SqueakNOS

squeaknos@lists.squeakfoundation.org

87 discussions

Networking is comming to a theater close to you!
by Gerardo Richarte 11 Jul '06

11 Jul '06

hah! after fixing a very (very!) stupid bug, the networking code I wrote on the weekend just worked!!! now, if I do ping from outside vmware to the box inside vmware, a value changes from 128 to 3 (belonging to the card = 128, belonging to the host + first buffer + last buffer). So, we are now officially receiving packets, and if we inspect the packet, we see the source mac address, the destination mac address, and all the byte of the packet! happinesss!! gera

3 7

Memory, ideas?
by Gerardo Richarte 03 Jul '06

03 Jul '06

ok, the time has come. Hardware devices usually need to know some memory addresses, for example, in this case, the network device I'm implementing (AMD Lance, works on vmware) needs to know a pointer to a "configuartion block" which in turns has a pointer to a "receive ring" and a "transmit ring", which have pointers to several receive and transmit buffers. Problem: If we use, lets say, ByteArrays for buffers, and a garbage collect moves the buffer to a different memory position, the hardware may try to incorrectly access it in the old memory address. Ideas (3): . Use some kind of "Object Locking" to mark unmovable objects, so the garbage collector doesn't move our buffers around. Is this possible? How will this affect the performance of the garbage collector? what special care should we have if we do this? . Create a a map of the external memory (including, for example, where the Interrupt Descriptor Table and the Global Descriptor Table are located, what memory is available (mapped), where's the video memory, where in memory is loaded the interpreter (kernel), and where starts the .image). Then create the ExternalMemoryManager class in Squeak and use it to administer this external memory) . Probably the most interesting idea, collaborated by [uhm... shit, forgot your name and can't find your emails, I'll fill this space asap] Create a second ObjectMemory to take care of the biggest block of external memory available. I like this idea, however we either need to be able to lock some objects in this externalObjectMemory, or we seldom (never?) do a garbage collect in it. How do we do this? Craig? :-) What do you think? this is what has me stoped right now. I have all the information needed (and more) to implement simple networking (ethernet) support (for a single NIC now). If too long passes without solving this (1 week), I will implement it anyway, and latter see how we solve this. But we seriously need to solve it. all ideas are more than welcome! thanks! gera

1 2

Re: [SqNOS] PCI Bus Scan
by Johnathon Meichtry 01 Jul '06

01 Jul '06

Gera, As usual you were right, rebuilding the kernel & plugins solved the primitive issue. Nice! All I have to do now is sort out the data in vendorsAndDevicesZLIBData and the orderedcollection should be as expected. Johnathon ----- Original Message ----- From: "Gerardo Richarte" <gera(a)corest.com> To: "Johnathon Meichtry" <johnathon-dev(a)meichtry.org> Sent: Friday, June 30, 2006 10:29 PM Subject: Re: [SqNOS] PCI Bus Scan > the primitive error is due to a bug (already fixed) in > primitiveOutPort:dword: > > you need to either rebuild the kernel from the SqueakNOS plugin in > Monticello, (I think it's gera-6) or ask me and I'll send you a build. > > the CRC error is due to something weird in the method PCIDevice > class>>vendorsAndDevicesZLibData, which is a method returning a 65351 > bytes string who's hash is 16rBF5756D (PCIDevice vendorsAndDevicesZLibData > hash hex). > > This is a very big string, I know, but it's the compressed form of a 300kb > string, and I thought having it compressed was a good idea. Apparently, > storing this strings in source form is not safe. > > Lets try to find how to solve it. If you are interested in solving this > problem, I will send you all what's required to rebuild this string, so > you can work on what's the best why of storing this information (PCI > vendors and devices strings) in the image and in the sources. > > gera > > >

1 0

PCI Bus Scan
by Gerardo Richarte 30 Jun '06

30 Jun '06

hi, I just uploaded a screenshot showing SqueakNOS scanning the pci bus. click on the images in the swikit: http://minnow.cc.gatech.edu/squeak/1762 My next thing on the list is to implement vmware's ethernet card, I have three options (as you can see from the screenshot): AMD Am79C790, PCnet Lance PCI Ethernet Controller Intel 82545EM Gigabit Ethernet Controller VMWare Inc. VMXNET PCI Ethernet Adapter I'm going to work on the first (AMD), mainly because I think it's the most common of the three to exist outside vmware, and also, and probably more important, because GRUB has support for it, and I can get the minimum information to make it work out of GRUB's sources :-) any ideas, as usual, are welcome. gere PS: I'd love to know if anybody managed to recompile SqueakNOS, or to rebuild or reboot with a custom .image. Or if anybody is working on something, or anything else related

3 5

Operating Systems on the Rise; Embedded Systems Design (06/06) Vol. 19, No. 6, P. 53; Turley, Jim
by Klaus D. Witzel 24 Jun '06

24 Jun '06

List, this is perhaps interesting. /Klaus <quote> Embedded Systems Design's annual survey of embedded systems developers finds that around 28 percent of all embedded systems currently under development will have no operating system (OS), and this absence is especially prominent among developers of consumer, automotive, and industrial electronics; conversely, computer peripherals are most likely to feature OSes. A lack of need was the top reason provided by respondents for not including an OS, followed by the pressure an OS would put on the system's processor and/or RAM, cost, and difficulty of use. According to the poll, OSes are more likely to appear in products at larger companies than at smaller companies, while more experienced developers tend to use an OS in the current project. Of the respondents who do use an OS, 51 percent employ a commercial, off-the-shelf system; 21 percent use a proprietary, in-house, or internally developed OS; 16 percent use an open-source OS; and 11.8 percent use a commercial Linux distribution. From these findings, it can be surmised that the popularity of commercial OSes is growing dramatically, and that such OSes are taking the place of in-house OSes. Developers who opted for a commercially available OS said the choice of OS was most heavily influenced by the software staff, although the software manager also ranked highly as a decision maker. Top-ranking criteria for assessing OSes include real-time performance, processor support, software tool availability, a lack of royalties, cost, memory footprint, simplicity, and middleware availability. The survey indicates a precipitous drop in commercial and noncommercial distributions of open-source OS usage over the last year, with poor performance and/or real-time capability, support concerns, memory usage, legal ambiguousness, the state of development tools, and price cited as reasons for the decline. Over 36 percent of respondents said they would use a different OS in future projects than the one they currently use, while around 63 percent said they would keep using the same OS. </quote> http://www.embedded.com/showArticle.jhtml?articleID=187203732

1 0

16 bits display fixed
by Gerardo Richarte 17 Jun '06

17 Jun '06

Thanks to Tim Rowledge and Andreas Raab I just fixed the 16 bits display :-) I basically took a peep at the framebuffer implementation and copied the function mangling the bits. Here I'm attaching a new sqGlue.c which can handle 16 bits properly. as usual, please feedback everything you can gera

1 0

Compiling
by Gerardo Richarte 17 Jun '06

17 Jun '06

Hi, today we had a great time at #squeak (IRC). Craig was trying to compile a new VM for SqueakNOS (I guess he's trying some kind of spoon and SqueakNOS integration!), and he have some problems. So, if anybody wants to compile, this is how to do it (and how to apply some patches). All from linux, sorry. >From the CD (or .ISO image) get the copy trunk to your hard drive, and make it writable (if it's not): this is a working pseudoscript, read all before trying, although nothing bad can happen if you try: mount -o loop SqueakNOS-today.iso /cdrom cp -R /cdrom/trunk . chmod -R w+ trunk cd trunk/platforms/SqueakNOS cp /cdrom/Squeak*image boot/iso.template cp /cdrom/Squeak*changes boot/iso.template make distro this should leave SqueakNOS-today.tar.bz2 and SqueakNOS-today.zip in directory release. Now, "make distro" may fail, if so, or if you prefer try "make iso" instead, this'll leave SqueakNOS.iso in release and may be enough for you. If you have gcc < 4 you may experience problems related to strlen, memcpy and memcmp. Edit the Makefile in and change this line to: LIBCOBJS = strcpy.o strncmp.o strcmp.o strlen.o memcpy.o memcmp.o (add strlen.o memcpy.o memcmp.o to the end of the existing) try to make again, this should be it. let me know if you experience any problems gera

1 1

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

SqueakNOS