users August 2020

users@lists.genode.org

14 participants
17 discussions

Start a nNew thread

Network issue
by Dima Sokolov 01 Feb '21

01 Feb '21

3 3

mmap implementation for ANON mapping with desired address
by Alexander Tormasov 21 Dec '20

21 Dec '20

I am trying to implement not-yet supported (as mentioned by nfeske@ in mailing list) option for mmap() related to anonymous mapping with pre-defined (desired) address. Like libc function with desired address out_addr. void *Libc::Mem_alloc_impl::alloc_at(void *out_addr, size_t size) I use a couple of useful function in avl allocator. Anyway, I found that straight - forward implementation do not work because of allocation gaps filled together with main allocation. In particular, if we want to allocate e.g. 0x10000 from address 0x500000 then we typically allocate 0x13000 bytes, including ones for avl metadata/etc stored in the same space, and receive virtual address range used, e.g. 0x500000 - 0x513000: /* * Calculate block size of needed backing store. The block must hold the * requested 'size' with the requested alignment, a new Dataspace structure * and space for AVL-node slab blocks if the allocation above failed. * Finally, we align the size to a 4K page. */ size_t request_size = size + max((1 << align_log2), 1024) + Allocator_avl::slab_block_size() + sizeof(Dataspace); Software I used (golang runtime) sometimes try to expand previous allocation, e.g. it can request allocation of 0x8000 bytes before previous range, like 0x4f8000 - 0x500000. In that case I can’t perform the operation - because it will try to allocate 0xa000 bytes from desired address 0x4f8000 - and it obviously fail due to Range_conflict. So, I split allocation to 2 parts - one for this gap by standard alloc() and random placement and second one by my own alloc_at() with size exactly correspond to the requested one. Implementation available as a patch to 20.05 at https://github.com/tor-m6/genode/commit/78fc751bea8acc2b515e04bec9f3a883461… Problem I see here is a free operation: it just free the block with requested address, and do not touch others, gap-related blocks. For me it is not obvious that they need to be free, because in the code (e.g. by unmap()) I do not see explicit free of other allocated structures, in particular, dataspace extension or related block metadata. Question: should I make additional efforts to trace and delete something else - except what standard free does here: /* forward request to our local allocator */ _alloc.free(addr); Like for dataspace structure or address ranges for second «auxiliary» allocation? PS Another potential problem is a probability to overlap of first «random address» allocation with desired address. This happens with first attempt to implement allocator - it allocates address from desired range, and immediately put here 0x18 bytes of Dataspace structure, and then fail because requested address is already busy by this structure! Anyway, I consider such a probability relatively small with proposed approach and ignore it (may be I am wrong)

4 23

Announcement: Genode OS Framework version 20.08 released
by Norman Feske 28 Aug '20

28 Aug '20

I'm happy to announce today's release of Genode 20.08. The two most prominent topics of the new version are our port of the Chromium web engine and the redesign of the low-level GUI stack. With the Chromium engine, we expose Genode to an extremely heavy-weight workload, which stresses the framework like never before on many levels, from the quality of our POSIX runtime, the ability to bridge 3rd-party build systems with Genode, to the performance and robustness of our native protocol stacks (TCP/IP, graphics, file-system access). So the collateral improvements while porting this behemoth of software are valuable for the users of the framework at large. The rework of the framework's low-level GUI stack is the result of a design and planning phase of over two years. With the architectural change conducted during this release cycle, we clear the path for the ability to swap-out graphics and input drivers on the fly without reboot, and to realize features like screen capturing and remote desktop scenarios in a privacy-protecting way. Besides these highlights, the release comes with plentiful improvements ranging from hardware-support (i.MX8), over the modularisation of CBE block encryptor, to CPU-affinity management for Sculpt OS. A quick summary of the most visible topics: - Redesigned low-level GUI stack - Removal of drivers from the trusted computing base - Consistent use of 32 bits per pixel - Improved Qt5 integration and work flows - Chromium engine and Falkon web browser - Pluggable crypto and trust anchor for the CBE block encrypter - Device drivers and platforms - ARM Trusted Firmware access - Power and clock management for i.MX 8M Quad SoC - New block server backed by VFS content - New port of SDL2 library - Interactive CPU-affinity management for Sculpt OS You can find all the details in the release documentation at https://genode.org/documentation/release-notes/20.08 Enjoy! Norman -- Dr.-Ing. Norman Feske Genode Labs https://www.genode-labs.com · https://genode.org Genode Labs GmbH · Amtsgericht Dresden · HRB 28424 · Sitz Dresden Geschäftsführer: Dr.-Ing. Norman Feske, Christian Helmuth

2 1

Unable to read from socket
by Daly, Connor G. 24 Aug '20

24 Aug '20

Hi all, Quick thanks to Alex for helping me get started a couple weeks ago. I'm trying to port/rewrite a C++ program I wrote to Genode and am running into a networking issue. The program is a server and accepts TCP connections from clients that do not wait for a response. I'm sending my data via a python 3.7 script that works with the Linux native version of the application and is essentially: import socket sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM) sock.connect(('127.0.0.1', 5555)) sock.sendall(b"test\r\n") sock.close() Basically, I've been unable to read from my listen sockets. I'm able to bind, listen for connections, accept an incoming connection, and call socket read(), but then it just hangs. I saw something in the 18.08 release notes that made me think LwIP might only support asynchronous mode, so I tried nonblocking sockets. Those no longer hang, but they also don't read any data. Next, I tried using tcp_terminal and modifying the tcp_terminal_echo program to fit my needs. This program also can't read any of the data I send it. It accepts the incoming connection, but never logs that any bytes are read. However, TCP terminal does accept telnet connections like in the example program. I then discovered that my synchronous socket reading would actually work for a telnet connection as well. Any idea what is going on here? For reference, my run script is essentially the same as tcp_terminal.run. Failing that, does anyone know what I would need to do to modify tcp_terminal/tcp_terminal_echo to make it accept regular tcp socket connections and not just telnet? Thanks so much! -Connor

1 0

Base-hw scheduler
by Chris Hofer 24 Aug '20

24 Aug '20

Hi everyone, I'm having some troubles in understanding how the concept of 'fills' and 'claims' of base-hw's scheduler [1] is expressed in the Xml syntax of Init's configuration. My concrete questions are: * how does the syntax indicate, whether the threads of a component are treated as fills or claims? * do priorities affect the scheduling of fills? The background of the second question is, that I'm confused by the priority assignment in the run-script vmm_arm.run. I expected all components to be fills and scheduled in a round-robin fashion. What is the reason for assigning the priority -1 to the Vmm? Best, Chris Hofer [1] 7.7.7 of Genode Foundations

2 1

Issue in booting on Lenovo x230
by Parfait Tokponnon 21 Aug '20

21 Aug '20

Hello, Please, may you help me figure out what makes the booting process hang on Lenovo x230? The booting process just hangs with no explicit error on the serial console. Freshly cloned git repository, This happens with both demo and seoul-net scenarios Log files are attached, Genode version: 20.05 Regards, -- Parfait T.

2 1

difference in genode memory allocation for NOVA and seL4
by Alexander Tormasov 19 Aug '20

19 Aug '20

Hello! I am trying to implement allocate_at() Libc:: function, which allocate memory region in given (desired) address. To implement I need to extend dataspace (as I understand, this is physical memory) at appropriate address, and later attach it to the local allocator. I found that memory allocation code from base/include/region_map/region_map.h /** * Shortcut for attaching a dataspace at a predefined local address */ Local_addr attach_at(Dataspace_capability ds, addr_t local_addr, size_t size = 0, off_t offset = 0) { return attach(ds, size, offset, true, local_addr); } does work for NOVA, but do not work for seL4 kernel for x86_64. This happens when I try to allocate 64k from pre-defined addresses like (try in order) 0xc000000000, 0x1c000000000, 0x2c000000000/etc… I create function int Libc::Mem_alloc_impl::Dataspace_pool::expand_at(size_t size, void * ptr, Range_allocator *alloc) and inside it, in my call to try { new_ds_cap = _ram->alloc(size); local_addr = _region_map->attach_at(new_ds_cap, addr_t(ptr), size); } in seL4 it catch exception catch (Region_map::Region_conflict) { while same calls in NOVA works ok. Problem happens in second call to attach_at(), new_ds_cap is ok. Are there significant difference in the memory allocation process (for particular addresses ranges) between NOVA and seL4? Thanks! Alexander

2 1

Re: Network issue
by Dima Sokolov 15 Aug '20

15 Aug '20

1 0

Genode's 12th anniversary
by Norman Feske 14 Aug '20

14 Aug '20

Dear Genode community, time flies by. We are now in the 12th year of Genode, following the very same vision we had when we founded the project. But in contrast to the early days, today Genode has long outgrown the status of an obscure research project and has gained a strong product focus. During the past year, this focus led us to embrace 64-bit ARM and NXP i.MX in particular, profound consolidation and optimization of the framework, the novel user interface of Sculpt OS, and the streamlining of work flows. Proud of these advances, we want to celebrate the Genode anniversary together with the community and friends around Genode by holding our traditional informal barbecue at the banks of the river Elbe: Eisenberger Straße, Dresden/Germany September 4, 2020 Google Maps: https://goo.gl/maps/C0luj If you are planning to attend, please let us know by writing a short note to: info(a)genode-labs.com Cheers Norman -- Dr.-Ing. Norman Feske Genode Labs https://www.genode-labs.com · https://genode.org Genode Labs GmbH · Amtsgericht Dresden · HRB 28424 · Sitz Dresden Geschäftsführer: Dr.-Ing. Norman Feske, Christian Helmuth

1 0

Re: Is Genode good for the Desktop?
by ttcoder＠netcourrier.com 11 Aug '20

11 Aug '20

Is Genode good for the Desktop and is it based on seL4? What does OS framework mean? Hi, Assuming you're looking for a personal take, rather than the official info available from genode.org (:-) : In my personal view, what makes Genode the "go-to" OS toolkit (or OS "framework" : a collection of modules that can work together or separately) is several things, but especially its modularity. A couple years back I used to be a happy camper as a user of a Be, Inc-derived OS, thinking if was "most" modular, "this is the place to be", rationalizing whatever problem I ran into as 'unavoidable with modern computing' and so on... Then I found out about Genode. It gives a new meaning to "lego like", assembling pieces together. In comparison, it makes my daily OS use look like child's play, a monolithic black box of crude, un-reusable, hard to maintain code. (talking of the internals, compared to the Genode internals, the APIs are another can of worms). The breakthrough for me was reading the PDF (Genode foundations.pdf). It takes some effort to comprehend it, it's not light reading... But if you really get "into" it, wow. As you read through it you'll keep thinking "why does so-and-so operating system not use the same mechanism as Genode-based OSes do ?". I believe seL4 is one of the micro-kernels that may be used by Genode, that info should be available on genode.org As to Desktop usage, if you ask around you'll probably get opinions over the whole spectrum, with many on the 'hopeful' scale, similar to that of the Genode Labs team : they reported a good while back that Sculpt-OS (their desktop OS based on Genode) has become their "daily driver" OS, which is a good milestone to have achieved. However, they still rely on VirtualBox for some things yet, so there are more milestones to reach. Some I remember being mentionned include self-hosting, i.e. ability for Sculpt OS to compile/build itself, having a native email client rather than relying on one embedded in VirtualBox, etc. Make yourself comfy and stay around a while, you'll see the people here are a quiet bunch, mainly reporting on their code once they have a deliverable. Cedric

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

users August 2020