Ros-dev

ros-dev@reactos.org

5789 discussions

pefixup doesnot set IMAGE_SCN_MEM_NOT_PAGED for .rdata WDM section

by Pavel Pisa

Hello all, I have noticed, that pxefixup does not set IMAGE_SCN_MEM_NOT_PAGED flag for read-only data section ".rdata". May it be, that it is problem for my own build of tools 586-mingw32-gcc (GCC) 4.4.2 and GNU ld (GNU Binutils) 2.20.51.20100608 I have fixed the problem by preparing modified LD script version which collect object files .rdata section into .text section for my driver build. But .rdata is legitimate section and it is automatically generated by GCC. I have not found information in Wiki or list archives, that there is some way/patch to deal with this situation. But if it is allowed, that WDM driver binaries contains readonly section then data paging could lead to horrible consequences. Possible fix diff --git a/reactos/tools/pefixup.c b/reactos/tools/pefixup.c index f908a9a..d6ec1f4 100644 --- a/reactos/tools/pefixup.c +++ b/reactos/tools/pefixup.c @@ -381,6 +381,7 @@ int main(int argc, char **argv) if (!strcmp((char*)section_header->Name, ".text") || !strcmp((char*)section_header->Name, ".data") || !strcmp((char*)section_header->Name, ".idata") || + !strcmp((char*)section_header->Name, ".rdata") || !strcmp((char*)section_header->Name, ".bss")) { section_header->Characteristics |= htodl(IMAGE_SCN_MEM_NOT_PAGED); If .rdata are not allowed or should be used only for pseudo_reloc, then the ldscript for kernel code should be provided. Missing flag for .rdata section in objdump ouput $ i586-mingw32-objdump --headers ul_wdm.sys BFD: ul_wdm.sys: Warning: Ignoring section flag IMAGE_SCN_MEM_NOT_PAGED in section .text BFD: ul_wdm.sys: Warning: Ignoring section flag IMAGE_SCN_MEM_NOT_PAGED in section .data BFD: ul_wdm.sys: Warning: Ignoring section flag IMAGE_SCN_MEM_NOT_PAGED in section .bss BFD: ul_wdm.sys: Warning: Ignoring section flag IMAGE_SCN_MEM_NOT_PAGED in section .idata ul_wdm.sys: file format pei-i386 Sections: Idx Name Size VMA LMA File off Algn 0 .text 0000cabc 00011000 00011000 00001000 2**2 CONTENTS, ALLOC, LOAD, READONLY, CODE 1 .data 000000e8 0001e000 0001e000 0000e000 2**2 CONTENTS, ALLOC, LOAD, DATA 2 .rdata 00003c20 0001f000 0001f000 0000f000 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 3 .bss 00000050 00023000 00023000 00000000 2**2 ALLOC 4 .edata 00000033 00024000 00024000 00013000 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 5 .idata 000008c4 00025000 00025000 00014000 2**2 CONTENTS, ALLOC, LOAD, DATA 6 .rsrc 0000035c 00026000 00026000 00015000 2**2 CONTENTS, ALLOC, LOAD, DATA 7 .reloc 00000cd4 00027000 00027000 00016000 2**2 CONTENTS, ALLOC, LOAD, READONLY, DATA 8 .rossym 0001576d 00028000 00028000 00017000 2**2 CONTENTS, READONLY, NEVER_LOAD, EXCLUDE

14 years, 7 months

reactos in lguest

by Enrico Weigelt

Hi folks, did anyone already do some research on whether ReactOS could run in lguest ? Maybe this could be a nice for development/debugging and attracting more people ;-) cu -- ---------------------------------------------------------------------- Enrico Weigelt, metux IT service -- http://www.metux.de/ phone: +49 36207 519931 email: weigelt(a)metux.de mobile: +49 151 27565287 icq: 210169427 skype: nekrad666 ---------------------------------------------------------------------- Embedded-Linux / Portierung / Opensource-QM / Verteilte Systeme ----------------------------------------------------------------------

14 years, 7 months

0.3.12 TODO list

by Olaf Siejka

As we progress into regression fixing, we are now in the good moment to decide upon upcoming release. It would be great if we could manage with the 0.3.12 for the October Chemnitz event, where we are invited again to participate, not to mention that its the time for such release. Right now, in my opinion we should decide without futher delay on the following: - reviewing the REGRESSION list, this is - to confirm entries with 0.3.12 milestone already set, as well as doublecheck the rest, to see if any should be promoted to said milestone. This is the action of most utter importance, as there is no way to know how much time will be spent on futher research, testing and fixing them; - decide upon bugs originated with last ole/comctl sync - are we haxing them locally, risking another winesync or perhaps reverting? - reviewing patches stored at http://www.reactos.org/wiki/Bug_Filters as well as deciding if any should be adapted for current release. Those actions, first of them especially, should be performed as swift as possible, else we might risk considerable delay and nervous finish before the Chemnitz.

14 years, 7 months

ASSERT while shutting down ReactOS with VirtualBox Guest Additions installed

by Javier Agustìn Fernàndez Arroyo

hello, this is to sir_richard, mostly, i got this crash (assert) when shutting down ReactOS with the VirtualBox Guest Additions installed. i think the video driver is guilty for some reason i cannot understand... probably you know :) http://www.reactos.org/paste/index.php/7731/

14 years, 7 months

uLan USB driver for ReactOS - query for help if possible

by Pavel Pisa

Hello to all ReactOS developers, thanks for great work. I introduce our case first. -------------------------------------------------------------------- We have developed, maintain and use uLan RS-485 multi-master protocol for years - sources traced to 1992 year. The protocol is used by multiple companies and individuals for broad range of applications from laboratory instruments control, home automation to agriculture feeding and milking systems. We started by Linux support. But due to customers demand, we have added Windows KDM and later WDM support as well. I remember, that pointer to our driver has been mentioned as one of little fully open-source drivers available for Windows on ReactOS list many many years ago. The uLan project homepage http://ulan.sourceforge.net/ The uLan driver code can be compiled for Linux, Windows NT KDM, Win98/2k/Vista WDM, plain DOS, system-less ARM and other devices. The old legacy UART, PCI addon cards and USB converters are supported for Linux and WDM version. All builds form same sources with unification layer of macros which unifies systems to something more close to Linux driver model. The code is not so elegant and readable, the portability layers are quite complex and sometimes hairy tasks. The traces of old age of project can be seen there as well. But it works and may it be, it could be some example and source of helpers functions for porting of Linux USB drivers to WDM model. The ReactOS has got to the stage where it is able to run one of our user applications utilizing our control protocol link. Because I believe in open technologies, I would be very happy if the project is usable even on RectOS - even that I do not expect that our users could and want to switch anytime soon or ever. The Open Chromatography Station - CHROMuLAN http://sourceforge.net/projects/chromulan/ If you think that it worth, it can be added to the list of applications working with ReactOS 0.4-git/svn. -------------------------------------------------------------------- As for the uLan driver. I have maneged it build for WDF with PCI and UART support only to run on ReactOS and whole CHROMuLAN setup seems to be functional. I have even interrest to build driver code with ReactOS to test it more, find incompatibilities and continue with 64-bit testing in future. The effort seems to be not so far form success. The driver builds with RectOS build under Linux when USB support is disabled. The build with USB seems to be near as well. There are two main problems: 1) We need to include usbdlib.h in the driver build, but DECLSPEC_EXPORT in this header is not declared/maintained. We use this file for next types declaration USBD_INTERFACE_LIST_ENTRY PUSBD_INTERFACE_LIST_ENTRY struct _USBD_INTERFACE_LIST_ENTRY' May it be, we should use different include headers, but next setup works for years with Microsoft WDF #include "wdm.h" #include "usbdi.h" #include "usbdlib.h" 2) If we overcome by some hack problem 1) the link fails on undefined references to __imp__USBD_ParseConfigurationDescriptorEx@28 __imp__USBD_CreateConfigurationRequestEx@8 There seems to be traces of implementation of these functions in ReactOS old/disabled USB code. The functions are even included in usbd.sys, usbd.exp and drivers/usb/usbd . But import library seems to be missing. Am I right? Is there way to use them or plan to solve that. Thanks for reply in advance. Even that we do not much expect to use ReactOS build in production environment the option has value for us. It allows to test Windows builds in Linux environment without need to start Wine for WDF. ReactOS build even helped us to find some real bugs in our code because rectos runtime is more strict in some checks then Windows systems. Please, keep my email addres in reply, I do not expect (have ability) to follow ReactOS development closely between my other duties. It is unfortunate a little, that only way to contact developers on the list is to subscribe. Best wishes, Pavel -- Pavel Pisa e-mail: pisa(a)cmp.felk.cvut.cz www: http://cmp.felk.cvut.cz/~pisa university: http://dce.fel.cvut.cz/ company: http://www.pikron.com/

14 years, 7 months

From what to start?

by Captain Sandro

What can you advice me to start from? Which module needed in programming? And which tools and methods are used for writing code?

14 years, 7 months

Errata

by Olaf Siejka

Altough i mentioned Chemitz, it should be of course OpenSource World Conference in Spain. Sorry

14 years, 7 months

Windows & ReactOS installed simultaneously

by Javier Agustìn Fernàndez Arroyo

Hello devs, Few users at spanish ReactOS blog ( http://reactos.wordpress.com ) are asking about having both Windows and ReactOS installed in different partitions at the same time... And i wanted you to answer, rather than me... :P What would you say? Is it currently possible?

14 years, 7 months

A room for improvement?

by Pierre Schweitzer

Hi, while working on GPT handling for ReactOS, one thing appeared: Microsoft (at least in Windows 2003 SP1) is using hardcoded partition entry size, ie it assumes that every partition entry in GPT will be 128 bytes big. Even if that assertion is largely right, it might exist cases where it's wrong. That's even why there's a field in GPT header that contains such size. Furthermore, reading Apple technical note(1) about GPT precise one more thing: "Do not hardwire the current size of the partition entry (128 bytes).". As we have no commercial issues with Apple, perhaps could we drop a bit compatibility with Windows 2003 and fix something obvious? In most cases, users won't see any difference. Only difference would be that ReactOS could properly handle GPT with exotic partition entry size when Windows can't. Any opinion on that subject? Best regards, P. Schweitzer (1): http://developer.apple.com/library/mac/#technotes/tn2006/tn2166.html

14 years, 7 months

48748 "Goodbye" Confusion

by Ros Arm

Hello, eVb checked in one of his final patches yesterday from his development machine in China, where we worked on a final pass on the driver. Unfortunately, my last two (unrelated) check ins were done from his machine (as most of you know, I was in China for the past month), and he was not used to SVN caching the credentials of the last user. Revision 48748 was a checkin from eVb, and his "goodbye" was in reference to my departure. Since it used my cached credentials from the same machine, it must've caused a lot of confusion. Many apologies and thanks for understanding. -r

14 years, 7 months

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Ros-dev