DOS binary doesn't work?

[max]

I tried to run the DOS binary found here: http://khepri.coresystems.de/~idwer/flashrom/

But it gave this message:

No coreboot table found.
DMI pipe open error

Is this version supposed to work? Something I'm doing wrong?

Thanks,
Dan

[Idwer Vollering]

2010/10/4 max
> I tried to run the DOS binary found here:
> http://khepri.coresystems.de/~idwer/flashrom/http://khepri.coresystems.de/%7Eidwer/flashrom/

I'll assume you used the DOS .exe from the r1181/ (which is a symlink to
latest/ ) directory.
> But it gave this message:
>
> No coreboot table found.
> DMI pipe open error

Are you running DOS and/or flashrom from write-protected media ?

Which variant of DOS are you using, MS-DOS, FreeDOS ?
And on what hardware ?
> Is this version supposed to work?

Versions 0.9.2, r1181 and r1184 are known to work, tested by me using
FreeDOS.
> Something I'm doing wrong?

Are you able to paste/show the verbose output (flashrom -V) ?
> Thanks,
> Dan
>
> flashrom mailing list
> [e-mail]
> http://www.flashrom.org/mailman/listinfo/flashrom

[Carl-Daniel Hailfinger]

On 04.10.2010 03:19, Idwer Vollering wrote:
> 2010/10/4 max
> > I tried to run the DOS binary found here:
> > http://khepri.coresystems.de/~idwer/flashrom/
>
> I'll assume you used the DOS .exe from the r1181/ (which is a symlink to
> latest/ ) directory.
> > But it gave this message:
> >
> > No coreboot table found.

That is OK and expected if your system is not running coreboot.
> > DMI pipe open error

That means flashrom could not find dmidecode.exe . This is not a real
problem unless flashrom tells you that it could not identify your mainboard.

Regards,
Carl-Daniel

[max]

Apparently DOS flashrom is trying to write to drive "C:", and if that fails, it exits with no error message. Using -V option does not give any additional messages.

Any idea why flashrom is trying to write to "C:", and how much disk space it requires?

Thanks,
Dan

[Rudolf Marek]

> Any idea why flashrom is trying to write to "C:", and how much disk space it requires?

Yes it is creating maybe a pagefile for virtual memory. I think we can 
use cwsparam to force it not to use any.

R.
> Thanks,
> Dan
>
> flashrom mailing list
> [e-mail]
> http://www.flashrom.org/mailman/listinfo/flashrom

[Rudolf Marek]

On 4.10.2010 18:44, Rudolf Marek wrote:
> > Any idea why flashrom is trying to write to "C:", and how much disk
> > space it requires?
>
> Yes it is creating maybe a pagefile for virtual memory. I think we can
> use cwsparam to force it not to use any.

use cwsparam.exe to edit the cwdpmi.exe and put there "" when it ask for 
filename. I think then it wont try to write anymore. If the dmidecode.exe 
already contains the DPMI server you can run it directly on the binary.

Maybe we should update the README and describe how to do that as well as how to 
create a binary which contains a DPMI server (plus way how to compress that with 
UPX).

Thanks,

Rudolf
> R.
> > Thanks,
> > Dan
> >
> > flashrom mailing list
> > [e-mail]
> > http://www.flashrom.org/mailman/listinfo/flashrom
>
> flashrom mailing list
> [e-mail]
> http://www.flashrom.org/mailman/listinfo/flashrom

[Idwer Vollering]

2010/10/4 Rudolf Marek
> On 4.10.2010 18:44, Rudolf Marek wrote:
> > Any idea why flashrom is trying to write to "C:", and how much disk
> > > space it requires?
> >
> > Yes it is creating maybe a pagefile for virtual memory. I think we can
> > use cwsparam to force it not to use any.
>
> use cwsparam.exe to edit the cwdpmi.exe and put there "" when it ask for
> filename.

For archiving purposes: this works on cwsdstub.exe, too. cwsparam
cwsdstub.exe swapfile=""

Dan: can you try the rebuilt .exe from
http://khepri.coresystems.de/~idwer/flashrom/r1184/ ? Rebuild against
forementioned change.
> I think then it wont try to write anymore. If the dmidecode.exe already
> contains the DPMI server you can run it directly on the binary.
>
> Maybe we should update the README and describe how to do that as well as
> how to create a binary which contains a DPMI server (plus way how to
> compress that with UPX).
>
> Thanks,
>
> Rudolf
> > R.
> >
> > Thanks,
> > > Dan
> > >
> > > flashrom mailing list
> > > [e-mail]
> > > http://www.flashrom.org/mailman/listinfo/flashrom
> >
> > flashrom mailing list
> > [e-mail]
> > http://www.flashrom.org/mailman/listinfo/flashrom
>
> flashrom mailing list
> [e-mail]
> http://www.flashrom.org/mailman/listinfo/flashrom

[max]

--- On Mon, 10/4/10, Idwer Vollering  wrote:
> Dan: can you try the rebuilt .exe from http://khepri.coresystems.de/~idwer/flashrom/r1184/
> ? Rebuild against forementioned change.

It works the same (i.e. doesn't work), except that it no longer tries to create a swap file. Evidently the swap file was not the problem.

Thanks,
Dan

[Idwer Vollering]

2010/10/4 max
> --- On Mon, 10/4/10, Idwer Vollering  wrote:
> > Dan: can you try the rebuilt .exe from
>
> http://khepri.coresystems.de/~idwer/flashrom/r1184/http://khepri.coresystems.de/%7Eidwer/flashrom/r1184/
> > ? Rebuild against forementioned change.
>
> It works the same (i.e. doesn't work),

It does not work on ? FreeDOS ? MSDOS ? ... ?
> except that it no longer tries to create a swap file. Evidently the swap
> file was not the problem.

Can you retry with the download from
http://khepri.coresystems.de/~idwer/flashrom/r1184-without-cwsdstub-and-upx/...
You need cwsdpmi for this build, the download location is
http://homer.rice.edu/~sandmann/cwsdpmi/csdpmi7b.zip
> Thanks,
> Dan
>
> flashrom mailing list
> [e-mail]
> http://www.flashrom.org/mailman/listinfo/flashrom

[max]

--- On Tue, 10/5/10, Idwer Vollering  wrote:
> Can you retry with the download from http://khepri.coresystems.de/~idwer/flashrom/r1184-without-cwsdstub-and-upx/
> ? You need cwsdpmi for this build, the download location is
> http://homer.rice.edu/~sandmann/cwsdpmi/csdpmi7b.zip

This version is uncompressed so I could search it for strings. I discovered that the environment variable TEMP must be set to a writable drive. If TEMP is not set, then it writes to "C:".

Thanks,
Dan

[Idwer Vollering]

2010/10/5 max
> --- On Tue, 10/5/10, Idwer Vollering  wrote:
> > Can you retry with the download from
>
> http://khepri.coresystems.de/~idwer/flashrom/r1184-without-cwsdstub-and-upx/http://khepri.coresystems.de/%7Eidwer/flashrom/r1184-without-cwsdstub-and-upx/
> > ? You need cwsdpmi for this build, the download location is
> > http://homer.rice.edu/~sandmann/cwsdpmi/csdpmi7b.ziphttp://homer.rice.edu/%7Esandmann/cwsdpmi/csdpmi7b.zip
>
> This version is uncompressed so I could search it for strings. I discovered
> that the environment variable TEMP must be set to a writable drive. If TEMP
> is not set, then it writes to "C:".

So it is working fine now ? Do you know what is being written to the TEMP
dir, since you used a default cwsdpmi.exe and/or cwsdstub.exe ?
What variant of DOS are you using ?
> Thanks,
> Dan
>
> flashrom mailing list
> [e-mail]
> http://www.flashrom.org/mailman/listinfo/flashrom

[max]

--- On Tue, 10/5/10, Idwer Vollering  wrote:
> So it is working fine now ? Do you know what is being
> written to the TEMP dir, since you used a default
> cwsdpmi.exe and/or cwsdstub.exe ?

I don't think it has anything to do with cwsdpmi. cwsdpmi gives a warning message if it can't create a swap file on "C:", but that is harmless. Flashrom seems to be trying to execute an external program, and the temp file may be related to that.

I can't say whether it's working fine yet because I'm still learning how to use it.
> What variant of DOS are you using ?

MS-DOS 7.1 mostly, but I also tried FreeDOS.

Thanks,
Dan

[Idwer Vollering]

2010/10/5 max
> --- On Tue, 10/5/10, Idwer Vollering  wrote:
> > So it is working fine now ? Do you know what is being
> > written to the TEMP dir, since you used a default
> > cwsdpmi.exe and/or cwsdstub.exe ?
>
> I don't think it has anything to do with cwsdpmi. cwsdpmi gives a warning
> message if it can't create a swap file on "C:", but that is harmless.
> Flashrom seems to be trying to execute an external program, and the temp
> file may be related to that.

There is a gdb for djgpp binaries:
ftp://ftp.univie.ac.at/applications/djgpp/v2gnu/gdb611b.zip
http://ap1.pp.fi/djgpp/gdb/6.8/gdb68b.zip
> I can't say whether it's working fine yet because I'm still learning how to
> use it.
> > What variant of DOS are you using ?
>
> MS-DOS 7.1 mostly, but I also tried FreeDOS.

You are seeing the same behaviour on FreeDOS too ?
Have you tried running flashrom with dmidecode.exe (
http://assembler.cz/flashrom/dmidecode.exe ) in the same directory ?
> Thanks,
> Dan
>
> flashrom mailing list
> [e-mail]
> http://www.flashrom.org/mailman/listinfo/flashrom

[max]

I'm having another problem now. I get the error:

"tried to set 0x4e to 0x3 on ICH4/ICH4-L failed"

I get the error whether dmidecode.exe is present or not.

flashrom under Linux works on the same machine. I tried setting "plug and play OS" option in BIOS off, but that didn't help.

Thanks,
Dan

[Idwer Vollering]

2010/10/8 max
> I'm having another problem now. I get the error:
>
> "tried to set 0x4e to 0x3 on ICH4/ICH4-L failed"

Can you paste/attach the full verbose log (flashrom -V > log) from DOS ?
> I get the error whether dmidecode.exe is present or not.
>
> flashrom under Linux works on the same machine.

On what hardware are you using flashrom ?
What is the manufacturer's URL of the board ?

I'd like to see the verbose output when running on linux.
> I tried setting "plug and play OS" option in BIOS off, but that didn't
> help.

[max]

Looking at the ICH4 data sheet, an SMI interrupt is generated when you enable BIOS writing.

Does flashrom-DOS disable SMI? If not, then the BIOS is probably resetting the bit.

(Under Linux the BIOS write enable bit is already set).

Thanks,
Dan

[Idwer Vollering]

2010/10/9 max
> Looking at the ICH4 data sheet, an SMI interrupt is generated when you
> enable BIOS writing.
>
> Does flashrom-DOS disable SMI? If not, then the BIOS is probably resetting
> the bit.

It should disable SMI, yes; I don't see why the cross compiled build
wouldn't.
It is the same code that is being compiled: see
http://www.flashrom.org/trac/flashrom/browser/trunk/README#L97
> (Under Linux the BIOS write enable bit is already set).

Perhaps linux handles SMI "I/O" (I couldn't think of a better name) better
dan DOS does ?
> Thanks,
> Dan

Idwer
> flashrom mailing list
> [e-mail]
> http://www.flashrom.org/mailman/listinfo/flashrom

[max]

--- On Sat, 10/9/10, Idwer Vollering  wrote:
> It should disable SMI, yes; I don't see why the cross
> compiled build wouldn't.

I searched the source code for "SMI" but didn't find anything. Can you point me to the source code where SMI is disabled?

Using -V did not give any information about SMI.

Thanks,
Dan

[Idwer Vollering]

2010/10/10 max
> --- On Sat, 10/9/10, Idwer Vollering  wrote:
> > It should disable SMI, yes; I don't see why the cross
> > compiled build wouldn't.

What I mean is that, in theory, the dos .exe does the same as a linux binary
-- in practice, it doesn't.
> I searched the source code for "SMI" but didn't find anything. Can you
> point me to the source code where SMI is disabled?
>
> Using -V did not give any information about SMI.

You are right at both points, have a look at how coreboot handles SMI:
http://tracker.coreboot.org/trac/coreboot/browser/trunk/src/southbridge/inte...
> Thanks,
> Dan
>
> flashrom mailing list
> [e-mail]
> http://www.flashrom.org/mailman/listinfo/flashrom