I spent some time yesterday staring at makefiles and trying to figure out why the AmigaOneG3SE errors out in make. Far as I can tell at this point, the missing .depend / _depend target comes from the rules.mk located in the top uboot dir, and is included by the board's makefile. This seems to work for he sbc8641d board, but fails for amigaone. I don't see a significant difference in things, the amigaoen Makefile seems to include things the same way as the seemingly good sbc8641d run, and it seems that should also work for the amigaone as well. Until we get sources from someone else for a known good build, or at least a successful make run, I'll fiddle with figuring out why it fails the way I got it from the current uboot repository.

Anyone know if there is a debugger for make? I've learned some new things about make yesterday, but am still wading around in there trying to figure out some things, if a debugger exists it should help me understand variable values and what's actually happening in there without having to parse the Makefiles myself so much.

_________________
All glory to the Hypnotoad!

@billt

Quote:

billt wrote: Anyone know if there is a debugger for make? I've learned some new things about make yesterday, but am still wading around in there trying to figure out some things, if a debugger exists it should help me understand variable values and what's actually happening in there without having to parse the Makefiles myself so much.

What about these two:
http://gmd.sourceforge.net/ and
http://bashdb.sourceforge.net/remake/mdb.html (Debian Lenny has a package for this make clone)

BTW: I ordered the PLCC plugs today. Let's hope they arrive until the weekend!

Last edited by Geri on 16-Feb-2010 at 03:39 PM.

_________________
A1SE: G3@600MHz, 2GB, 1GBit network card
A1XE: G4@933MHz, 2GB, refitted AC'97 codec
microA1: G3@800MHz, 1GB

- A1 Linux support -

@billt

I'm in favor of waiting to see what m3x comes up with. I'm still waiting for sources from Thomas as well.

Trev

_________________
Sam440ep-flex 733 MHz/1 GB RAM/Radeon 9250/AmigaOS4.1 Update 2
borked A1200/Blizzard1260+SCSI-IV/Z4+MediatorZIV/Deneb/Voodoo3/CatweaselMk3
more borked A1200/MBX1200z/Indivision
A500/clockport/RRNet
A600/A603

@Trev

Quote:

Trev wrote: @billt I'm in favor of waiting to see what m3x comes up with. I'm still waiting for sources from Thomas as well.

Well, at least it won't hurt, if we take a look at the official code. Support for the A1 is broken in the official repository and IMHO we should fix it there. But I hope that m3x comes up with a nice source code package.

_________________
A1SE: G3@600MHz, 2GB, 1GBit network card
A1XE: G4@933MHz, 2GB, refitted AC'97 codec
microA1: G3@800MHz, 1GB

- A1 Linux support -

@Trev

I'm sure m3x's stuff will be very helpful, but it's something to do and learn my way around in the uboot environment a bit. I requested the 1.1.1 sources from Hans-Joerg and Ben. I know they've got a lot going on, hopefully they'll find time to respond to that soonish. And I do think that the official U-Boot repository should have the fixed and working build when we get there.

Last edited by billt on 16-Feb-2010 at 06:35 PM.

_________________
All glory to the Hypnotoad!

Anyone know of a linting tool for Makefiles? I'm having trouble tracking down an issue with the current u-boot sources from git repository, the AmigaOneG3SE Makefile seems to have a problem with recodnizing the .depend target, which is triggered by the _depend target.

But it seems that this .depend target should be present and available. It's located in top-level file rules.mk which is included by the AmigaOne makefile, same as how its done for the sbc8641d board which seems to work fine.

I've fiddled around with rules.mk and made the target a double-name target, changing
$(obj).depend:

to

.depend $(obj).depend:

and I do get a warning message from that telling me the target has two identical names, since $(obj) is not supposed to be defined to anything here. (I really wonder why it's all over the place at all since it seems supposed to be an empty variable)

I've tried the gmd make debugger as was suggested somewhere, but am having trouble being productive with that. It seems gmd lives and runs inside the make run environment. This particular error has make stop running when the AmigaOne makefile is made due to this error, so the debugger ceases to be and can't help me find the issue.

Is there a linting tool for makefiles? Google points me to PC-Lint, but I can't tell if that lints makefiles as well as C/C++ files, or if it's just stuff about running pclint commands inside of a makefile.

I've not yet downloaded the main u-boot 1.1.1 sources, maybe I should try that and see if it's an answer to getting the sources that made our AmigaOne XE images while we wait for responses from Hyperion folks, or if it at least runs and compiles.


edit:
just learned about the -d and --debug and -p flags to make. Will have to play with those too.

Last edited by billt on 18-Feb-2010 at 05:40 PM.
Last edited by billt on 18-Feb-2010 at 05:16 PM.

_________________
All glory to the Hypnotoad!

@billt

Did you try "remake"? It's description says:

Quote:

remake is a patched and modernized version of GNU make utility that adds improved error reporting, the ability to trace execution in a comprehensible way, and a debugger.

_________________
A1SE: G3@600MHz, 2GB, 1GBit network card
A1XE: G4@933MHz, 2GB, refitted AC'97 codec
microA1: G3@800MHz, 1GB

- A1 Linux support -

@Geri

never heard of remake. I'll give it a try when I get back to this. Thanks!

_________________
All glory to the Hypnotoad!

@billt

iirc there are missing files, that's why make fails on .depend target.

I managed to build a binary image (1.3.1) and an updater, but didn't had time to test it yet.

_________________
Massimiliano Tretene, ACube Systems, Soft3

@m3x

Thanks for the tip. Someone on u-boot mailing list suggested to look at the COBJS var regarding the bios emulator files, since the emulator was moved out of boards/MAI and into drivers/ instead for other boards to also use. I just wish that the Make errors suggeted more about what the missing files were...

I look forward to hearing more about your 1.3.1 work, but I'll also continue tinkering around in current. Have you made any bugfix or other updates to the code, or just fixing the build?

_________________
All glory to the Hypnotoad!

@billt

Quote:

Have you made any bugfix or other updates to the code, or just fixing the build?

Fixed the build, and merged, where possible, with code from the Sam-flex version. It's still too early for any bug fixes.

_________________
Massimiliano Tretene, ACube Systems, Soft3

@billt

I got the PLCC plugs now. I'll try to draw a Eagle symbol/package for the plugs until the weekend, then I can start with the layout.

_________________
A1SE: G3@600MHz, 2GB, 1GBit network card
A1XE: G4@933MHz, 2GB, refitted AC'97 codec
microA1: G3@800MHz, 1GB

- A1 Linux support -

@Geri

Cool. I've managed to get further with compile but not complete yet. Last complaint was something about includes that I didn't expect to be part of things, might be looking at the host Linux includes instead of just the DENX environment. But pretty sure it's getting the correct (CROSS_COMPILE)-gcc executable.

_________________
All glory to the Hypnotoad!

@m3x

Quote:

Fixed the build, and merged, where possible, with code from the Sam-flex version. It's still too early for any bug fixes.

This is very appreciated, thanks

_________________
Site admins are people too..pooff!

@m3x

Do have access to anything in the unofficial and hard to find AmigaOne uboot 1.2 build (or newer?) as well as whatever is included in your source tree? Or is that already part of your tree?

Are there even any AmigaOne/Articia specific updates in this compared to the public 1.1.1 release, or is it just changes in device drivers and other things outside of the board/MAI/AmigaOneG3SE area?

Last edited by billt on 24-Feb-2010 at 05:28 PM.

_________________
All glory to the Hypnotoad!

I have received another version of some sources. Also still need to find time to figure out problem on hello_world.srec dependency, it has a no target rule error. But getting further.

This problem seems to come from the % target which is dependency of %.srec target. The %.o dependency of the % target seems to cause this, the $(LIB) seems OK left there, removing %.o made ti try to go further. So this is hello_world.o but that file is present in my examples dir. Strange. There is a %.o target, I'll try to spend more time looking at that one too.

While this is in examples dir and we may not include it or have a true need for it, it would be nice to have everything compile out of the box. And it may be a nice thing to use in checking that our new uboot stuff works, being able to see that the computer is alive enough to run the examples when we're testing. And if we leave it out, there may be a similar problem that still needs fixed anyway.

While I and perhaps some others have checked out the official uboot sources, who here knows git well enough to make a branch that we can collaborate on to later submit to master branch? How does this all work with git? I've only used CVS and SVN before, and even there have never created or merged a branch.

Last edited by billt on 25-Feb-2010 at 06:22 PM.

_________________
All glory to the Hypnotoad!

I was just browsing on amiga.org and found a picture of some sort of A1 bios flashing board for an A500 some one created. http://www.amiga.org/gallery/index.php?n=973

Its probably not very useful for what you are working on but I thought I would post the link just in case :)

_________________
Sam440ep 667mhz, 512MB, 120GB 2.5" HD, OS4.1FE
WinUae 3.0.0, OS 3.9, BB3, Catweasel MkIV
Amiga 1200, Blizzard 040/40 (BlizzardPPC 060/200 with SCSI removed at present), mediatorSX pci, Voodoo3, PCI network card os 3.9BB2
4MB Minimig with ARM addon boar

@Dwyloc

Interesting, if not a little odd. :) But it will certainly be nice to be able to program the flash outside of the AmigaOne if there's no other way and somehting bad in the test code bricks the machine. I think Geri's switcher will do nicely, similar idea I'd imagine but all in the AmigaOne.

_________________
All glory to the Hypnotoad!

A very short status update:

I almost finished a first layout of the new PCB. It more or less looks like the layout I already posted here, but it is the best way to get a small and cheap dual layer PCB. I also created a library for the PLCC plug (reuses the symbol of the 29F040 flash). I can put the library and schematic online, if anyone is interested to start working on a MicroA1 version of the U-boot switcher PCB in the meantime.

_________________
A1SE: G3@600MHz, 2GB, 1GBit network card
A1XE: G4@933MHz, 2GB, refitted AC'97 codec
microA1: G3@800MHz, 1GB

- A1 Linux support -

*buuuuummmp*

@m3x

Any news on the U-boot source code?

@all

Who is actually interested in a microA1 version of the U-boot switcher board (for development purposes)?

_________________
A1SE: G3@600MHz, 2GB, 1GBit network card
A1XE: G4@933MHz, 2GB, refitted AC'97 codec
microA1: G3@800MHz, 1GB

- A1 Linux support -