Problems building OMAP ROM

Looks like the version of egrep you’re using isn’t ARMv7-safe.

Thanks – referring back to Chris’s post should rectify that.

...I think BuildEnv is only really relevant if you’ve got the entire CVS tree on your machine…

OK, I’ll ignore it then. Thanks.

It’s an improvement but I now have a fonts problem!

[...]

ROMFonts...
amu -E  -k rom COMPONENT=ROMFonts TARGET=ROMFonts 
do mkdir -p o
objasm -depend !Depend -Stamp -o o.Iyonix s.ROMFonts
ARM AOF Macro Assembler 3.32 (Acorn Computers Ltd) [10 Feb 2005]
Error: File 'Fonts.Trinity.Medium.Italic.Outlines0' not found at line 44 in macro ResourceFile
 at line 247 in macro MyDesk
 at line 66 in file s.<UserIF>_Fonts
 included by GET/INCLUDE directive at line 252 in file "s.ROMFonts" 
   44 0004b988         DCD     :FLOAD:"Fonts.Trinity.Medium.Italic.Outlines0"  ; Load address of file
Error: File 'Fonts.Trinity.Medium.Italic.Outlines0' not found at line 45 in macro ResourceFile
 at line 247 in macro MyDesk
 at line 66 in file s.<UserIF>_Fonts
 included by GET/INCLUDE directive at line 252 in file "s.ROMFonts" 
   45 0004b98c         DCD     :FEXEC:"Fonts.Trinity.Medium.Italic.Outlines0"  ; Exec address of file
Error: File 'Fonts.Trinity.Medium.Italic.Outlines0' not found at line 46 in macro ResourceFile
 at line 247 in macro MyDesk
 at line 66 in file s.<UserIF>_Fonts
 included by GET/INCLUDE directive at line 252 in file "s.ROMFonts" 
   46 0004b990         DCD     :FSIZE:"Fonts.Trinity.Medium.Italic.Outlines0"  ; The length of the file
Error: File "Fonts.Trinity.Medium.Italic.Outlines0" not found at line 52 in macro ResourceFile
 at line 247 in macro MyDesk
 at line 66 in file s.<UserIF>_Fonts
 included by GET/INCLUDE directive at line 252 in file "s.ROMFonts" 
   52 0004b9c4         BIN     Fonts.Trinity.Medium.Italic.Outlines0           ; The file itself
Fatal error: Structure mismatch at line 52 in macro MyDesk
 at line 66 in file s.<UserIF>_Fonts
 included by GET/INCLUDE directive at line 252 in file "s.ROMFonts" 
   52 0004b9c4         BIN     Fonts.Trinity.Medium.Italic.Outlines0           ; The file itself
Assembly terminated: 
5 Errors, 0 Warnings
AMU: *** exit (8) ***

[...]

ROMFonts...
amu -E  install_rom INSTDIR=SCSI::HardDisc0.$.RISCOS.Install.ROOL.OMAP3.Video COMPONENT=ROMFonts TARGET=ROMFonts 
do mkdir -p o
objasm -depend !Depend -Stamp -o o.Iyonix s.ROMFonts
ARM AOF Macro Assembler 3.32 (Acorn Computers Ltd) [10 Feb 2005]
Error: File 'Fonts.Trinity.Medium.Italic.Outlines0' not found at line 44 in macro ResourceFile
 at line 247 in macro MyDesk
 at line 66 in file s.<UserIF>_Fonts
 included by GET/INCLUDE directive at line 252 in file "s.ROMFonts" 
   44 0004b988         DCD     :FLOAD:"Fonts.Trinity.Medium.Italic.Outlines0"  ; Load address of file
Error: File 'Fonts.Trinity.Medium.Italic.Outlines0' not found at line 45 in macro ResourceFile
 at line 247 in macro MyDesk
 at line 66 in file s.<UserIF>_Fonts
 included by GET/INCLUDE directive at line 252 in file "s.ROMFonts" 
   45 0004b98c         DCD     :FEXEC:"Fonts.Trinity.Medium.Italic.Outlines0"  ; Exec address of file
Error: File 'Fonts.Trinity.Medium.Italic.Outlines0' not found at line 46 in macro ResourceFile
 at line 247 in macro MyDesk
 at line 66 in file s.<UserIF>_Fonts
 included by GET/INCLUDE directive at line 252 in file "s.ROMFonts" 
   46 0004b990         DCD     :FSIZE:"Fonts.Trinity.Medium.Italic.Outlines0"  ; The length of the file
Error: File "Fonts.Trinity.Medium.Italic.Outlines0" not found at line 52 in macro ResourceFile
 at line 247 in macro MyDesk
 at line 66 in file s.<UserIF>_Fonts
 included by GET/INCLUDE directive at line 252 in file "s.ROMFonts" 
   52 0004b9c4         BIN     Fonts.Trinity.Medium.Italic.Outlines0           ; The file itself
Fatal error: Structure mismatch at line 52 in macro MyDesk
 at line 66 in file s.<UserIF>_Fonts
 included by GET/INCLUDE directive at line 252 in file "s.ROMFonts" 
   52 0004b9c4         BIN     Fonts.Trinity.Medium.Italic.Outlines0           ; The file itself
Assembly terminated: 
5 Errors, 0 Warnings
AMU: *** exit (8) ***

AMU: *** 'install_rom' not re-made because of errors ***

Error running make install_rom on module 'ROMFonts'.
Fatal error running make install_rom on module 'ROMFonts'.
Batched errors...
Error running make install_rom on module 'ROMFonts'.
------------------------------------------------------------------------------
Closing log file 'SCSI::HardDisc0.$.RISCOS.BuildSys.Logs.dUI800-00'.

Here’s the full log in case there’s some other useful info in it – no syntax highlighting specified because I didn’t know what (if anything) to choose.

[Edit: My first log from yesterday (which pastebin.com isn’t accepting right now) doesn’t seem to indicate exactly the same font issues.]

Any more pointers, please?

Interesting – can you check the contents of RiscOS.Sources.Video.Render.Fonts.ROMFonts.Fonts? The fonts it’s looking for should all be there.

Funnily enough, I’m just there (because it’s where pwd pointed me to).

I have two files:

• SCSI::HardDisc0.$.RISCOS.Sources.Video.Render.Fonts.ROMFonts.s.Iyonix_Fonts
• SCSI::HardDisc0.$.RISCOS.Sources.Video.Render.Fonts.ROMFonts.s.ROMFonts

There’s no Fonts directory! [Edit: yes, of course there is!]

Presumably there should also be something relating to the final contents of Resources:$.Fonts.

[Edit: I’ll add these and see what happens.]

[Edit: In fact, it looks like I can copy most of them directly from my current installed Resources:$.Fonts.]

[Edit: My RISCOS.Sources.Video.Render.Fonts.ROMFonts.s.Iyonix_Fonts only refers to Corpus, Homerton and Trinity, whereas the online live CVS has a couple more (and the ROMFonts CVS doesn’t have my two files but does have a lot more! For the record, my currently installed Resources:$.Fonts (2010-05-05 build) contains Corpus, Homerton, Trinity, WIMPSymbol, Encodings and Messages1.]

Have now found Corpus, Homerton and Trinity directories in the (unmerged) original ‘castle’ download folder. (I’m working on a merged copy here.)

Really sorry Jeffrey, I was in the ’s’ directory. In the directory you asked about Trinity etc. are there (and with an Aug 2005 filestamp, which corresponds to the live CVS ‘4 years’ age).

In that case, I’m not really sure why it would be failing. It’s odd how it’s only Trinity.Medium.Italic.Outlines0 that it’s complaining about, which is the second-to-last file it tries to include (see s.Iyonix_Fonts). Can you check that that file exists? If it couldn’t find any of the files at all then I’d expect it to have complained about more than just one of them.

The only other thing I can think of is that you were doing something else at the time of the build an inadvertantly changed the CSD. But since the exact same error is reported twice in your pastebin I doubt that would be the case.

Thanks so much for bearing with me. I’ve just checked in the original downloaded unmerged SCSI::HardDisc0.$.Public.Dev.ROOL.OMAPROM.2010-05-27.OMAP3Dev.castle.RiscOS.Sources.Video.Render.Fonts.ROMFonts.Fonts.Trinity.Medium.Italic directory. The Outlines0 file appears to have suffered a problem in setting the filetype. It’s called Outlines0,ff6000640 and is Text. I guess I’ll just rename it and change the filetype to ff6!

[Edit: I’ll also download the latest OMAP Tar/BZ2 file and see if the same error remains.]

That’s sorted it. ROM is built and working.

Thanks so much for all your assistance Jeffrey and sorry to have distracted you from the VFP work you’re on with.

I’ll follow up here after putting this morning’s OMAP source through UnTarBZ2 – I note that Outlines0 in the live CVS sources doesn’t seem to indicate anything out of the ordinary to me.

Jeffrey, I have just merged the latest download CVS archive and the build fails at the link stage. The error is

OMAPVideo…
amu -E rom_link ADDRESS=4229322220 LINKDIR=ADFS::Backup.$.Sources.omap.RiscOS.Install.ROOL.OMAP3.Video
COMPONENT=OMAPVideo TARGET=OMAPVideo DEBUG=TRUE
do mkdir -p linked
link -o linked.OMAPVideo -rmf -base 4229322220 aof.OMAPVideo RISC_OSLib:o.abssym
ARM Linker: (Error) Undefined symbol(s).
ARM Linker: dss_reset, referred to from aof.OMAPVideo.
ARM Linker: dsi_pll_program_for_dispc, referred to from aof.OMAPVideo.
ARM Linker: venc_set_tvmode, referred to from aof.OMAPVideo.
ARM Linker: calculate_dss_clock_divider, referred to from aof.OMAPVideo.
ARM Linker: Errors in link, no output generated.
ARM Linker: finished, 5 informational, 0 warning and 1 error messages.
AMU: exit (1)
AMU: ‘rom_link’ not re-made because of errors
Error running make rom_link on module ‘OMAPVideo’.
Fatal error running make rom_link on module ‘OMAPVideo’.
Batched errors…
Error running make rom_link on module ‘OMAPVideo’.

Is there a problem you are aware of ? I notice one of the last changes to the CVS was your fix for a similar problem, but this is included in the source I used. Is there anything else I might be missing ?

Cheers

I’ve got no idea what’s gone wrong there. The functions which it can’t find are from several different source files, and I’m fairly certain all of them have existed right from the first version of the module. Were there any other errors reported for OMAPVideo during the other ROM build stages? Have you checked the makefile and some of the source files (e.g. dss.c, which has dss_reset right at the start) against CVS to make sure the files aren’t corrupt?

SharedRISC_OSLib...
amu -E  rom_link ADDRESS=4229093708 LINKDIR=SCSI::HardDisc1.$.RISCOS.Install.ROOL.OMAP3.Lib COMPONENT=SharedRISC_OSLib TARGET=RISC_OSLib 
do <Perl$Dir>.perl build:xtentries > syms.C_Entries kernel.s.k_entries kernel.s.k_entries2 clib.s.cl_entries clib.s.cl_entry2
print rlib.swi { >> syms.C_Entries }
do <Perl$Dir>.perl build:xtentries > syms.A_Entries1 kernel.s.k_entries clib.s.cl_entries
egrep -v "^(0x00000000 . )?_swix?$" < syms.A_Entries1 > syms.A_Entries

Internal error: undefined instruction at &000212BA

Postmortem requested
212b6 in anonymous function
  Arg2: 0x000216c8 136904 -> [0xe1a0c00d 0xe92ddbf0 0xe24cb004 0xe15d000a]
  Arg1: 0x00025e20 155168 -> [0x65726765 0x762d2070 0x285e2220 0x30307830]
fc131778 in shared library function
83ec in function ___init
AMU: *** exit (1) ***

AMU: *** 'rom_link' not re-made because of errors ***

Strictly they’re not the very last lines, but the rest are no more than a consequence of this error.

Any ideas, please?

egrep -v ”^(0×00000000 . )?_swix?$” < syms.A_Entries1 > syms.A_Entries

Is this not the very same question as posted on 30 May 2010 9:36am in this very same thread? Keep up at the back!

(Looks sheepish…) Thank you, Sprow!

Having got the 2009 November versions of egrep etc., now the error changes to:

Cannot open syms.A_Entries1 for I/O redirection

Now I can’t be sure I understand the statement that caused the error, but I think it means that the file syms.A_Entries1 is being used as input, although I don’t know the path it’s expecting it to be on.

There is a file SCSI::HardDisc1.$.RISCOS.Sources.Lib.RISC_OSLib.syms.A_Entries1, whose access details show public read (and public write) as No.

There’s no reference to the above error line in this thread anyway :-)

Any more cluse, please?

Try *set UnixEnv$egrep$nonametrans "" (as suggested here)

Thanks, Jeffrey, that worked too.

Then I had the problem with the strange suffix to one of the Outlines0 files too. Does anyone know why that happens? There appears to be no such odd name in the result of the untar process, so I can’t understand at what stage it happens.

Finally, after umpteen attempts, it built the ROM. The HAL is identical to that in the last build but one, which is the one I’m using… which reminds me, I must update… and I must check whether the HAL is any different.

Building with the latest RISC_OSLib changes fails when it gets to Paint.

The two newley enabled functions are not available in the link stage.

These need to be uncommented in the Sources.Lib.RISC_OSLib.rlib.s.rl_entries file at lines 309 and 468.

The big step forward is that the C now compiles. (I have no idea yet whether it will work.)

The step back: SDMA fails with one ADR out of range error (no surprise, and easy to fix), but five ASSERT errors. I’ve just wasted a long time failing to get grep to work, but I’ve got it going now, and I’m on the case.

Oops, wrong thread… wrong forum… sorry.

Building with the latest RISC_OSLib changes fails when it gets to Paint.

Hi Tank,

Yes, this was a temporary breakage by one of our contributors which he fixed as soon as I let him know. This is something to keep in mind: there is no guarantee that the “Dev” builds even build correctly, let alone function correctly, at any given time. This is because they use the HEAD revision of nearly every component rather than (for example) the last known stable revision. If you want a stable build, you don’t check out the “dev” sources.

I’m sure everyone knows and understands that but I thought it was worth pointing out.

As for letting us know about the problem with RISC_OSLib, that’s useful stuff and the sooner we hear about breakages, the sooner we can get them fixed so thanks for that! :) If you spot anything else breaking, drop an email straight to us – mailto:code@riscosopen.org – so we can look into it more quickly than we might if it’s only reported on the forums.

I’m getting a build failure building the latest OMAP rom. It looks like this change hasn’t been added to the OMAP branch.

Ah, the joy of branches. I’ve updated the Cortex branch, so it should work now.

Steve: Can you think of any good reason why we can’t get rid of the Cortex branch of HdrSrc? There are barely any changes between it and the main branch, and I don’t have any immediate plans for making changes that could cause problems for Iyonix/IOMD builds.

No. I can’t see any reason to keep it.

I had build errors with the extracted sources. I’m currently looking for a pastebin which will accept the large logfile…

• ‘RiscOS.Images’ is empty
• Most of ‘RiscOS.Install.ROOL.OMAP3’ is also empty, with the exception of:
  • Desktop
  • FileSys
  • HWSupport
  • Internat
  • RISC_OS

[Edit: If it’s not a silly question, do I need the full ‘cvs-repos’ tarball too, and merge that with the ‘src-omap.5.17’ one?]

‘RiscOS.Images’ is empty

It starts off empty. That’s where your successfully built images should go.