[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [MiNT] XaAES 1.55 + NetSurf



Am Mittwoch, den 27.03.2013, 16:12 +0100 schrieb "Helmut Karlowski" <helmut.karlowski@ish.de>:
Helmut Karlowski, 27.03.2013 08:20:35:

With MyAES+MP I get:

pid  25 (ns020): get_cookie(): tag=45644449 (EdDI) ret=033A13CC
pid  25 (ns020): get_cookie(): value 0111C5AC returned at 033A13CC
pid  25 (ns020): GET_COOKIE: return 111C5AC
pid  25 (ns020): [1][ PROCESS  "ns020"  KILLED: | MEMORY VIOLATION.
(PID  025) | Type: super     PC: 0111C5AC | Addr: 0111C5AC  BP:
02E14000 |
pid  25 (ns020): MEMORY VIOLATION: type=super     RW=r  AA=111C5AC
PC=111C5AC BP=2E14000
pid  25 (ns020): signal #10 raised [syscall_pc 0x111C5AC,
exception_pc  0x111C5AC]


NetSurf requires the Eddi cookie to be present (to read screen info), is there anything interesting about your system (anything could be interesting =))?

It looks like the eddi function which is called (eddi version info) is causing the error (or maybe the wrong function address was returned).

Here the cookie is loaded (line 392):
http://git.netsurf-browser.org/netsurf.git/tree/atari/plot/plot.c

This is the code which is called to call the eddi function:
http://git.netsurf-browser.org/netsurf.git/tree/atari/plot/eddi.s

tos_getcookie is implemented here:
http://git.netsurf-browser.org/netsurf.git/tree/atari/osspec.c

So all that looks clean to me, but yes, calling (reading) a function which doesn't belong to the process memory is maybe forbidden in some mmu setups,...

Can you compare the eddi value with the value 111C5AC (cookies.cpx)? if it differs, then netsurf may smash that memory,... otherwise I would expect that the eddi cookie is wrong.

Hm... trying to compile XaAES 1.6.0 for coldfire now...