users

users@lists.genode.org

6 participants
2306 discussions

Linker Issues in base-hw for an Armv8 Implementation
by Bob Stewart 18 Jul '18

18 Jul '18

Hi, I realize my issues are far from the current mostly-Sculpt roadmap, but perhaps someone can throw some light on my problems...at least I'm dealing with current hardware devoid of parallel ports :o I've been carefully creating a build structure in base-hw that will support armv8 and at least AArch64 over the past several months. I'm at the point where I have made all the architectural additions to support armv8 and the arm64 assembler which includes support for the assembler instruction changes from armv7, execution modes, and exception levels (only EL0 and EL1, initially). Everything builds for a cortex A-53 platform but fails in the final linking steps in two ways as described below. But first, my build tools are a copy of the Linaro toolchain, version 6.4.1, which used binutils 2.27, and my CROSS_DEV_PREFIX points to its location. I'm building in Genode 18.02 1. The final linking step fails initially with the "_impure_ptr" global not being resolved (see the build output in build_impure.log). Typically this global is supplied via libc, but Genode's special libc does not contain it. This global is a pointer to a structure that contains "errno" from failed C-code execution, along with some concurrency data for threading, if remember correctly. The pointer is only used in vterminate.cc which comes from the toolchain. In the currently used Genode toolchain the library supc++.o contains vterminate.o which is stripped, so the linker would not complain about the pointer. Is this by design because the process is about to die anyway when vterminate is called? To temporarily work around this problem I created a dummy void * _impure_ptr global in base/lib/cxx/misc.cc. The linker now continues and eventually completes. 2. The second problem is the size of the executable image files that the linker produces, several gigabytes! The actual memory image of the executable appears correct based on a dump of the section header but the header information shows that the File Offsets to the sections appears to be based off the link address. For example, in the creation of the boot image, build_core creates the binary log.bootstrap, the dump of which is: objdump -h ../build/dragonboard/var/run/log.bootstrap |more ../build/dragonboard/var/run/log.bootstrap: file format elf64-little Sections: Idx Name Size VMA LMA File off Algn 0 .note.gnu.build-id 00000024 0000000000000000 0000000000000000 81042000 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 1 .text 0003b840 0000000081000000 0000000081000000 81001000 2**4 CONTENTS, ALLOC, LOAD, READONLY, CODE 2 .data 00000718 000000008103c000 000000008103c000 8103d000 2**3 CONTENTS, ALLOC, LOAD, DATA 3 .got 000000e0 000000008103c718 000000008103c718 8103d718 2**3 CONTENTS, ALLOC, LOAD, DATA 4 .got.plt 00000018 000000008103c7f8 000000008103c7f8 8103d7f8 2**3 CONTENTS, ALLOC, LOAD, DATA 5 .eh_frame 00003084 000000008103c810 000000008103c810 8103d810 2**3 CONTENTS, ALLOC, LOAD, READONLY, DATA 6 .gcc_except_table 00000968 000000008103f894 000000008103f894 81040894 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 7 .bss 00008f09 0000000081040200 0000000081040200 810411fc 2**3 ALLOC It would appear that all the "File off" values are the expected value plus the bootstrap_link_address which is 0x81000000. Additionally, there is no stack unwind sections present, that is .ARM.exid/extab. The linker script being used is base/src/ld/genode.ld, as it should be. I don't have sufficient knowledge about linker functioning to begin to understand how the File offset values are determined or how the c++ stack unwind section is added, or not. Any suggestions would be very helpful. Thanks, Bob Stewart

2 2

difference between lwip and lxip
by Guido Witmond 16 Jul '18

16 Jul '18

Hi All, I've been playing with lwip and lxip and I noticed a discrepancy between them. I made a tester-application [1] to pinpoint these differences. The application gets a DHCP-address, opens a berkeley socket, binds to it, sets listen mode and accepts an incoming request. Then it sends a simple string back and waits for a new request. With lwip it works as expected. With lxip it runs until the accept-step. There it times out after roughly 10 seconds with errno 21 (EINVAL) Invalid argument. However if I connect to it with a client before the timeout hits, it serves a reply and resets the timer. I can keep the application alive for as long as I keep requests coming in. I understood that lxip was plug-in compatible with lwip but somehow lxip does not have the patience that lwip has. Is this a bug in lxip or is there a missing dependency for lxip? I've added the make run/tester_l[wx]ip log files in the repo. The lxip one has scary warning about incompatible type conversions. Cheers, Guido. 1: https://github.com/gwitmond/tester

2 2

scupt-tc real hardware: 0/3
by Paul Dufresne 16 Jul '18

16 Jul '18

Oh well, I am like a returning Genode user, for almost ... one and an half year I guess. I am glad to see that there is now an image (a USB one... I would guess it would not work on a DVD?). I have downloaded it yesterday, and tried it on 3 computers (2 almost the same: Dell Optiplex 745, not sure the other is a 745 too). One of the two is my own computer, and I have try many BIOS config, but it does always do the same: show the Logo, screen become black for about 7 secs... then it reboots and loop. On a laptop I own but keep at my friend location, it would show the logo, screen would become black, and it would hang there. I think I remembered that debug messages was sent on serial port, but sadly, I have dump my serial null-modem, thinking that it was outdated. I don't think the Dell Optiplex support AMT? ... that technology that is supposed to keep a log in memory that would have been sent to serial port. I was thinking about the fact, that one should expect things not to work rather than to work. Was thinking about what I would probably design. I was thinking that an hard disk partition could be used as a log partition. I was thinking about giving step numbers while booting. Was thinking how screen is small, not so many lines, and that writing only error or warning messages, avoiding went good messages is a good default. Was thinking about boot options: -Show only warning and error messages -Show text description of all steps, one by line + one line by error messages -Show all steps numbers done, just separated by a space, giving only error messages -Write log to disk partition (USB drivers is hard to get right on real hardware, from my experience with ReactOS), with warning and errors only -Write log to disk partition with all messages, even info messages Sure, the step where you go from text mode to graphic mode, is a bit tricky... guess in "debug mode", I would do like when changing screen resolutions: change resolution Show "Are you happy with this screen?" Wait 5 secs If not yes, return to text mode, and ... go to a terminal. But of course, I am more thinking than doing.

7 31

How To Query Remote Repository
by John J. Karcher 12 Jul '18

12 Jul '18

Hello, Apologies in advance if this is a stupid question, but is there a way to query a remote repository for what packages are available (and what versions)? I couldn't find any way to do this, in either the development tools or the Sculpt runtime. Thanks! -- John J. Karcher devuser(a)alternateapproach.com

2 1

Service denied when server is updated
by Stefan Thöni 10 Jul '18

10 Jul '18

Hello Genodians, I observed that when a runtime configuration with a app already present (downloaded and unpacked) accessing a service on new server (version) is deployed, the app gets a Service_denied exception while the new server (version) is being downloaded, verified or unpacked. Wouldn't it be better to start all components in the runtime configuration only when all possible downloading is complete? Otherwise we either need to make components tolerant to Service_denied exceptions or detect such error conditions and restart the runtime entirely. Cheers Stefan -- Freundliche Grüsse Stefan Thöni Senior Security Architect +41 79 610 64 95 gapfruit AG Baarerstrasse 135 6300 Zug https://gapfruit.com

2 3

Sculpt built from current master - availability of packages
by Tomasz Gajewski 07 Jul '18

07 Jul '18

Hi, after release Sculpt TC I've made another attempt to use it. I've downloaded official image and attempted to boot on my computers. On one it rebooted soon after splash screen probably due to issue with IOMMU investigated some time ago in issue 2700 to which I hope to get back to in near future. On another one it booted but failed to initiate ethernet controller. I've checked that it is an intel i219v-2 which is handled by dde_ipxe in newer version than the one in genode I've made some changes that resulted in commit: https://github.com/tomga/genode/commit/7ce2291796c2752974c182e42aebfc8d8d31… that allowed to get ip address from my router so it that networking is ok. This commit contains a simple change to intel.c patch and three changes to hashes that were needed to build sculpt after this change. Is it normal that each change needs update of hashes of all dependent components? Do I correctly understand that this will force many changes of hashes in repository? Next I attempted to perform next step from Sculpt TC manual and uncommented part from config/deploy about fonts_fs and it attempted to download some packages from depot.genode.org. Files that were attempted to be downloaded do not exist on this server. There are similar but with different timestamps e.g. download of .../fonts_fs/2018-07-02.tar.xz was attempted but the latest one in this location is 2018-06-12.tar.xz This date, if I correctly understand, comes from: b485caf33c (Christian Helmuth 2018-06-25 12:51:03 +0200 1) 2018-07-02 a31d98a7dfb8cd7913ec38b0dfe080ae3382f8b6 So the second question is: should I assume that if this "version" is in the git repository it should be available on depot.genode.org? If not, what would you suggest to proceed? Regards Tomasz Gajewski

2 3

Error after running run/printf. I built hw_imx6
by Prakash A 06 Jul '18

06 Jul '18

Hi, I am getting error while making a run/printf and my build platform is hw_imx6 and genode version 13.02. *$ make run/printf* .... . . .Skip target core/vea9x4/trustzone because it requires platform_vea9x4 trustzone Library platform Library cxx Library kernel_interface Library base-common Library perf_counter Program core/imx6/no_trustzone/core ASSEMBLE boot_modules.o LINK core make[1]: Leaving directory '/home/prakash-a/secureos/tool/Genode_imx6' SHOULD I BUILD UIMAGE BUILD UIMAGE *spawn qemu-system-arm -nographic -m 64 -kernel var/run/printf/image.elf* *qemu-system-arm: -kernel var/run/printf/image.elf: No machine specified, and there is no default* *Use -machine help to list supported machines* *Error: Spawned process died unexpectedly* *Makefile:237: recipe for target 'run/printf' failed* *make: *** [run/printf] Error 253* I dont know what is the machine for imx6and how to include her. Regards, Prakash A

2 1

sel4: Need to have python module 'future' installed
by Paul Dufresne 06 Jul '18

06 Jul '18

Is it my system that is supposed to give Python modules, or is it supposed to be part of the toolchain? I get it on: paul@paul-OptiPlex-745:~/genode$ ./tool/ports/prepare_port sel4 /home/paul/genode/repos/base-sel4/ports/sel4.port:11: *** Need to have python module 'future' installed.. Arrêt. /home/paul/genode/tool/ports/mk/prepare_single_port.mk:38: recipe for target '_install_in_port_dir' failed make[1]: *** [_install_in_port_dir] Error 2 tool/ports/prepare_port:33: recipe for target 'sel4' failed make: *** [sel4] Error 2 paul@paul-OptiPlex-745:~/genode$

2 1

depot/build creates target directory with missing dependencies
by Guido Witmond 06 Jul '18

06 Jul '18

Hi Genodians, I think I start to get the hang of the depot system. However, I get this behaviour with missing depencencies: $ tool/depot/create guido/src/feather created guido/src/feather/2018-07-04-c $ tool/depot/build guido/bin/x86_64/feather/2018-07-04-c KEEP_BUILD_DIR=1 ... Library description file libc_lwip.mk is missing Library description file libc_lwip_nic_dhcp.mk is missing created guido/bin/x86_64/feather/2018-07-04-c But that directory is completely empty. Running the command a second time give me: make[1]: Nothing to be done for 'all'. As there are missing dependencies, I'd expect the second run to give the same error as the first run, not a message that everything is fine. To retry I currently rm -r depot/guido and create and build again or add the REBUILD=1 flag. My questions: - How can I add these lwip-libraries included? I don't see any examples of these included in used_apis. - What's going wrong that it mentions the missing deps but still creates the target directory? Cheers, Guido

2 1

make fun: optimize depot/create by using depot/download
by Roman Iten 06 Jul '18

06 Jul '18

Hello, Merely as an experiment to avoid building already published/uploaded archives, I tried to integrate depot/download into the depot/create front-end. Having basically zero make experience, I came up with [1]. If the environment variable DOWNLOAD is set, it 'seems' to do what I intend: - extract API/source/raw archives (make sure we use the correct version) - download as many archives as possible (best-effort) - build everything else, that couldn't be downloaded Still I have some questions: - Can you think of any (corner) cases which would result in an incomplete/inconsistent build regarding the content of the working directory? - How can I deal with the following warning (when calling depot/download): make[2]: warning: jobserver unavailable: using -j1. Add '+' to parent make rule. - Is there a simple way to prevent the ugly log output as it may contain (expected) error messages like: Error: failed to download 'https://...' make[2]: Target '...' not remade because of errors. - Anything else I didn't consider? Comments are highly appreciated, thanks! Roman [1] https://github.com/rite/genode/commit/1b50ac73dc3b4986c1df1694b88ecc49757d0…

1 0

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

users