www.delorie.com/archives/browse.cgi   search  
Mail Archives: djgpp/2003/02/06/17:47:18

From: "Francisco Pastor" <Francisco DOT Pastor AT uv DOT es>
Newsgroups: comp.os.msdos.djgpp
References: <eebb6570 DOT 0302051128 DOT 4ca566c AT posting DOT google DOT com> <9lj44vg0bh43925odl7eu977806vv0ft31 AT 4ax DOT com>
Subject: Re: remote debug stub and Windows 2000
Lines: 75
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2800.1106
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1106
Message-ID: <KXz0a.9618$md.598748@news.ono.com>
Date: Thu, 06 Feb 2003 20:54:02 GMT
NNTP-Posting-Host: 81.202.87.254
X-Complaints-To: abuse AT ono DOT com
X-Trace: news.ono.com 1044564842 81.202.87.254 (Thu, 06 Feb 2003 21:54:02 MET)
NNTP-Posting-Date: Thu, 06 Feb 2003 21:54:02 MET
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Reply-To: djgpp AT delorie DOT com

As Andrws Cottrell say, there is some problem with the fifo UART in a MS-DOS
virtual machine under WinNT or Win2K. The effect is that the packets that
gdb transmit are wrong.

The code work well under MS-DOS or Win95/98.

I solved this problem avoiding to use the fifo. Eli Zaretskii put a
conditional code with the #define  UART_FIFO_WORKS in the module ser-go32.c
to do that in gdb v5.1. I have not seen the sources in gdb v5.3.

You can try it rebuilding the code with this modification.

Good luck

Francisco Pastor
Francisco DOT Pastor AT uv DOT es

"Andrew Cottrell" <andnews AT ihug DOT com DOT oz DOT au> escribió en el mensaje
news:9lj44vg0bh43925odl7eu977806vv0ft31 AT 4ax DOT com...
> >I have built a demo program that uses i386-stub.c to remote debug from
> >a Windows 2000 machine.  My target machine appears to work correctly.
> >However I am having problems with GDB on the host.  I am using V5.3.
> >
> >The problem is as follows:
> >
> >+ I start the target - which contains a breakpoint() instruction.
> >
> >+ I start GDB on the host, then type "target remote COM1".  I know the
> >host the transmits data correctly as I can see it arriving on the
> >target.
> >
> >+ The first few commands transmitted by GDB are not supported by the
> >stub,
> >so I return $#00, as per the DGB FAQ.  I know the target transmits
> >this okay as I can see them arriving if I use a dumb terminal instead
> >of GDB.
> >
> >+ GDB reports errors like "Ignoring packet error, continuing..." * 3,
> >then
> >"Couldn't establish connection to remote target ... Malformed response
> >to
> >offset query, timeout".
> >
> >These errors make it look like GDB does not understand the reply -
> >however, if I unplug the serial cable the error messages are the same
> >so I think it is just that GDB is not seeing the replies at all.  On
> >the target Tx and Rx appear to work.  On the Windows 2000 host only Tx
> >appears to work.
> >
> >Has anybody any experience of this?  Also, does GDB assume hardware
> >flow control over the serial port?
> This has been reported by Nvidia in 2001 while DJGPP was being fixed
> so it woudl work under 2K.
>
> The problem is still on the list of items to be looked at, but I have
> not got to it because of more important issues that have poped up. I
> had a look a the code in GDB 5.1 & just now the 5.3 code to handle the
> serial ports and if I remember correctly (please correct me if I am
> wrong) I suspect that the code has not been updated in a while.
>
> If you have a look at PMCOM or DZCOMM and comapre the functions in the
> GDB 5.3 file gdb\ser-go32.c  with the simialr funcitons and I suspect
> that you may find the problem. I would look at the #ifdef
> UART_FIFO_WORKS conditional code first up and any other conditional
> code in  the ser-go32.c file.
>
> On some ALI chipset based motherbaords the PMCOM library will not, bit
> DZCOMM will.
>
> Let me know how you go and if you get stuck I may be able to point you
> in the right direction.
>
> Andrew


- Raw text -


  webmaster     delorie software   privacy  
  Copyright © 2019   by DJ Delorie     Updated Jul 2019