let's see if it can show anything. Debug kernel with level 15 hasn't shown anything via the serial port when the problem appears.
Are you sure about 15? Because it hardly can boot even with 7 because of gazillion of output. My bet is that it wrongly set in firmware (maybe you forget add =, while should). I.e. the whole line should be for x5000:
Quote:
setenv os4_commandline serial munge debuglevel=5 saveenv
Because with level 15 you for real can't boot. Or it will take many minutes.
I had a "=", but at the wrong place on the line. Fixed now. Thanks.
Here's my Snoopy output. This is what happens after a clean boot of the system (Sam460ex@1150 MHz, Radeon HD 5450, RadeonHD.chip 1.21 and OS 4.1 FE update 2), starting Odyssey beta 7 from May 2020, browsing to eab.abime.net from the bookmarks:
What can actually be seen is that when the problem occurs the number of I/O Writes quickly raises up to the maximum number of 32 (shown with the shownetstatus ppc460ex_eth command).
At least 17 out of 32 I/O Writes doesn't make the ethernet stall.
By limiting the number of active network connections in Odyssey to 2 it will take longer before the eth stall happens.
If you have Radeon RX you may have hard times: there no RadeonRX lite driver in the OS4. So you need to modify install ISO to add Radeon RX driver + add kick layout line.
If RadeonHD, then all fine, you will have a lite driver (limited to 1024x768), which you can replace just by copying RadeonHD 3.7 driver over old lite on in sys: kickstart/ and reboot (then screen mode, check 1920x1080 here, if not, choose "detect automatically" or add mode manually, all in screen mode prefs).
After that warp3dnova, warp3d SI (for old warp3d/minilg, not part of Enhancer), and ogles2.library, and so on.
Thanks for your advice. Got everything up and running with just the bare minimum from the Enhancer pack. Ended up in a moment 22 due to my graphics card (my modified install ISO turned out to be crap), but luckily I had an older one laying around that I could use to boot up FE without the Enhancer drivers. Other than that it was smooth sailing and everything seems rock solid so far.
What I really wanted to say was thanks to all involved in this update. And thanks for OWB b7. Overall OS4 and X5000 is a really nice kit.
@sTix Maybe another advice will be for the time being to switch to the older kernel from update1: at least for now we don't know if it new kernel guilty for some random crap happens (like network stalls, unexpected messages about hanging forbids on exit from some programs, something with priorities/tasks handling, etc), or it just new kernel unhide bugs in some other os components.
Just if all is ok as it now, keep it, but if you will experience issues of the kind I describe, then switch back to the kernel from update1 for time being.
ps. and don't' forget to maximize your amidock with tray bar - put there dockies from zzd10h: netdock, ramdock, gfx dock, cpu dockies and from Javier: keymap switcher, mixer & datetime dockies. AmiDock can be not very intuitive to work with at first, but all you need can be done. Without them, the default setup looks a bit empty :)
@sTix In update2 there was a bit of SDK update just for those for want manually update some bits. I.e. no fresh SDK, just something which better than nothing. There you can find in update2 what is important enough:
1). updated SDK for DOS (as it quite massively improved) 2). updates to pthread lirbary (some critical bugs were fixed) 3). updates to newlib (with lots of fixes from Frederik)
Also some "C" tools (mostly for new DOS and to help developer stuff for, but also few others, check SDK/C). And that all. MUI and MiniGL SDK also included, but that nothing to worry about they all publicly online for those who need them.
You install it, and on top add from the latest SDK stuff from "base.lha" and what else you may need later + those parts of SDK from update2 on top, and you are ready.
@broadbluesQuote: That sounds like a bug in SimpleMail, it shouldn't be adding a new app icon each time.
Maybe SimpleMail isn't doing it right but the resulting behaviour is different from update1 to update2. In update1 the app icon gets placed over the existing app icon, effectively replacing it. In update2 the app icon gets placed in the next available position due to the new positioning code, thus creating a new app icon each time. The new behaviour is not backward compatible. Not suggesting Simplmail is doing it right though.
So sounds like SimpleMail has fallen foul of a fix that fixed arrangement of icons, which should never overlap.
Quote:
Are applications supposed to remove the old app icon before creating a new one?
Yes, except in rare circumtances, sounds like simple mail is just piling them up on the same spot, slowing draining resources.
Nice, I should have though of that. I'm using Marlon Beijers Docker image with 8.3 for cross-compiling, a very neat solution. I wouldn't mind having an SDK_Balbalba.lha though
Combination of X-Dock and third party program called CPU_Watcher (version 0.7) cause a lockup on my Sam460ex when using the copy menu in AmigaShell . No issue with the copy function if X-Dock is running but not CPU_Watcher. And no issue with the copy if CPU_Watcher is running but not X-Dock .
I was trying to duplicate the issue with the copy menu that NinjaCyborg was having with his clean AMigaOS 4.1 Update 2. Could not duplicate with my own clean Update 2.
What version of textclip-handler do you have installed? Could be unrelated but I have reported a bug with the latest version of this (53.3) that causes pasting to hang after writing to TEXTCLIP:0