*(&5X=&5R;B!U
M,S(@("`@(&=T-C0R-C!?<F5V:7-I;VX["B!E>'1E<FX@=3@@("`@("!G=#8T
M,C8P7W!C:5]E>&-L=61E7V)R:61G93L*(`H]/3T]/2!A<F-H+W!P8R]K97)N
M96PO9W0V-#(V,%]P:6,N8R`Q+C4@=G,@961I=&5D(#T]/3T]"BTM+2`Q+C4O
M87)C:"]P<&,O:V5R;F5L+V=T-C0R-C!?<&EC+F,)5&AU($1E8R`R,"`Q.#HP
M,CHP-"`R,#`Q"BLK*R!E9&ET960O87)C:"]P<&,O:V5R;F5L+V=T-C0R-C!?
M<&EC+F,)5&AU($1E8R`R-R`Q.#HP,CHS-"`R,#`Q"D!`("TV-RPV("LV-RPX
M($!`"B`)3E5,3`D)"0DO*B!S971?869F:6YI='D@*B\*('T["B`**W4S,B!G
M=#8T,C8P7V9I<G-T25)1(#T@,#L@("`@("`O*B!'5#8T,C8P(&AA;F1L97,@
M=&AE(&YE>'0@.38@25)1<R!F<F]M(&AE<F4@*B\**PH@+RH@9W0V-#(V,%]I
M;FET7VER<2@I"B`@*@H@("H@(%1H:7,@9G5N8W1I;VX@:6YI=&EA;&EZ97,@
M=&AE(&EN=&5R<G5P="!C;VYT<F]L;&5R+B!)="!A<W-I9VYS"D!`("TQ,#,L
M,3$@*S$P-2PQ,2!`0`H@"4=4-C0R-C!?4D5'7U=2251%*$=4-C0R-C!?24-?
M0U!57TE.5%)?34%32U](22P@<'!C7V-A8VAE9%]I<G%?;6%S:ULQ72D["B`*
M(`DO*B!U<V4@=&AE(&=T-C0R-C`@9F]R(&%L;"`H<&]S<VEB;&4I(&EN=&5R
M<G5P="!S;W5R8V5S("HO"BT)9F]R("AI/3`[(&D\.38[(&DK*RD**PEF;W(H
M(&D@/2!G=#8T,C8P7V9I<G-T25)1.R`@:2`\("AG=#8T,C8P7V9I<G-T25)1
M("L@.38I.R`@:2LK("D@('L*(`D):7)Q7V1E<V-;:5TN:&%N9&QE<B`]("9G
M=#8T,C8P7W!I8SL**PE]"B`*(`EI9B`H('!P8U]M9"YP<F]G<F5S<R`I('!P
M8U]M9"YP<F]G<F5S<R@B9W0V-#(V,%]I;FET7VER<3H@97AI="(L(#!X,"D[
M"BT)<F5T=7)N.PH@?0H@"B`*0$`@+3$U,RPQ-"`K,34U+#$W($!`"B`)"0D)
M"6ER<2`](&ER<5]G<'`@*R`V-#L*(`D)"0D)1U0V-#(V,%]214=?5U))5$4H
M1U0V-#(V,%]'4%!?24Y44E]#055312P@?B@Q/#PH:7)Q+38T*2DI.PH@"0D)
M"7T*+0D)"7T*+0D)"65L<V4@>PHK"0D)?2!E;'-E('L*(`D)"0EI<G$@*ST@
M,S(["B`)"0E]"B`)"7T*(`E]"B`*+0ER971U<FX@:7)Q.PHK"6EF*"!I<G$@
M/"`P("D@('L**PD)<F5T=7)N*"!I<G$@*3L**PE](&5L<V4@('L**PD)<F5T
M=7)N*"!G=#8T,C8P7V9I<G-T25)1("L@:7)Q("D["BL)?0H@?0H@"B`O*B!G
M=#8T,C8P7W5N;6%S:U]I<G$H*0I`0"`M,3@P+#8@*S$X-2PW($!`"B!S=&%T
M:6,@=F]I9`H@9W0V-#(V,%]U;FUA<VM?:7)Q*'5N<VEG;F5D(&EN="!I<G$I
M"B!["BL):7)Q("T](&=T-C0R-C!?9FER<W1)4E$["B`):68@*&ER<2`^(#,Q
M*2!["B`)"6EF("AI<G$@/B`V,RD@>PH@"0D)+RH@=6YM87-K($=04"!I<G$@
M*B\*0$`@+3(Q-2PV("LR,C$L-R!`0`H@<W1A=&EC('9O:60*(&=T-C0R-C!?
M;6%S:U]I<G$H=6YS:6=N960@:6YT(&ER<2D*('L**PEI<G$@+3T@9W0V-#(V
M,%]F:7)S=$E243L*(`EI9B`H:7)Q(#X@,S$I('L*(`D):68@*&ER<2`^(#8S
M*2!["B`)"0DO*B!M87-K($=04"!I<G$@*B\*0$`@+3(S-"PV("LR-#$L-2!`
M0`H@"6EF("AI<G$@/3T@,S8I('L@+RH@4V5E;7,@;F5C97-S87)Y(&9O<B!3
M1$U!(&EN=&5R<G5P=',@*B\*(`D)=61E;&%Y*#$I.PH@"7T*+0HM"7)E='5R
M;CL*('T**PH]/3T]/2!D<FEV97)S+VYE="]G=#8T,C8P7V5T:"YC(#$N-B!V
M<R!E9&ET960@/3T]/3T*+2TM(#$N-B]D<FEV97)S+VYE="]G=#8T,C8P7V5T
M:"YC"51U92!.;W8@,C`@,3,Z,C4Z-#8@,C`P,0HK*RL@961I=&5D+V1R:79E
M<G,O;F5T+V=T-C0R-C!?971H+F,)5&AU($1E8R`R-R`Q-SHS.3HQ,B`R,#`Q
M"D!`("TX-S(L,3(@*S@Y."PQ,B!`0`H@("`@("`@("`@159"-C0R-C!?151(
M7VER<5MP;W)T72LK("`@("`I"B`@("`@>PH@(V5N9&EF"BT@("`@("`@(')E
M='9A;"`](')E<75E<W1?:7)Q*"!'5#8T,C8P7T542%]I<G%;<&]R=%TL"BL@
M("`@("`@(')E='9A;"`](')E<75E<W1?:7)Q*"!'5#8T,C8P7T542%]I<G%;
M<&]R=%T@*R!G=#8T,C8P7V9I<G-T25)1+`H@("`@("`@("`@("`@("`@("`@
M("`@("`@("`@("`@9W0V-#(V,%]E=&A?:6YT7VAA;F1L97(L"B`@("`@("`@
M("`@("`@("`@("`@("`@("`@("`@("A305])3E1%4E)54%0@?"!305]304U0
M3$5?4D%.1$]-*2P*("`@("`@("`@("`@("`@("`@("`@("`@("`@("`@(D=4
M-C0R-C!?171H(BP@9&5V("D["B`@("`@("`@(&EF*"`A<F5T=F%L("D@('L*
M+2`@("`@("`@("`@(&1E=BT^:7)Q(#T@1U0V-#(V,%]%5$A?:7)Q6W!O<G1=
M.PHK("`@("`@("`@("`@9&5V+3YI<G$@/2!'5#8T,C8P7T542%]I<G%;<&]R
M=%T@*R!G=#8T,C8P7V9I<G-T25)1.PH@("`@("`@("`@("`@<')I;G1K*"`B
M9W0V-#(V,%]E=&A?;W!E;B`Z($%S<VEG;F5D($E242`E9"!T;R!G=#8T,C8P
M7V5T:"5D7&XB+`H@("`@("`@("`@("`@("`@("`@("`@("`@("`@("`@("`@
M("`@("`@("`@("`@("`@("`@("`@("`@("`@("!D978M/FER<2P@<&]R="`I
4.PH@(VEF(#`@+RH@6%A86"`J+PH`
`
end

Re: Changes to gt64260 interrupt support code

[Troy Benjegerdes]

On Thu, Dec 27, 2001 at 12:44:24PM -0500, mod+linuxppc-dev@MissionCriticalLinux.com wrote:
>
>
> Please introduce these changes into the linuxppc
> sources in order that the existing gt64260 interrupt
> support code be usable on platforms other than the
> Galileo Eval Board.
>
> The current gt64260 interrupt support code forces
> use of particular Linux irqs (starting from 0) for
> particular GT64260 interrupt sources.  These changes
> allow use of that support code in other platforms and
> (of particular interest in my situation) in combination
> with other interrupt controllers (like the elegant
> and flexible 8259) that also want to map *their*
> interrupt sources to the Linux irqs starting at 0...

Ugh. Did I mention I hate 8259's?

We **REALLY** need a better way to deal with cascaded 8259's than adding
an offset.. I think Dan Malek had some good ideas on what to do about
this, but I don't know that anyone has got any currently running code that
solves this problem nicely.



--
Troy Benjegerdes | master of mispeeling | 'da hozer' |  hozer@drgw.net
-----"If this message isn't misspelled, I didn't write it" -- Me -----
"Why do musicians compose symphonies and poets write poems? They do it
because life wouldn't have any meaning for them if they didn't. That's
why I draw cartoons. It's my life." -- Charles Schulz

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Changes to gt64260 interrupt support code

[hollis]

On Wed, Jan 02, 2002 at 12:30:46PM -0600, Troy Benjegerdes wrote:
>
> We **REALLY** need a better way to deal with cascaded 8259's than adding
> an offset.. I think Dan Malek had some good ideas on what to do about
> this, but I don't know that anyone has got any currently running code that
> solves this problem nicely.

The ppc64 port has an array of virtual irq's which is used to index to the
real irq. In the interests of easy debugging (eg when drivers print out their
irq number) the virtual irq is kept identical to the actual whenever possible.

arch/ppc64/prom.c sets up the map (virt_irq_to_real_map), and
arch/ppc64/xics.c (the interrupt controller) uses it.

-Hollis

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Changes to gt64260 interrupt support code

[Mark A. Greer]

mod+linuxppc-dev@MissionCriticalLinux.com wrote:

> Please introduce these changes into the linuxppc
> sources in order that the existing gt64260 interrupt
> support code be usable on platforms other than the
> Galileo Eval Board.

Michael,

I'm pushing your patch as I write this.  I made one change, though.
I changed the variable name gt64260_firstIRQ to gt64260_irq_base since it
seems more descriptive--firstIRQ makes me think of a special case for the
very first intr to happen.

You will need to make the appropriate changes to your internal code.

Mark


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Changes to gt64260 interrupt support code

[Troy Benjegerdes]

On Wed, Jan 02, 2002 at 03:26:23PM -0700, Mark A. Greer wrote:
>
> mod+linuxppc-dev@MissionCriticalLinux.com wrote:
>
> > Please introduce these changes into the linuxppc
> > sources in order that the existing gt64260 interrupt
> > support code be usable on platforms other than the
> > Galileo Eval Board.
>
> Michael,
>
> I'm pushing your patch as I write this.  I made one change, though.
> I changed the variable name gt64260_firstIRQ to gt64260_irq_base since it
> seems more descriptive--firstIRQ makes me think of a special case for the
> very first intr to happen.

err, if you're going to commit this, let's be consistent with openpic
and use 'gt64260_pic_irq_offset'....

(arch/ppc/kernel/open_pic_irq_offset)

Also see openpic_init() .... the offset is passed in


--
Troy Benjegerdes | master of mispeeling | 'da hozer' |  hozer@drgw.net
-----"If this message isn't misspelled, I didn't write it" -- Me -----
"Why do musicians compose symphonies and poets write poems? They do it
because life wouldn't have any meaning for them if they didn't. That's
why I draw cartoons. It's my life." -- Charles Schulz

** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Changes to gt64260 interrupt support code

[Michael O'Donnell]

>> The current gt64260 interrupt support code forces
>> use of particular Linux irqs (starting from 0) for
>> particular GT64260 interrupt sources.  These changes
>> allow use of that support code in other platforms and
>> (of particular interest in my situation) in combination
>> with other interrupt controllers (like the elegant
>> and flexible 8259) that also want to map *their*
>> interrupt sources to the Linux irqs starting at 0...
>
>
>Ugh. Did I mention I hate 8259's?
>
>We **REALLY** need a better way to deal with cascaded
>8259's than adding an offset..  I think Dan Malek
>had some good ideas on what to do about this, but I
>don't know that anyone has got any currently running
>code that solves this problem nicely.


We're all agreed: the current interrupt subsystem is
too brittle and needs to be reworked.  Meanwhile, can we
get that patch applied?  I'm in the process of porting
to a new platform and I'm trying to be a good boy and
feed y'all my platform-independent stuff in bite-sized
pieces prior to the larger platform-specific bolus...

Regards,
          Michael O'Donnell
        Mission Critical Linux


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/

Re: Changes to gt64260 interrupt support code

[mod+linuxppc-dev]

>The ppc64 port has an array of virtual irq's which is used to index to
>the real irq.  In the interests of easy debugging (eg when drivers print
>out their irq number) the virtual irq is kept identical to the actual
>whenever possible.
>
>arch/ppc64/prom.c sets up the map (virt_irq_to_real_map), and
>arch/ppc64/xics.c (the interrupt controller) uses it.


Simple translation of irqs doesn't really solve the
problem; a general solution would provide us with the
ability to initialize and service interrupt requests
routed through multiple, nested, heterogeneous
controllers no matter what the HW designers were
smoking when they laid out the system.  See earlier
posts to this list for good discussion on this topic.


** Sent via the linuxppc-dev mail list. See http://lists.linuxppc.org/