Hijack v386: extra memory support for mk1/mk2/mk2a

Hijack v386 is now available.

This version includes automatic detection/use of upgraded memory on Mk1, Mk2, and Mk2a players. This still requires Hugo's upgraded 0e000 flash loader (to enable the DRAM before we execute from it), but is automatic from that point on.

The follow memory configurations are automatically detected/supported:

• Mk1 players with 8MB (factory) or 16MB memory (8MB extra).
• Mk2 players with 12MB (factory) or 16MB memory (4MB extra).
• Mk2a players with 16MB (factory) or 32MB memory (16MB extra).

I hope to also remove the need for the updated 0e000 flash loader at some point, but for now it is still a prerequisite to use any extra installed memory. Note that 0e000 gets overwritten whenever installing a software .upgrade, unless using version 0.9 or higher of my ethernet upgrader utility (available from the Hijack site).

Cheers

On my upgraded Mk2a player (32MB total), the player software seems to make use of at least a few MB of the extra memory for holding the database. But it does not use most of it yet -- need a newer alpha release (from peter or rob) for that to happen.

Cheers

Okay, Hijack v387 is now available.

This version removes any requirement for a special 0e000 flash loader when extra memory is installed in the player. The stock loader (which gets restored like it or not on any software upgrade) works just fine with my 32MB player now.

Cheers

EXCELLENT!!! Thanks!!! I can't wait to try it out!

For some reason, I cannot help but think this release number is appropriate for add-on chips to expand the capability of the existing silicon... (historical use only)

I seem to have missed the memory upgrade thread. Any ideas who's doing these? I'd like at least 1 of my 2a's upgraded.

Mmm.. yes, I missed that !

Very observant, Paul!

I cannot help but think this release number is appropriate for add-on chips to expand the capability of the existing silicon

True! Now, that's funny.

At this point I only have to catch up with the mem expansion thread...

Well, so far nobody is doing them.

I appear to have the only confirmed upgraded Mk2a player in existance, but I don't have a supply of DRAM chips for other people's players (these cost about $40 per Mk2a player, I think).

And at the moment I'm unsure if I have the time to do the upgrades. The actual chip install took only about an hour on the one player I've worked on so far -- suprisingly easy to do after all of that procrastination! Much easier, for instance, than assembling the knob LED boards. But also much riskier.

EDIT: Note that I do have a temperature controlled soldering station (required for this mod), and a 0.01" tip on the iron (which helps greatly).

Stu has a queue of players to upgrade, but has yet to verify success (expected soon, though).

Cheers

Oh, and a side-benefit: while doing this mod, I finally learned how/when to use copper braid to remove solder bridges! Life-skills!!

Hijack v388 is now available.

This version features a much more exhaustive memory test for expanded DRAM. It cycles several patterns through each location over the first 1MB of any installed "extra" DRAM.

This doesn't noticeably slow down player startup, but will very much aid in diagnosis of issues when testing a newly installed memory upgrade.

On upgraded players where all tests pass, no special messages are output -- you'll just see the player reporting more total system memory than before the upgrade.

On players without the upgrade, and players with a botched upgrade, one may notice a one-liner like this (below) near the top of the serial output at boot time:

Checking for extra DRAM at c1000000: wrote 00000000, read c00e78e0

A message like this (above) indicates either no extra DRAM found or test failed on the extra DRAM. In either case, the player startup continues WITHOUT the extra DRAM being used. The wrote and read values can be used to troubleshoot the memory upgrade if necessary (I suggest first simply resoldering the data lines which differ).

Cheers

Hello Mark,

The V386 - V388 it works on my MK2a as expected. Normal booting Empeg. But on the MK2 I can't boot the device any more until downgrading to V385.

The last lines looks like that!

ide0 at 0x000-0x007,0x038 on irq 6
hda: IC25N080ATMR04-0, 76319MB w/7884kB Cache, CHS=9729/255/63
empeg-flash driver initialized
smc chip id/revision 0x3349
smc9194.c:v0.12 03/06/96 by Erik Stahlman ([email protected])

SMC9194: SMC91C94(r:9) at 0x4008000 IRQ:7 INTF:TP MEM:6144b MAC 00:02:d7:10:02:2
e
Partition check:
hda: hda1 < hda5 hda6 > hda2 hda3 hda4
RAMDISK: ext2 filesystem found at block 0
RAMDISK: Loading 320 blocks [1 disk] into ram disk... <0>Kernel panic: VFS: Free
block list corrupted

bye

Oh good, I thought perhaps the entire BBS was asleep at the wheel this week.

I'm happy to know that *somebody* tries out this stuffl

I mean, come on! You folks know the deal: I write it, you test it!

Thomas, I'm betting that higher up in the log there's more info somewhere. Can I see the whole log, please.

Thanks

You meen crazymelki, instead of Thomas?

ok..Maybe I will get a free upgrade as the first power tester ;-)

Strange I got right now two different errors..Here the first one. In the meantime I am trying to reproduce the other one.


empeg-car bootstrap v1.00 20000601 ([email protected]) CHS=19485/16/63
On DC power, quick boot
e
Partition check:

e000 v1.04
empeg-fl
Copying kernel...ized3] hda1
Calling linux kernel...
smc chip id/rev
Uncompressing Linux..................................... done, booting the kerne
smc9194.c:v0.12 03/06/96 by Erik Stahlman ([email protected])0 blocks [1 disk] into
l.m
Linux version 2.2.17-rmk5-np17-empeg52-hijack-v388 (root@ibbm) (gcc version 2.95

SMC9194: SMC91C94(r:9) at 0x4008000 IRQ:7 INTF:TP M
.3 20010315 (release)) #2 Thu May 6 12:52:42 EDT 2004ded
Processor: Intel StrongARM-1100 revision
e

Dentry hash table entries: 2048 (order 2, 16k)rfs but got error 19
Buffer cache hash table entries: 16384 (order 4, 64k)on irq 6
Mounting s
Page cache hash table entries: 4096 (order 2, 16k)
hda: TOSHIBA MK1016GAP, 9590MB
POSIX conformance testing by UNIFIX4 but got error 6
Linux NET4.0 for Linux 2.2r mounting partitions (pos
pla
TCP: Hash tables configured (ehash 16384 bhash 16384)eg-car 3.00-alpha7 2004/02/13. Loading 320 blocks [1
IrDA (tm) Protocols for Linux-2.2 (Dag Brattli)
Starting kswapd v 1.5
SA1100 serial driver version 4.27 with no serial options enabled
hdstress.cpp 189 ( 14):
ttyS00 at 0xf8010000 (irq = 15) is a SA1100 UARTg-car bootstrap v1.00 20000601 ([email protected])c
ttyS01 at 0xf8050000 (irq = 17) is a SA1100 UART
If there is anyone present w
ttyS02 at 0xf8030000 (irq = 16) is a SA1100 UART,
Signature is 20706d65 'emp '

VFS: Mounted root (ext2
Found custom animation at offset 0
empeg dsp initialised
empeg audio-in initialised, CS4231A revision a0rved for the super user
empeg remote control/panel button initialised.-1100 revision 11s.cpp 189 ( 14): Poll re
empeg usb initialised, PDIUSBD12 id 1012
NetWinder
empeg state support initialised 0089/88c1 (save to d0004d80).
empeg RDS driver initialisedg=0)

empeg power-pic driver initialised (first boot)serial number 80000558)upsdev/hda: Spinning
RAM disk driver initialized: 16 RAM disks of 4096K size
8192 bl
Command line: mem=1
empeg single channel IDE
smc chip id/revision 0x3349
TCP: Hash ta
smc9194.c:v0.12 03/06/96 by Erik Stahlman ([email protected])
hdstr
hdstr

SMC9194: SMC91C94(r:9) at 0x4008000 IRQ:7 INTF:TP MEM:6144b MAC 00:02:d7:10:02:2
hdst
Drive
ttyS02 at 0xf8030000 (irq =
e6
Partition check:T
hda: hda1 < hda5 hda6 > hda2 hda3 hda4ss.cp
Signature is 20706d65
VFS: Mounted root (ext2 filesystem) readonly.
empeg display initialised.
Freeing unused kernel memory: 4k init
empeg dsp audio
Probing primary interface...0
r10: 00000005 r9 : 00000000 r8 : 00330003A DISK drive9): Poll result=0 on fd=5
r7 : c0ffdfc0 r6 : c0159740 r5 : c0ffd040 r4 : c0ffdd3cda: TOSHIBA MK1016GAP, ATA DISK driveeed checking /dev/hda
r3 : 0000007c r2 : 00000000 r1 : c0157fa0 r0 : 00000009
hda: TOSHIBA MK1016GAP,
Flags: nZCv IRQs on FIQs on Mode SVC_32 Segment kerneless.cpp 189 ( 29): Po
St
Control: C000517D Table: C000517D DAC: 0000001D
Stressi

a8t.
c0003b80: c0003b8c c002ab20 c002a358 00000000 c0103fb8 00000000 00000400 c0003besent who wants to upgr

SMC9194: SM
c4C9
c0003ba0: c0003bac c0033ed4 c002a900 c015bf60 00000000 00000000 c0003bfc c0003b
hdstress.cpp 1
c8(
c0003bc0: c0033f9c c0033e68 c0ffe000 c0002000 00000000 00000000 c0ffe000 000004
hda: [PTBL] [1222/255/63] hda1 < hda5 hda6 > hda2 hda3 hda4 Penguins of Linux!
00
c0003be0: 00000305 c0103f78 c0103f18 00002105

Mounting second music partition8)ress.cpp 189
ec
c0003c80: 00000008 00000000 c0003cd0 c0003c98 c004adcc c004a874 c0003cec c00155ount /dev/hdc4 but got error 6
Error mou
ccng
c0003ca0: 00000000 00000010 c0003d10 00000000 c01879c8 00000000 c01879c8 c01007
hdstress.c
Linux N
g player Lin
982
08
c0003d00: c0185060 c01852e0 c0003d2c c0003d18 00000000 00000000 00000000 000000pinning up.p.t=0 on fd=
00
c0003d20: 00000000 00000000 00000000 00000000 00000000 c0185360 c01879c8 c01879lash, let them speak now,TCP
c8
c0003d40: c0003da4 c0100798 00000001 c0003d78 c0003d5c c004c12c c004bdd0 000000
or forever hold their peace...it seems not.
00t
c0003d60: c0185360 fffffff4 c01852e0 c0003da0 c0003d7c c003ac24 c004c0fc c0003d


a4
00
c0003de0: c0003e8c c0100770 00000001 c0100798 00000000 c0003f6c c0003e04 c0039b8 Corel Computer Corp.
c4
c0003e00: c003b050 c0003e38 c0003f98 c004265c c00155cc 00000000 c01fda00 000000 (hardware revision 7, serial number 80000558)
30
c0003e20: c01176c8 c0003ea4 00000001 c0003e58 c0003e3c c0042820 c0042598 c0003e
empe
Calibrating delay loop... 207.67 B
6c
ircomm_tty: virtual tty driver for IrCOMM ( revision:Wed May 26 0
649:
c0003e80: 4401a11b 00000000 c0003ec0 00000000 00000000 00000000 00000000 000000
Based upon Swanse
eU
Part
Starting kswapd v 1.5NE
000
c0003ea0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 000000ns enabledda2 hda3 hda4
00
c0003ec0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 000000
empeg-
ttyS01 at 0xf8050000 (irq = 17) i
00 S
c0003ee0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 000000irq = 16) is a SA1100 UARTd root has d_count=1
00
c0003f40: c0100798 c0008000 c0008000 c0100770 c0100798 c0003f98 4401a11b c00085driver initialized: 16 RAM disks of 4096K size92 blockstialised (first boot)
c0
c0003f60: c0003f90 c0003f70 c000eef0 c0039b64 c0100798 c000a248 0000000b c00020peg single channel IDE blocks (0.00%) reserved
00
c0003f80: c0003fd4 00000000 c0003f94 c000a0a0 c000eec0 c0100798 c00e730c c01007C25N080ATMR04-0, ATA DISK drive disk dr
70

c0003fa0: c0100798 c00e730c c0100798 c0100770 c0100798 c0100770 c0100988 4401a1
Block size=1024 (log=
hdb: I
1bN0
hdb: hdb1 < hdb5 hdb6 > hdb2 hdb3 hdb4
94
Backtrace:
RAMDIS
Function entered at [<c002a34c>] from [<c002ab20>]
RAMDISK: Loading 320 blo
r10 = 00000000 r9 = C0103F18 r8 = C0003BCC r7 = 00000005
VFS: Mounted root (ext2 fi
r6 = 00000000 r5 = C0159748 r4 = C0159740
empeg-pump v0.03 (19980601)
Function entered at [<c002a8f4>] from [<c0033ed4>]
Freeing unused kernel
r7 = 00000400 r6 = 00000000 r5 = C0103FB8 r4 = 00000000
I see this is a developer image!
Function entered at [<c0033e5c>] from [<c0033f9c>]
Mounting first music partition
Function entered at [<c004a868>] from [<c004adcc>]p v1.00 20000601 ([email protected])
r10 = 00000000 r9 = 00000008 r8 = C0003CEC r7 = 00000000

e000 v1.04
r10 = 00000006 r9 = C0100798 r8 = C01879C8 r7 = 00000000
r6 = C01879C8 r5 = 00000000 r4 = C0003D10
Function entered at [<c004bdc4>] from [<c004c12c>]
r10 = 00000001 r9 = C0100798 r8 = C0003DA4 r7 = C01879C8
r6 = C01879C8 r5 = C0185360 r4 = 00000000
Function entered at [<c004c0f0>] from [<c003ac24>]
r6 = C01852E0 r5 = FFFFFFF4 r4 = C0185360
Function entered at [<c003ab98>] from [<c003af10>]
r10 = 00000001 r8 = C000800C r7 = C01852E0 r6 = 00000001
r5 = 00000000 r4 = C0003DA4
Function entered at [<c003ad9c>] from [<c003b0a4>]
r10 = 00000000 r8 = 00008000 r7 = C0100770 r6 = 00000000
r5 = C0008000 r4 = 00000000
Function entered at [<c003b044>] from [<c0039bc4>]
r10 = 00000000 r9 = C0100798 r8 = 00000001 r7 = C0100770
r6 = C0003E8C r5 = C0008000 r4 = C0008000
Function entered at [<c0039b58>] from [<c000eef0>]
r10 = C00085C0 r9 = 4401A11B r8 = C0003F98 r7 = C0100798
r6 = C0100770 r5 = C0008000 r4 = C0008000
Function entered at [<c000eeb4>] from [<c000a0a0>]
r8 = C0003FD4 r7 = C0002000 r6 = 0000000B r5 = C000A248
r4 = C0100798
Code: e3520000 1a00000b (e5884000) e2844004 ea00000a

Here is the second one:

empeg-car bootstrap v1.00 20000601 (hugo@e
Dentry hash table entries: 2
b4 (
c0003fe0: c0009b88 c012b7ec c0100760 c0100764 c00fffd8 c0004000 c000b6cc c0009ble entries: 16384 (order 4, 64k)0
94ia
Backtrace
94
Backtrace:c0003a6c c
Function entered at [<c002a34c>] from [<c002ab20>]58 000000
Calling
POSIX conformance te
r10 = 00000000 r9 = C0103F18 r8 =
01

r4 = 000004000000001 r9 =
Function entered at [<c003376c>] from [<c00375cc>]0007 00000000 00000000 000001fa 5c2d2f7c 00000000
r10 = 00000069 r9 = 00000000 r8 = 00000400 r7 = 00000000
empeg dsp initialised
04ta
c0003b80: 00000000
r6 = 00000000 r5 = 00000001 r4 = 00001000 00000000 000000 with no serial op
empeg rem
Function entered at [<c0037458>] from [<c00a10c0>]
00
c0003ba0: 00000000 00000000 00000000 000000
r10 = C0003C54 r9 = 00000140 r8 = 00000000 r7 = C0003CAC
empeg state support initialised 0089/88c1 (save to d0
00d
r6 = C0161400 r5 = 00000069 r4 = 0000006900 00000000 00000000 00000000 000000 initia
Function entered at [<c00a0cfc>] from [<c00a12a4>]river initialised (fir
00bo
c0003be0: 00000000 000
r10 = C00085C0 r9 = 4401A11B r8 = C0122520 r7 = C012AE80RAM disks of 4096K size
r10 = 00000000 r9 = C01
r6 = 00000040 r5 = 0000

empeg-car bootstrap v1.00 20000601 ([email protected])
e
Partition ch
or forever hold their

If there is anyone present who wants to upgrade the flash, let them speak now,000000

e000 v1.04
Copying k

e000 v1.04
Copying k
or forever hold their peace...it seems not. Let fly the Penguins of Linux!000 00000000 000000................................... done, booting the k

e000 v1.04
Copying kernel... 00000000 0000000
Calling linux kernel...00000 00000000 00000000
Uncompressing Linux..................................... done, booting the kerne
00l.
c0003d60: 00000000 00000000 00000000 00000000 00000000 00000000 000
l.00
Linux version 2.2.17-rmk5-np17-empeg52-hijack-v388 (root@ibb
00

TCP: Hash tab
IP Protocols: ICMP, UDP, TCPh
c0003e80: 00000000 0000000
TCP: Hash tables configured (ehash 16384 bhash 16384)ls for Linux-2.2 (Dag Brattli)0 on fd=5
IrDA (tm) Protocols for Linux-2.2 (Dag Brattli)000 00000000 00000000 00000000 00000000 000000
Starting kswapd v 1.5options enablednce te
SA1100 serial driver version 4.27 with no serial options enabled3ec0: 00000000 00000000 00000000 00000000 00000000 00000000 000
ttyS00 at 0xf8010000 (irq = 15) is a SA1100 UARTirq = 17) is a SA1100 UART
00
ttyS01 at 0xf8050000 (irq = 17) is a SA1100 UART0000000 00000000 00000000 000000utton initialise
ttyS02 at 0xf8030000 (irq = 16) is a SA1100 UART'emp '
00

empeg display initialised.00000000 00000000 00000000
empeg dsp audio initialised
Could not find CS4231A (version=80)
empeg remote control/panel button initialised.
empeg usb initialised, PDIUSBD12 id 1012
empeg state support initialised 0089/88c1 (save to d0004e80).
empeg RDS driver initialised
empeg power-pic driver initialised (first boot)
RAM disk driver initialized: 16 RAM disks of 4096K size
empeg single channel IDE
Probing primary interface...
hda: IC25N080ATMR04-0, ATA DISK drive
ide0 at 0x000-0x007,0x038 on irq 6
hda: IC25N080ATMR04-0, 76319MB w/7884kB Cache, CHS=9729/255/63
empeg-flash driver initialized
smc chip id/revision 0x3349
smc9194.c:v0.12 03/06/96 by Erik Stahlman ([email protected])

SMC9194: SMC91C94(r:9) at 0x4008000 IRQ:7 INTF:TP MEM:6144b MAC 00:02:d7:10:02:2
e
Partition check:
hda: hda1 < hda5 hda6 > hda2 hda3 hda4
RAMDISK: ext2 filesystem found at block 0
RAMDISK: Loading 320 blocks [1 disk] into ram disk... <0>Kernel panic: VFS: Free
block list corrupted

Whoops! Sorry about that, Melkon.. The orange colour avatar fooled my eyes!

I think I found the bug.. Hijack v389 will be out shortly, with a fix for all non-Mk2a players.

Thanks (and let me know if it works or not).

Cheers

Okay, go wild.

WORKS!

empeg-car bootstrap v1.00 20000601 ([email protected])
On DC power, quick boot

e000 v1.04
Copying kernel...
Calling linux kernel...
Uncompressing Linux..................................... done, booting the kerne
l.
Linux version 2.2.17-rmk5-np17-empeg52-hijack-v389 (root@ibbm) (gcc version 2.95
.3 20010315 (release)) #2 Thu May 6 16:13:00 EDT 2004
Processor: Intel StrongARM-1100 revision
Checking for extra DRAM at c0c00000: wrote 00000000, read c00e78e0
NetWinder Floating Point Emulator V0.94.1 (c) 1998 Corel Computer Corp.
empeg-car player (hardware revision 7, serial number 80000558)
Command line: mem=12m noinitrd
Calibrating delay loop... 207.67 BogoMIPS
Memory: 10960k/12M available (980k code, 20k reserved, 324k data, 4k init)
Dentry hash table entries: 2048 (order 2, 16k)
Buffer cache hash table entries: 16384 (order 4, 64k)
Page cache hash table entries: 4096 (order 2, 16k)
POSIX conformance testing by UNIFIX
Linux NET4.0 for Linux 2.2
Based upon Swansea University Computer Society NET3.039
NET4: Linux TCP/IP 1.0 for NET4.0
IP Protocols: ICMP, UDP, TCP
TCP: Hash tables configured (ehash 16384 bhash 16384)
IrDA (tm) Protocols for Linux-2.2 (Dag Brattli)
Starting kswapd v 1.5
SA1100 serial driver version 4.27 with no serial options enabled
ttyS00 at 0xf8010000 (irq = 15) is a SA1100 UART
ttyS01 at 0xf8050000 (irq = 17) is a SA1100 UART
ttyS02 at 0xf8030000 (irq = 16) is a SA1100 UART
Signature is 20706d65 'emp ': [<c002a6d4>] lr : [<c00
Found custom animation at offset 0x98b80
Tuner: loopback=1, ID=-1 : c00037ec ip : 400000
Forced AC/Home mode 00000000 r4 = 000
empeg display initialised.
empeg dsp audio initialised
c0003a60:
r10: 0
empeg dsp mixer initialised 00000000Function entered a
empeg dsp initialisedc00
empeg audio-in initialised, CS4231A revision a0
r7 : c0ffafe0 r6 : c01597
empeg remote control/panel button initialised.a80: 0082000
empeg usb initialised, PDIUSBD12 id 1012
Function entered at [<c003376
empeg state support initialised 0089/88c1 (save to d0005100).r1 : c0157f28 r0 : 000000097, serial n
empeg RDS driver initialised
r10 = 00
empeg power-pic driver initialised (first boot)

hda: IC25N080ATMR04-0, 76319MB w/7884kB Cache, CHS=9729/255/63
Process swapper (pid: 1, stackpage=c0003000)
empeg-flash driver initialized r10 = C0003C54 r9 = 00000140
smc chip id/revision 0x3349
Stac
smc9194.c:v0.12 03/06/96 by Erik Stahlman ([email protected]) c002a50 r5 = 00

SMC9194: SMC91C94(r:9) at 0x4008000 IRQ:7 INTF:TP MEM:6144b MAC 00:02:d7:10:02:2
e
Partition check:
Func
empeg init 0.8
I see this is a developer image!
Mounting proc
Mounting first music partition
Tried to mount /dev/hda4 as reiserfs but got error 19
Mounting second music partition
Tried to mount /dev/hdc4 but got error 6
Error mounting partitions (possibly already mounted)
Remounting first music partition read-only
Remounting second music partition read-only
No secondary hard disk
Press 'q' now to go into development mode. You Have Zero Seconds To Comply...
Starting player
Timezone: /usr/share/zoneinfo/Atlantic/South_Georgia
Hijack: intercepting config.ini

hijack: removed menu entry: "Serial Port Assignment"
khttpd: listening on port 80
kftpd: listening on port 21
Using non-standard cache size 38 (adjustment 12)
player.cpp : 564:empeg-car 3.00-alpha7 2004/02/13.
Prolux 4 empeg car - 2.1490 Feb 13 2004
Vcb: 0x4032e000

Okay, good stuff!

Now all you need is some DRAM and a steady soldering hand.

Anyone who tries this hardware upgrade should do it BEFORE having their daily caffeine intake! Much easier that way!

Cheers

ahhh I was thinking about, that this is the price for the best and fastest Hijack code tester.

bye

It's probably easier to get a steady hand than the DRAM.

>ahhh I was thinking about, that this is the price for the best and fastest Hijack code tester.

Actually, Melkon: If you can supply the chips (Hugo and Stu both have them, I think), then send me the mainboard of your player and I'll install them for you, no charge.

Cheers

Mark

I wonder if the DRAM is the same as the ram used for the Series 1 SA TiVo upgrades.

See Hugo's post here. I believe it is the same used for upgrading the original Tivo at least for the Mk2a upgrade.

Specifically, Hugo said:

The Mk1/Mk2 chips are 1M x 16. I have shedloads of these (at least 1000 pcs)

.

Cheers

Ok, ok. It seems to work on Jolene (Mk2a@16MB). Consider this a before picture.


e000 v1.04
Copying kernel...
Calling linux kernel...
Uncompressing Linux..................................... done, booting the kernel.

Linux version 2.2.17-rmk5-np17-empeg52-hijack-v389 (root@ibbm) (gcc version 2.95
.3 20010315 (release)) #2 Thu May 6 16:13:00 EDT 2004
Processor: Intel StrongARM-1100 revisi
Checking for extra DRAM at c1000000: wrote 00000000, read c00e78e0
NetWinder Floating Point Emulator V0.94.1 (c) 1998 Corel Computer Corp.
empeg-car player (hardware revision 9, serial number 10101670)
Command line: mem=16m
Calibrating delay loop... 207.67 BogoMIPS
Memory: 15016k/16M available (980k code, 20k reserved, 364k data, 4k init)
Dentry hash table entries: 2048 (order 2, 16k)
Buffer cache hash table entries: 16384 (order 4, 64k)
Page cache hash table entries: 4096 (order 2, 16k)
POSIX conformance testin
Linux NET4.0 for Linux 2.2

-snip-

Starting player
Timezone: /usr/share/zoneinfo/Atlantic/South_Georgia
Hijack: intercepting config.ini

hijack: removed menu entry: "Hard Disk Detection"
hijack: removed menu entry: "Serial Port Assignment"
khttpd: listening on port 80
kftpd: listening on port 21
player.cpp : 385:empeg-car 2.00 2003/04/01.
Prolux 4 empeg car - 2.1434 Mar 26 2003
Vcb: 0x4086d000

Hiyas,

This is way cool Mark! I'm not sure what I need more memory for yet, but it could be great for future use

Does anyone know a source for the MK2a memory modules? If not and there is an interest in getting a bunch, I could try calling up a friend I have working for Micron and see if they still have them somewhere...

// Jonathan

If you have a friend in Micron, that would be helpful...

If that doesn't work out, I am in the middle of a bulk order. I am getting a quote now, so I will know more soon. Please hold on individual orders (unless you really want to...).

I can now confirm that the added memory on my Mk2a (Rita) is NOT used at all by the player software. It has exactly the same memory footprint with and without the extra 16MB installed.

But third-party apps.. and the Linux page (disk) cache.. and the built-in webserver.. and most likely player SYNC operations... all of these will take full advantage of added memory space.

And on Mk2 players, the extra 4MB probably *is* used by the player, though this will require confirmation.

Eventually, somebody will give us an alpha release that can make use of more memory, rather than using totally hardcoded internal limits (which were fine, until we started hacking things ).

EDIT: I can also confirm that negative ReserveCache=xxx values are completely ignored by the player

Cheers

...Eventually, somebody will give us an alpha release that can make use of more memory, rather than using totally hardcoded internal limits...

Um, that sounds like a big hint to all those folks out there...

hmm, is it possible to HEX the amount of ram allocated by the player? a dirty hack.

I've sent the guy at Micron a mail, hopefully I will have a reply by tomorrow evening.

// Jonathan

I suspect it may be possible to edit the player binary, if someone wants to disassemble it and find where they access the ReservedCache variable, and then patch it to be a bigger default number.. or to permit '-' signs..

Thank you Mark! Let's see, what we can do.....

bye

I actually am working at Micron in Manassas, VA. If your contact doesnt work out let me know maybe I could look around at work and find someone to help us out.

Greetings!

Just in case, I am following through with the alternate order. You can never have too much memory!!!

Link to memory bulk order info in the For Sale forum.

Got the reply from Micron,
they are out of stock world wide on them.
There's plenty of "MT4LC16M4A7TG-5" in the US, but I didn't find a proper datasheet though I expect it to not be compatible.
There's also a "planned" build of the chip, MT4LC4M16R6TG-5 in the US, due end of July.

// Jonathan

Quote:
On my upgraded Mk2a player (32MB total), the player software seems to make use of at least a few MB of the extra memory for holding the database. But it does not use most of it yet -- need a newer alpha release (from peter or rob) for that to happen.

Suppose, hypothetically speaking, a new alpha release was in the works, and one of the target features was out-of-the-box support for bonus RAM. Rather than duplicating all the effort, it'd seem wise to use the mighty power of the GPL to add the relevant bits of Hijack back into the standard kernel (with full Mark Lord attribution of course). Would I be right in saying that the bonus RAM detection is localised in arch/arm/mm/mm-sa1100.c, include/asm-arm/sa-1100/memory.h, and arch/arm/kernel/setup.c? Or are there some other bits hiding somewhere I've not found? (head-armv.S seems to have been patched and then unpatched).

Peter