Hi everybody,
thanks for your responses... but finally I think it's a more general
problem, and more exactly a hardware problem, too bad :-(
Why ?
In fact, my setup is not as stable as yours seems to be. I've always
had software crashes in CT60 mode, but attributed this to buggy
software. But now, i'm almost sure it's a specific issue with my CT60.
I've not been able to get similar crashes in 030 mode with MiNT.
Here are some crashes (always signal 10) I've been able to catch with
mintdeb kernel through the serial port:
Example with HypView:
pid 49 (HYP_VIEW): get_cookie(): tag=44485354 (DHST) ret=020639D4
pid 49 (HYP_VIEW): get_cookie(): value 0000000A returned at 020639D4
pid 49 (HYP_VIEW): GET_COOKIE: return A
pid 49 (HYP_VIEW): attach_region: reg 18, return loc 1ACC000 (1ACC000)
pid 49 (HYP_VIEWid 49 (HYP_VIEW): [1][ PROCESS "HYP_VIEW" KILLED:
| MEMORY VIOLATION. (PID 049) | Type: free PC: 00FC535A | Addr:
01AB8000 BP: 0201A000
pid 49 (HYP_VIEW): signal #10 raised [syscall_pc 0x202F51C,
exception_pc 0xFC535A]
Taskbar, on shutdown
pid 10 (taskbar): [1][ PROCESS "taskbar" KILLED: | MEMORY
VIOLATION. (PID 010) | Type: free PC: 0114B3CE | Addr: 10002584
BP: 016B2000 |
pid 10 (taskbar): signal #10 raised [syscall_pc 0x16BB9AC,
exception_pc 0x114B3CE]
This one when starting Conholio 0.8:
pid 13 (conholio060): get_cookie(): tag=45644449 (EdDI) ret=01F7DD64
pid 13 (conholio060): get_cookie(): value 00FC59DE returned at 01F7DD64
pid 13 (conholio060): GET_COOKIE: return FC59DE
pid 13 (conholio060): signal #10 raised [syscall_pc 0xEC5126,
exception_pc 0xFC535A]
pid 13 (conholio060): system stack not in proc structure
pid 13 (conholio060): Fatal MiNT error: adjust debug level and hit a key...
FATAL ERROR. You must reboot the system.
When networking, MiNT is also killed with signal 10...
It seems that at moments, any program could try to read/write some
random addresses :-(
Does anyone has an idea why I have this sort of crashes ?
Here are some details about my setup:
Falcon 030 (of course) in original case
CentRAM 14 with EDO 60ns DIMM (recently changed to try...)
CT60 (one a the first batch in 2003) with CTCM like modification by
R.Czuba, but no solder and no bus speed enhancement
MC 68060 71E41J @ 95Mhz (recently, it was @ 66.6Mhz until some weeks ago)
256Mo SDRAM PC133 CL2
Flex ATX power 110W
CTPCI + Radeon 9200
IDE devices are CF and SD readers
I will contact R.Czuba to ask him if he knows something about that.
Thanks
Sébastien Liénard.
Le 23 février 2011 21:45, Sébastien Liénard (Gmail) <slix95@gmail.com> a écrit :
Hi everybody,
thanks for your responses... but finally I think it's a more general
problem, and more exactly a hardware problem, too bad :-(
Why ?
In fact, my setup is not as stable as yours seems to be. I've always
had software crashes in CT60 mode, but attributed this to buggy
software. But now, i'm almost sure it's a specific issue with my CT60.
I've not been able to get similar crashes in 030 mode with MiNT.
Here are some crashes (always signal 10) I've been able to catch with
mintdeb kernel through the serial port:
Example with HypView:
pid 49 (HYP_VIEW): get_cookie(): tag=44485354 (DHST) ret=020639D4
pid 49 (HYP_VIEW): get_cookie(): value 0000000A returned at 020639D4
pid 49 (HYP_VIEW): GET_COOKIE: return A
pid 49 (HYP_VIEW): attach_region: reg 18, return loc 1ACC000 (1ACC000)
pid 49 (HYP_VIEWid 49 (HYP_VIEW): [1][ PROCESS "HYP_VIEW" KILLED:
| MEMORY VIOLATION. (PID 049) | Type: free PC: 00FC535A | Addr:
01AB8000 BP: 0201A000
pid 49 (HYP_VIEW): signal #10 raised [syscall_pc 0x202F51C,
exception_pc 0xFC535A]
Taskbar, on shutdown
pid 10 (taskbar): [1][ PROCESS "taskbar" KILLED: | MEMORY
VIOLATION. (PID 010) | Type: free PC: 0114B3CE | Addr: 10002584
BP: 016B2000 |
pid 10 (taskbar): signal #10 raised [syscall_pc 0x16BB9AC,
exception_pc 0x114B3CE]
This one when starting Conholio 0.8:
pid 13 (conholio060): get_cookie(): tag=45644449 (EdDI) ret=01F7DD64
pid 13 (conholio060): get_cookie(): value 00FC59DE returned at 01F7DD64
pid 13 (conholio060): GET_COOKIE: return FC59DE
pid 13 (conholio060): signal #10 raised [syscall_pc 0xEC5126,
exception_pc 0xFC535A]
pid 13 (conholio060): system stack not in proc structure
pid 13 (conholio060): Fatal MiNT error: adjust debug level and hit a key...
FATAL ERROR. You must reboot the system.
When networking, MiNT is also killed with signal 10...
It seems that at moments, any program could try to read/write some
random addresses :-(
Does anyone has an idea why I have this sort of crashes ?
Here are some details about my setup:
Falcon 030 (of course) in original case
CentRAM 14 with EDO 60ns DIMM (recently changed to try...)
CT60 (one a the first batch in 2003) with CTCM like modification by
R.Czuba, but no solder and no bus speed enhancement
MC 68060 71E41J @ 95Mhz (recently, it was @ 66.6Mhz until some weeks ago)
256Mo SDRAM PC133 CL2
Flex ATX power 110W
CTPCI + Radeon 9200
IDE devices are CF and SD readers
I will contact R.Czuba to ask him if he knows something about that.
Thanks
Sébastien Liénard.
2011/2/17 Stefan Niestegge <beetle@abbuc.de>:
Hi Slix,
Does your ct60 use the bus speedup option? Overclocked bus seems to make trouble
with Ethernec and other Romport network solutions.
If you have bus speedup, it might be worth a try to switch it off. RAM is always a possible cause of the funniest effects, but since its stable otherwise and you tried other SDRAM bars already, i don't think its the problem.
There is no general software problem on this setup. I use my Falcon ct63 ctpci and NetUSBee for hours and hours and get no network problems.
Greetings,
Stefan
Am 17.02.2011 um 21:37 schrieb Peter Slegg <p.slegg@scubadivers.co.uk>:
On Wed, 16 Feb 2011 15:09:59 , Sébastien Liénard (Kerlinux) <slix@kerlinux.org> wrote:
Hi FreeMiNT guys,
I have some networking problems with FreeMiNT.
My freemint setup seems stable but networking always crashes the
system after few KB or MB of data transferred, with a message like
"Mint killed, you must reboot..."
Does it happen when uploading or downloading ?
Does it matter which app you are using ?
Does crash after a single file over a certain size or is it crashing
after a cumulatve number of bytes are transferred ?
Peter
I actually use a CT60+CTPCI+Radeon configuration and an Ethernec, with
Freemint 1.17.0 and enec6.xif (I also tried recently with Assemsoft
Driver with same result)
But this problem was already present with previous versions, 1.16.x
and 1.17-CUR, and with different hardware (no CTPCI, using PPP on
serial port).
Anybody has an idea ? Could it be a hardware problem (RAM ? I tried
with different DIMMs) ?
Thanks in advance,
SLiX.