bandit Gangwere
|
system page, ppc, setup correctly
|
bandit Gangwere
07/31/2008 8:32 PM
post11302
|
system page, ppc, setup correctly
One of my frustrations is the all of the system page docs tell me all the pieces of how to setup the system page (ie use
as_add()), but nothing about the philosophy of the what/why.
I had the choice of creating a set of postings for my system page questions. I finally decided to put them altogether.
I hope this is not too big of a posting and pain in the xxx.
I get to the point of crashing in init_cpu() on line 77.
(I look forward to your answers. It may be a while before I can acknowledge them.)
Thanks ... bandit
mmu 1:1 for low ram and all devices
ppc 7447A - single cpu
MV64460
on-board ram: 00000000-0fffffff (ie 256 Meg)
MV64460: 60000000 (256MB)
PCI0: 70000000 (256MB)
PCI1: b0000000 (256MB)
flash: FF000000-FFFFFFFF
dump_bat(0): i up(00001ffe) lo(00000012) d up(00001ffe) lo(00000052)
dump_bat(1): i up(00001ffe) lo(00000012) d up(00001ffe) lo(00000052)
dump_bat(2): i up(00001ffe) lo(00000012) d up(00001ffe) lo(00000052)
dump_bat(3): i up(00001ffe) lo(00000012) d up(00001ffe) lo(00000052)
dump_bat(4): i up(70001fff) lo(70000022) d up(70001fff) lo(7000002a)
dump_bat(5): i up(b0001fff) lo(b0000022) d up(b0001fff) lo(b000002a)
dump_bat(6): i up(60001fff) lo(60000022) d up(60001fff) lo(60000022)
dump_bat(7): i up(ff001fff) lo(ff000022) d up(ff001fff) lo(ff000022)
---------------------------------------------------
Header size=0x0000009c, Total Size=0x00000580, #Cpu=1, Type=1
Section:system_private offset:0x000001c0 size:0x00000068
syspage ptr user:021f2000 kernel:021f2000
cpupage ptr user:021f3000 kernel:021f3000 spacing:4096
kdebug info:00000000 callback:00000000
boot pgms: idx=0
0) base paddr:02014000 start addr:0203d240
ramsize:00000000 pagesize:00001000
[not sure why the kdebug and callback == 00000000
or if this really is a problem]
Section:qtime offset:0x00000158 size:0x00000048
boot:00000000 CPS:0000000001fc9f08 rate/scale:3000030000/-17 intr:2147483648
[I think this is correct. slime out is: (hex)[dec]
set ppc600 clock to 133332000 (07f27c20)
ppc600_set_clock_freqs(ENTER) clock(07f27c20)[133332000]
ppc600_set_clock_freqs(EXIT) clock(2faee8c0)[799992000]
...
main() call ppc_dec_init_qtime()
]
Section:callout offset:0x000000a0 size:0x00000048
reboot:021f28d4 power:00000000
timer_load:021f2978 reload:021f2994 value:021f29e0
0) display:021f29f0 poll:021f29fc break:021f2a04
1) display:00000000 poll:00000000 break:00000000
[the reboot appears to be a reboot of the mv64460, not a reboot of the 7447A
however, it seems obvious to me that I do not have enough callouts.
BTW - the display char function hw_sermpsc_mv64360.c/put_mpsc_mv64360()
causes a trap when called. I replaced it with my own function
and it worked. This is not the callout putc() function, see next+1 sentence.
I also cut out of the init function because the serial ports were already setup.
The other issue seems to be a bug in the display_char_mpsc_mv64360() patch_debug_tx() where it was not replacing the
fake address with the real hardware address and commenting out the patch function.
However, this is going to be an issue with the other callout functions
]
Section:cpuinfo offset:0x000001a0 size:0x00000020
0) cpu:80030101 flags:c0000026 speed:0000031f cache i/d:1/0 name:43
[The Data cache is enabled. Not sure why (yet) the data cache flag == 0.
I have not looked at the other bits in the other fields.]
Section:cacheattr offset:0x00000540 size:0x00000040
0) flags:22 size:0020 #lines:0400 control:021f2588 next:255
1) flags:01 size:0020 #lines:0400 control:021f25d8 next:255
[I do not know if this is right or not]
Section:meminfo offset:0x00000580 size:0x00000000
Section:asinfo offset:0x00000340 size:0x00000140
0000) 0000000000000000-00000000ffffffff o:ffff a:0010 p:100 c:00000000 n:21
0020) 0000000060000000-000000006fffffff o:0000 a:0003 p:100 c:00000000 n:28
...
View Full Message
|
|
|