Login
Username:

Password:

Remember me



Lost Password?

Register now!
Main Menu
Who's Online
2 user(s) are online (2 user(s) are browsing Forum)

Members: 1
Guests: 1

mordesku, more...
   All Posts (twilen)


(1) 2 3 4 ... 8 »


Re: Max filesize

Joined:
2011/5/10 23:05
Group:
Member
Posts: 76
Offline
File access requires special care when reading >2G (or >4G, can depend on filesystem) files with 32-bit file functions: Seek()'s beyond 2G may not be possible, only linear reading.

But yes, 64-bit "wrapper" would be nice.

Posted on: 7/5 10:11
Transfer the post to other applications Transfer


Re: Max filesize

Joined:
2011/5/10 23:05
Group:
Member
Posts: 76
Offline
PFS3 "experimental" (Modifications by me and original PFS3 author) supports >4G files (and larger partition sizes) but it isn't production ready.

One problem is that it is not easy to test >4G files without special test programs because OS does not directly support it, at least not very well.

(If AROS 64-bit filesystem handling test cases is needed)

Posted on: 6/29 5:31
Transfer the post to other applications Transfer


Re: Vampire hyper threading and aros smp

Joined:
2011/5/10 23:05
Group:
Member
Posts: 76
Offline
Yes, it is perfectly fine if compiler does CPU-specific optimizations (just the way it should be) + some kernel changes so that all extra features work. The rest should be in machine-specific drivers anyway.

So there is no (at least no major) problems after all. I hope..

Posted on: 6/29 5:23
Transfer the post to other applications Transfer


Re: Vampire hyper threading and aros smp

Joined:
2011/5/10 23:05
Group:
Member
Posts: 76
Offline
That sounds fine, at least in theory.

In theory because there is no compiler or assembler support, "optimization" will unfortunately become something like "Lets do it in Apollo assembly in hexadecimal, we must micro-optimize everything in assembly, we have new instructions, they should be used!!1!" which won't help other builds.

Posted on: 6/29 0:03
Transfer the post to other applications Transfer


Re: Vampire hyper threading and aros smp

Joined:
2011/5/10 23:05
Group:
Member
Posts: 76
Offline
Bleh, not again those kinds of non-technical and useless replies from new posters..

There is big problem with AROS support. It has far too many features (really weird ones) that simply don't fit with AmigaOS design without a) losing compatibility b) writing each module twice (one m68k, one Apollo optimized). And aros-m68k main reason is compatibility with AmigaOS!

Apollo/Vampire stops being Amiga m68k compatible if any of its CPU extensions need support. It needs separate AROS platform. But good luck finding developers to do it, but as I said in other forum, it is far too developer hostile. None of existing tools support it, none of debugging tools support it, can't debug it in emulation. Nothing really supports it. So from any serious 680x0 Amiga -only developer point of it view it is and will be "only" very fast 680x0, nothing more, nothing less.

I am not going to accept (and if it happens, I will be permanently out of aros development) any Vampire/Apollo specific code to aros-m68k platform unless it makes 100% sense and won't add much extra code (rom space!), change is localized, something like CPU detection (that is already done so that user space programs can detect it without needing hacks) or optimized memory copy functions are fine.

Anything more complex -> separate aros platform, just like aros was designed to work: main tree/specific platform modules can be inherited directly or overridden with platform specific parts or even single files can be overridden.

Posted on: 6/28 0:54
Transfer the post to other applications Transfer


Re: (JANUS-UAE) Is today's nightly build for 68K running for you?

Joined:
2011/5/10 23:05
Group:
Member
Posts: 76
Offline
JIT problem = anything is possible, unrelated change can cause unexpected side-effects. Adding or removing debug statements most likely only moves the problem around really mysteriously.


Posted on: 6/19 7:32
Transfer the post to other applications Transfer


Re: (JANUS-UAE) Is today's nightly build for 68K running for you?

Joined:
2011/5/10 23:05
Group:
Member
Posts: 76
Offline
(I missed your log post..)

Hmm.. It is UAE autoconfig board init that breaks, strange.

Does it stop crashing if you remove all harddrives and all UAE expansions? (Not sure if it makes any difference with this UAE port)

Posted on: 6/19 2:49
Transfer the post to other applications Transfer


Re: (JANUS-UAE) Is today's nightly build for 68K running for you?

Joined:
2011/5/10 23:05
Group:
Member
Posts: 76
Offline
It is not just A4 but A5 and A6 that also have string data. (A6 should contain some library base in most situations)

It still sounds stack related if removal of one subroutine call changes the behavior. Original problem may have happened long time ago before it causes any visible differences..

(Adding serial logging does not need anything else than opening some file for writing, each time SERDAT is written, SERDAT low byte to file)

Posted on: 6/19 2:45
Transfer the post to other applications Transfer


Re: (JANUS-UAE) Is today's nightly build for 68K running for you?

Joined:
2011/5/10 23:05
Group:
Member
Posts: 76
Offline
Same reply here, serial log needed (should be easy to hack in emulator, if not already, to save all SERDAT writes to a file), both with and without JIT.

A4: 72617469 6E672053 79737465 is very suspicious part, ascii characters in address registers ("rating Syste"). This sounds like stack related problem. If it only happens when booting normally, narrow down it in program(s) in startup-sequence first.

Posted on: 6/15 23:15
Transfer the post to other applications Transfer


Re: current state of aros 68k..

Joined:
2011/5/10 23:05
Group:
Member
Posts: 76
Offline
Any aros-m68k problem: serial log (115200 8N1) must be included. (at least I won't do any debugging without serial log).


Posted on: 6/15 12:23
Transfer the post to other applications Transfer



 Top
(1) 2 3 4 ... 8 »




Search
Top Posters
1 paolone
paolone
4347
2 magorium
magorium
4095
3 phoenixkonsole
phoenixkonsole
3891
4 nikolaos
nikolaos
3677
5 deadwood
deadwood
2923
6 ncafferkey
ncafferkey
2635
7 mazze
mazze
2214
8 clusteruk
clusteruk
2105
9 Kalamatee
Kalamatee
2024
10 damocles
damocles
1789
© 2004-2017 AROS Exec