Click Here
home features news forums classifieds faqs links search
6065 members 
Amiga Q&A /  Free for All /  Emulation /  Gaming / (Latest Posts)
Login

Nickname

Password

Lost Password?

Don't have an account yet?
Register now!

Support Amigaworld.net
Your support is needed and is appreciated as Amigaworld.net is primarily dependent upon the support of its users.
Donate

Menu
Main sections
Home
Features
News
Forums
Classifieds
Links
Downloads
Extras
OS4 Zone
IRC Network
AmigaWorld Radio
Newsfeed
Top Members
Amiga Dealers
Information
About Us
FAQs
Advertise
Polls
Terms of Service
Search

IRC Channel
Server: irc.amigaworld.net
Ports: 1024,5555, 6665-6669
SSL port: 6697
Channel: #Amigaworld
Channel Policy and Guidelines

Who's Online
64 crawler(s) on-line.
 11 guest(s) on-line.
 0 member(s) on-line.



You are an anonymous user.
Register Now!
 Marcian:  8 mins ago
 thellier:  8 mins ago
 t0lkien:  15 mins ago
 agami:  33 mins ago
 Karlos:  33 mins ago
 Rob:  42 mins ago
 DC_Edge:  59 mins ago
 Musashi5150:  1 hr 8 mins ago
 AF-Domains.net:  1 hr 12 mins ago
 AMIGASYSTEM:  1 hr 17 mins ago

/  Forum Index
   /  Amiga OS4 Software
      /  AmigaOS4.1 OWB Development whilst we wait?
Register To Post

Goto page ( Previous Page 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | 32 | 33 | 34 | 35 | 36 | 37 | 38 | 39 | 40 | 41 | 42 | 43 | 44 | 45 | 46 | 47 | 48 | 49 | 50 | 51 | 52 | 53 | 54 | 55 Next Page )
PosterThread
Tomppeli 
Re: AmigaOS4.1 OWB Development whilst we wait?
Posted on 15-Feb-2011 20:38:05
#381 ]
Super Member
Joined: 18-Jun-2004
Posts: 1646
From: Home land of Santa, sauna, sisu and salmiakki

@Metalheart

I think all RAM should not come back because it's cached.

_________________
Rock lobster bit me. My Workbench has always preferences. X1000 + AmigaOS4.1 FE
"Anyone can build a fast CPU. The trick is to build a fast system." -Seymour Cray

 Status: Offline
Profile     Report this post  
OldFart 
Re: AmigaOS4.1 OWB Development whilst we wait?
Posted on 15-Feb-2011 20:51:43
#382 ]
Elite Member
Joined: 12-Sep-2004
Posts: 3009
From: Stad; en d'r is moar ain stad en da's Stad. Makkelk zat!

@TiredofLife

Quote:
That sounds quite extreme.

Started out this session with about 100 million bytes of free memory left. At this very moment only about a third is remaining, so it's just a matter of short time that my micro will be fully locked up again.

Quote:
Are any particular sites in use when this happens?

Well, I must admit that I visit some rather taxing sites off late: mostly AW.net, OS4Depot.net, UtilityBase, IntuitionBase, etc., etc. Quite heavy stuff.

On the good side: looking into the upload queue of OS4Depot I discovered OWB.lha...

OldFart

_________________
More then three levels of indigestion and you're scroomed!

 Status: Offline
Profile     Report this post  
TiredofLife 
Re: AmigaOS4.1 OWB Development whilst we wait?
Posted on 15-Feb-2011 21:05:30
#383 ]
Super Member
Joined: 6-Jul-2005
Posts: 1688
From: Here

@OldFart

That is weird, none of those sites should having that effect I would not have thought

Upgrading to the latest version is supposed to stop the lockups

Maybe that will partially help you

Last edited by TiredofLife on 15-Feb-2011 at 09:10 PM.

_________________
If your nose runs and your feet smell, you're upside down.

 Status: Offline
Profile     Report this post  
whose 
Re: AmigaOS4.1 OWB Development whilst we wait?
Posted on 15-Feb-2011 21:57:06
#384 ]
Cult Member
Joined: 21-Jun-2005
Posts: 893
From: Germany

Hmm... seems most people here resemble to the way memory allocations worked in OS3.x, right?

AFAIK this way changed massive, so that you cannot rely on the free memory display on Workbench anymore, or CLI display after a "Avail FLUSH", i.e. to catch memory leaks.

As Tompelii said, some things are "cached" (and I bet that this is true for these freakin´ sobjs, too), although these things are "freed" by the only program that used them.

The lockups may result from the pager, it is said to still have some bugs in certain states. I wouldnt wonder if the pager kicks in when OWB with a bunch of tabs open is running.

 Status: Offline
Profile     Report this post  
kas1e 
Re: AmigaOS4.1 OWB Development whilst we wait?
Posted on 15-Feb-2011 22:07:37
#385 ]
Elite Member
Joined: 11-Jan-2004
Posts: 3535
From: Russia

@whose
Quote:

As Tompelii said, some things are "cached" (and I bet that this is true for these freakin´ sobjs, too), although these things are "freed" by the only program that used them.


If i remember right, Joerg's owb not use sobjes anymore, and all done statically in binary (but can be wrong here).

_________________
Join us to improve dopus5!
zerohero's mirror of os4/os3 crosscompiler suites

 Status: Offline
Profile     Report this post  
whose 
Re: AmigaOS4.1 OWB Development whilst we wait?
Posted on 15-Feb-2011 22:24:25
#386 ]
Cult Member
Joined: 21-Jun-2005
Posts: 893
From: Germany

@kas1e

Quote:
If i remember right, Joerg's owb not use sobjes anymore, and all done statically in binary (but can be wrong here).


Thats why I wrote "too", because one might think of another bunch of things, used, freed, but still in memory (for caching reasons).

RA libs, for example. Or normal libraries. Or sobjs used by statically linked support libraries.

 Status: Offline
Profile     Report this post  
TiredofLife 
Re: AmigaOS4.1 OWB Development whilst we wait?
Posted on 15-Feb-2011 22:44:45
#387 ]
Super Member
Joined: 6-Jul-2005
Posts: 1688
From: Here

@all

Wunderground no longer seems to cause the memory leak.
Can someone provide a link to a site that does?

Cheers

_________________
If your nose runs and your feet smell, you're upside down.

 Status: Offline
Profile     Report this post  
tonyw 
Re: AmigaOS4.1 OWB Development whilst we wait?
Posted on 15-Feb-2011 23:16:23
#388 ]
Elite Member
Joined: 8-Mar-2003
Posts: 3229
From: Sydney (of course)

Whatever memory OWB (or any other properly-behaved program) uses, it will be removed from the available memory when it is allocated. If there isn't enough "available" memory, the kernel will initiate a cleanup and perhaps page out some stuff to get more. Not all programs use "pageable" memory, so that process might fail. Well-behaved, modern programs like OWB will accept the failure gracefully, but older or less well-written programs might crash when faced with an AllocMem() that failed. Some programs just assume that the AllocMem() will work and proceed with the result, even though it may be null. Of course, they will crash and might even lock up the system.

Of course, it is possible that the program asking for memory is something really important, something that can't continue without it. In that case, the system will lock up if that program can't deal with the failure.

Since the memory that is allocated can come from anywhere, you may not see the "Avail" memory change by the same amount as the program asked for. In some cases it can even increase, if the cleanup process delivers more than was asked for.

Wherever it came from, all that memory will be returned to the system when that program exits or frees the memory. You might expect the "Avail" to increase by the amount that the program releases, but don't bet on it. All or part of the freed memory might just be marked as "unused" and not necessarily cleaned and added to the "Avail" memory. Some may even be given to another program that has been waiting for it. Some might be used to page back in some other program that was paged out earlier.

That's why the Frieden brothers keep repeating (usually to deaf ears) that the "Avail" figures don't mean a thing. "Avail" memory only shows the amount of memory that is clean and ready for use (probably has never been used). There is usually a lot more that can be used if necessary.

_________________
cheers
tony

Hyperion Support Forum: http://forum.hyperion-entertainment.biz/index.php

 Status: Offline
Profile     Report this post  
TiredofLife 
Re: AmigaOS4.1 OWB Development whilst we wait?
Posted on 15-Feb-2011 23:24:28
#389 ]
Super Member
Joined: 6-Jul-2005
Posts: 1688
From: Here

@tonyw

I totally accept what you say and have seen this explanation from the Frieden brothers.
This explains the difference in memory showing before and after, opening and closing a tab.

What I would say though is two systems running the same software, I would expect to see simlar results.

Looking at some of the postings above, this doesn't appear to be the case.

Last edited by TiredofLife on 15-Feb-2011 at 11:25 PM.

_________________
If your nose runs and your feet smell, you're upside down.

 Status: Offline
Profile     Report this post  
Tomas 
Re: AmigaOS4.1 OWB Development whilst we wait?
Posted on 15-Feb-2011 23:28:02
#390 ]
Elite Member
Joined: 25-Jul-2003
Posts: 4286
From: Unknown

@whose

Quote:

whose wrote:
Hmm... seems most people here resemble to the way memory allocations worked in OS3.x, right?

AFAIK this way changed massive, so that you cannot rely on the free memory display on Workbench anymore, or CLI display after a "Avail FLUSH", i.e. to catch memory leaks.

As Tompelii said, some things are "cached" (and I bet that this is true for these freakin´ sobjs, too), although these things are "freed" by the only program that used them.

The lockups may result from the pager, it is said to still have some bugs in certain states. I wouldnt wonder if the pager kicks in when OWB with a bunch of tabs open is running.

Yeah i know available memory is unreliable, but when OWB does eat up memory it is definitely not freed for other tasks, as anything leads to swapping when this happens even after OWB has been closed.

The only way to make system usable again is a reboot..

 Status: Offline
Profile     Report this post  
whose 
Re: AmigaOS4.1 OWB Development whilst we wait?
Posted on 16-Feb-2011 0:34:03
#391 ]
Cult Member
Joined: 21-Jun-2005
Posts: 893
From: Germany

@Tomas

Maybe you (and others) should try C:memstat then (and prep yourself with some patience, important!) to see, how "deep" the pager goes into this game.

It is quite possible that OWB eats up all your memory, then the pager kicks in, and your system needs some time to resemble the memory lists after OWB is closed. Maybe some seconds, maybe a minute, or even more.

As I said, there is saying that there are some bugs in the paging system. Quite possible that this is one of them.

Nonetheless an OWB bug is possible, too.

Btw., I say this because I have real hard problems to track down memory leaks with OS4.x, using "simple" application software but no debugging tools, so it is a bit strange for me to read about memory leaks here

 Status: Offline
Profile     Report this post  
TiredofLife 
Re: AmigaOS4.1 OWB Development whilst we wait?
Posted on 16-Feb-2011 1:04:21
#392 ]
Super Member
Joined: 6-Jul-2005
Posts: 1688
From: Here

@whose

Just for the sake of completeness, I thought I'd open up tabs and sites until the system used paged memory.

Did get a DSI error but will do some more testing.

Other than that, there seemed to be no other issues.
Memory was released back after I shut OWB down.

_________________
If your nose runs and your feet smell, you're upside down.

 Status: Offline
Profile     Report this post  
TiredofLife 
Re: AmigaOS4.1 OWB Development whilst we wait?
Posted on 16-Feb-2011 1:27:43
#393 ]
Super Member
Joined: 6-Jul-2005
Posts: 1688
From: Here

@TiredofLife

2nd attempt, really pushed it this time.
No error and memory returned.

So it seems the memory not returning is either specific to a site(s) or some configurations.

For those having a lockup, is your swap partition set up correctly?
Seem to remember that causing a problem for myself in the past.

_________________
If your nose runs and your feet smell, you're upside down.

 Status: Offline
Profile     Report this post  
Deniil715 
Re: AmigaOS4.1 OWB Development whilst we wait?
Posted on 16-Feb-2011 9:29:24
#394 ]
Elite Member
Joined: 14-May-2003
Posts: 4226
From: Sweden

@tonyw

Just drop it! OWB has a mem leak bug.

It did happen for me again yesterday on csn.se where I logged in to pay my student loads. I had to surf a bit to find the place and (try) to do my payment (which required Java so I had to abort ) but it had eated 200MB RAM. When I exited OWB I expected to see about 150-180MB being given back to "Avail", which is normal when the leak does not happen, the rest being cached as you say. But it didn't happen. I was left with 80MB "avail" and subsequent programs, such as gcc did NOT get any share of this memory but was forced to swap. It was NOT free. OWB leaks!

BTW. New "modern" program are much more sloppy in memory management IMO because they come from Linux or windows where eventual leaks will be cleaned by the kernel when the process end. malloc is also never expected to return NULL. Older Amiga programs are better at this because they were developed in an environment with low memory, no swap and no kernel garbage collection.

@TiredofLife

It's definately related to some sites. Surfing "proper"(?) sites you can do for ages and puch it till it swappes. It will still free up everything and what is not free but cached will still be available for other programs.

_________________
- Don't get fooled by my avatar, I'm not like that (anymore, mostly... maybe only sometimes)
> Amiga Classic and OS4 developer for OnyxSoft.

 Status: Offline
Profile     Report this post  
NutsAboutAmiga 
Re: AmigaOS4.1 OWB Development whilst we wait?
Posted on 16-Feb-2011 9:45:35
#395 ]
Elite Member
Joined: 9-Jun-2004
Posts: 12343
From: Norway

@Deniil715

AmigaOS4 does not flush memory anymore, it’s impossible to know, it keeps previous allocated blocks in case an application needs the same amount again, actual amount of free memory is never reported.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

 Status: Offline
Profile     Report this post  
Deniil715 
Re: AmigaOS4.1 OWB Development whilst we wait?
Posted on 16-Feb-2011 9:50:32
#396 ]
Elite Member
Joined: 14-May-2003
Posts: 4226
From: Sweden

@all

I have made a small but very fast memory monitor a looong time ago that shows the "avail" memory in "realtime". It gives you a nice overview: http://aminet.net/util/wb/MemWatch.lha

A unique feature is that the more frequent the free memory changes the faster it updates. Default is 5 times a seconds IIRC. Having this running you can see how memory is freed and consumed when the pager kicks in in a way that any shell command or the WB title cannot.

_________________
- Don't get fooled by my avatar, I'm not like that (anymore, mostly... maybe only sometimes)
> Amiga Classic and OS4 developer for OnyxSoft.

 Status: Offline
Profile     Report this post  
Tomppeli 
Re: AmigaOS4.1 OWB Development whilst we wait?
Posted on 16-Feb-2011 10:36:51
#397 ]
Super Member
Joined: 18-Jun-2004
Posts: 1646
From: Home land of Santa, sauna, sisu and salmiakki

@Deniil715

Quote:
Just drop it! OWB has a mem leak bug.

Why I reacted after Metalheart's post is when people are watching WB titlebar and reporting memory leaks all such reports are not necessarily real memory leaks. But that does not mean any application could not have a real memory leak. So any of this talk wasn't meant to say anything against anybody. So you can have a memory leak in any app but every reported memory leak is not necessarily real memory leak.

@tonyw
Isn't all memory allocations resource tracked so if any app leaks memory then when you close that app all memory, which that app allocated, is freed (or marked as unused at least) ? Also if app leaks memory it means it allocates memory but doesn't use it ever (or uses it only once and then forgets it or something like that). So such unused but allocated memory should be eventually paged out when system runs out of free memory and doesn't harm anybody.

_________________
Rock lobster bit me. My Workbench has always preferences. X1000 + AmigaOS4.1 FE
"Anyone can build a fast CPU. The trick is to build a fast system." -Seymour Cray

 Status: Offline
Profile     Report this post  
NutsAboutAmiga 
Re: AmigaOS4.1 OWB Development whilst we wait?
Posted on 16-Feb-2011 10:40:40
#398 ]
Elite Member
Joined: 9-Jun-2004
Posts: 12343
From: Norway

@Deniil715

Nice but how does it work?

You can’t patch exec main interface from a 68k program, and stuff has changed in AmigaOS, so can you trust the numbers you are getting?

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

 Status: Offline
Profile     Report this post  
Deniil715 
Re: AmigaOS4.1 OWB Development whilst we wait?
Posted on 16-Feb-2011 10:46:41
#399 ]
Elite Member
Joined: 14-May-2003
Posts: 4226
From: Sweden

@NutsAboutAmiga

Quote:
AmigaOS4 does not flush memory anymore, it’s impossible to know, it keeps previous allocated blocks in case an application needs the same amount again, actual amount of free memory is never reported.


It flushes most memory, which is very obvious using an old-fashion memory meter, such as mine here above. The rest is also free (but cached) and used by applications, which can be seen when you start an app and the memory hardly moves at all - it only use cached memory.

This however is very much not the case when OWB has eaten the memory. No other program is allowed to use this "supposadly cached" memory. It is not cached, it is leaked by OWB. If the system truly cached these 200MB and didn't flush any of it, then wouldn't subsequent programs seemingly use no memory at all because they would use this cached memory instead? If so, why doesn't this happen? Why is any new allocation just using more memory rather than this "cache" if OWB now is so completely perfect and "modern" as you ppl seem to believe.

_________________
- Don't get fooled by my avatar, I'm not like that (anymore, mostly... maybe only sometimes)
> Amiga Classic and OS4 developer for OnyxSoft.

 Status: Offline
Profile     Report this post  
whose 
Re: AmigaOS4.1 OWB Development whilst we wait?
Posted on 16-Feb-2011 11:01:41
#400 ]
Cult Member
Joined: 21-Jun-2005
Posts: 893
From: Germany

@Tomppeli

Quote:
@tonyw Isn't all memory allocations resource tracked so if any app leaks memory then when you close that app all memory, which that app allocated, is freed (or marked as unused at least) ? Also if app leaks memory it means it allocates memory but doesn't use it ever (or uses it only once and then forgets it or something like that). So such unused but allocated memory should be eventually paged out when system runs out of free memory and doesn't harm anybody.


Ok, Im not tonyw, but I think I can answer this question to the best of my knowledge, too.

The resource tracking in OS4.x is actually organized in a cooperative way. Programs that allocate memory in the "simple" way wont get resource tracking for it. You have to use dedicated OS functions for this (AllocSysObject()), so that the actual resource tracking can work. As you see, this is restricted to "SysObjects" atm., memory pools and such are said to be not tracked (well, actually I cant find any information regarding tracking, reading the AutoDocs for the standard memory allocation functions, that is).

I dont think that e.g. the libraries used for OWB supports such a mechanism. Excuse me for this, but *nix style coding is dumb, most of the time: "Just grab memory, dont care for attributes. The kernel will do it right. No NULL pointers possible". In the light of this, its quite improbable that e.g. OWB uses the cooperative RT memory allocation scheme all over its code.

Code written with OS4.x in mind should adhere this scheme. But it mustnt.

For the unused memory thing: How should the OS know, if this memory is used or not? In an ideal world, the OS should not need to know what the application is doing with requested memory. It simply manages the memory pool, and cares for out-of-memory situations in an appropriate way (returning NULL or swapping). If the OS should indeed know whats going on within requested memory pools for a certain time, it would have to look after every memory access, which would be a rather hard job for the main CPU...

 Status: Offline
Profile     Report this post  
Goto page ( Previous Page 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | 32 | 33 | 34 | 35 | 36 | 37 | 38 | 39 | 40 | 41 | 42 | 43 | 44 | 45 | 46 | 47 | 48 | 49 | 50 | 51 | 52 | 53 | 54 | 55 Next Page )

[ home ][ about us ][ privacy ] [ forums ][ classifieds ] [ links ][ news archive ] [ link to us ][ user account ]
Copyright (C) 2000 - 2019 Amigaworld.net.
Amigaworld.net was originally founded by David Doyle