Hi, as nobody replied to this thread started 22 dec 2008,
I am following up myself with a more dramatic subject ;-)

However, trying FTE-HX in dosemu, it seems that we did
NOT break ctrl-c handling in 2038. Instead, it seems that
ctrl-c handling is broken in FTE-HX for both FreeDOS 1.0
kernel 2036 and kernel 2038, which makes it unlikely that
some change after 1.0 is to blame.

Here is what dosdebug produces after "fte readme.txt",
shift cursordown and ctrl c for me:

EAX: 116702ce EBX: 116c30ff ECX: 00000097 EDX: 00000019
eflags: 00210282 ESI: 116a8a07 EDI: 116a3024 EBP: b35f8fd0
DS: 00d7 ES: 00d7 FS: 00f7 GS: 0000
CS:EIP= 00cf:116798a7 SS:ESP= 00d7:b35f8f50
#00cf:116798a7 61               popad
DPMI: do_cpu_exception(0x0e) at 0xcf:0x1167a31f,
ss:esp=d7:b35f8f40, #cr2=0xb35f8f40, err=0x4

00cf: 00000000 fffff000 Code32 3 P RA  (cache mismatch)
00d7: 00000000 fffff000 Data32 3 P WA  (cache mismatch)
00df: 00000d10 0000ffff Data32 3 P WA
00e7: 10e10000 00000fff Data32 3 P WA  (cache mismatch)
00ef: 000c5000 0000ffff Data32 3 P WA
(there seems to be no f7 in the LDT)

According to EBP, ESP and CR2, maybe something broke
the upper half of ESP or EBP here?

At a random moment during FTE operation, I have other
values for the stack (a7:0000nnnn) and for FS (e7) but
EBP is already > 64k (122aedfc). LDT is shown below:

008f: 080e9d7d 00000149 Code16 3 P R
0097: 080e9d7d 00000149 Code32 3 P RA
009f: b7bfd000 00000fff Data16 3 P W
00a7: 087577e0 0000efff Data32 3 P WA
00af: 0000b1d0 0000ffff Code16 3 P RA
00b7: 00002980 0000ffff Data32 3 P WA
00bf: 00002980 000000ff Data32 3 P WA
00c7: 00002850 0000ffff Data32 3 P WA
00cf: 00000000 fffff000 Code32 3 P RAa (cache mismatch)
00d7: 00000000 fffff000 Data32 3 P WA  (cache mismatch)
00df: 00000d10 0000ffff Data32 3 P WA
00e7: 1080c000 00000fff Data32 3 P WA  (cache mismatch)
00ef: 000c5000 0000ffff Data32 3 P WA



I wrote on 2008/12/22 on freedos-kernel:

> Forwarding to the kernel list...
> Original Mail is from Japheth.
>
> You can check the SVN for 2036
> to 2038 changes, probably in
> conio or the ctrl-c handling.

Japheth wrote on 2008/12/21 on freedos-user:

> Date: 2008/12/21
> Subject: [Freedos-user] Tiny Kernel incompatibility introduced in 2038?
> To: FreeDOS User <[email protected]>
>
> Hello,
>
> in the latest FreeDOS kernels (from 03/2008 and 09/2007,
> AFAIU they are called 2038 and 2038pre) there might be a
> tiny incompatibility concerning the Ctrl-C handling.
>
> As a test case I use this version of the FTE editor:
>
> http://www.japheth.de/Download/FTE-HX.zip
>
> Once a text file has been loaded and one wants to press
> Ctrl-C to copy a line, the program terminates due to an
> exception. The kernel 2036 (date of binary is 05/2007)

Page fault in dosemu... Can anybody reproduce it outside?

> has no problems, and other DOS versions (MS-DOS) also work
> flawless.

I get the same page fault with 2036, does it run somewhere?
Can somebody test with MS DOS or EDR DOS or similar?

Eric




------------------------------------------------------------------------------
Create and Deploy Rich Internet Apps outside the browser with Adobe(R)AIR(TM)
software. With Adobe AIR, Ajax developers can use existing skills and code to
build responsive, highly engaging applications that combine the power of local
resources and data with the reach of the web. Download the Adobe AIR SDK and
Ajax docs to start building applications today-http://p.sf.net/sfu/adobe-com
_______________________________________________
Freedos-kernel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/freedos-kernel

Reply via email to