Ros-dev

ros-dev@reactos.org

5788 discussions

Re: [ros-dev] [ros-diffs] [ion] 52807: [KERNEL32]: Optimize SwitchToFiber to simply use "ret" to jump between fibers, instead of saving EIP and doing a JMP. Bug #50: SwitchToFiber needs to check if FXSR is *NOT* present in ...

by Timo Kreuzer

This won't work, since new Fibers, created with CreateFiber(Ex) don't push a "return address" on the stack, but set the Eip member to BaseFiberStartup. Am 23.07.2011 14:08, schrieb ion(a)svn.reactos.org: > Author: ion > Date: Sat Jul 23 12:08:36 2011 > New Revision: 52807 > > URL: http://svn.reactos.org/svn/reactos?rev=52807&view=rev > Log: > [KERNEL32]: Optimize SwitchToFiber to simply use "ret" to jump between fibers, instead of saving EIP and doing a JMP. > Bug #50: SwitchToFiber needs to check if FXSR is *NOT* present in order to skip using ldmxcsr/stmxcsr. Previously, it would check if it's unsupported, and jump past the instruction if it was (resulting in invalid opcode instructions on older systems) > 50 bugs. Penance has been paid. > > Modified: > trunk/reactos/dll/win32/kernel32/client/i386/fiber.S > > Modified: trunk/reactos/dll/win32/kernel32/client/i386/fiber.S > URL: http://svn.reactos.org/svn/reactos/trunk/reactos/dll/win32/kernel32/client/… > ============================================================================== > --- trunk/reactos/dll/win32/kernel32/client/i386/fiber.S [iso-8859-1] (original) > +++ trunk/reactos/dll/win32/kernel32/client/i386/fiber.S [iso-8859-1] Sat Jul 23 12:08:36 2011 > @@ -26,20 +26,16 @@ > mov [eax+FIBER_CONTEXT_EDI], edi > mov [eax+FIBER_CONTEXT_EBP], ebp > > - /* Save the return address */ > - mov ebx, [esp] > - mov [eax+FIBER_CONTEXT_EIP], ebx > - > /* Check if we're to save FPU State */ > cmp dword ptr [eax+FIBER_CONTEXT_FLAGS], CONTEXT_FULL OR CONTEXT_FLOATING_POINT > jnz NoFpuStateSave > > /* Save the FPU State (Status and Control)*/ > fstsw [eax+FIBER_CONTEXT_FLOAT_SAVE_STATUS_WORD] > - fstcw [eax+FIBER_CONTEXT_FLOAT_SAVE_CONTROL_WORD] > + fnstcw [eax+FIBER_CONTEXT_FLOAT_SAVE_CONTROL_WORD] > > /* Check if the CPU supports SIMD MXCSR State Save */ > - cmp byte ptr ds:[PROCESSOR_FEATURE_FXSR], 0 > + cmp byte ptr ds:[PROCESSOR_FEATURE_FXSR], 1 > jnz NoFpuStateSave > stmxcsr [eax+FIBER_CONTEXT_DR6] > > @@ -103,7 +99,7 @@ > ControlWordEqual: > > /* Load the new one */ > - cmp byte ptr ds:[PROCESSOR_FEATURE_FXSR], 0 > + cmp byte ptr ds:[PROCESSOR_FEATURE_FXSR], 1 > jnz NoFpuStateRestore > ldmxcsr [ecx+FIBER_CONTEXT_DR6] > > @@ -121,7 +117,8 @@ > mov [edx+TEB_FLS_DATA], eax > > /* Jump to new fiber */ > - jmp dword ptr [ecx+FIBER_CONTEXT_EIP] > + mov esp, [ecx+FIBER_CONTEXT_ESP] > + ret 4 > +END > > -END > /* EOF */ > > >

13 years, 2 months

Re: [ros-dev] [ros-diffs] [ion] 52806: [KERNEL32]: Implement BaseFiberStartup in ASM, just like the thread/process thunks, so we get a "naked" thunk without any compiler-generated crap.

by Timo Kreuzer

Am 23.07.2011 14:05, schrieb ion(a)svn.reactos.org: > Author: ion > Date: Sat Jul 23 12:05:38 2011 > New Revision: 52806 > > URL: http://svn.reactos.org/svn/reactos?rev=52806&view=rev > Log: > [KERNEL32]: Implement BaseFiberStartup in ASM, just like the thread/process thunks, so we get a "naked" thunk without any compiler-generated crap. I don't see the point in this change. That "crap" is setting up a stack frame, which is pretty useful for debugging. It consists of 3 instructions: push ebp mov ebp, esp sub esp, 8 Your change also broke compilation with MSVC (the END directive is only at the end of the file!) and deleted the function for amd64. The function is not performance critical and the asm code doesn't really add any performace gain anyway. So why are we going back? I'm asking, because I've been working on converting more asm code to C code and IMO thats the way to go. Regards, Timo

13 years, 2 months

Re: [ros-dev] [ros-diffs] [dgorbachev] 52557: [FREELDR] - Move read-only data into data section (allows to boot with GRUB again). - Discard .drectve sections. - Silence "set but not used" warnings.

by Timo Kreuzer

Am 07.07.2011 21:19, schrieb dgorbachev(a)svn.reactos.org: > Author: dgorbachev > Date: Thu Jul 7 19:19:44 2011 > New Revision: 52557 > > URL: http://svn.reactos.org/svn/reactos?rev=52557&view=rev > Log: > [FREELDR] > - Move read-only data into data section (allows to boot with GRUB again). > - Discard .drectve sections. > - Silence "set but not used" warnings. I assume this is for rbuild. Did you check cmake builds, too?

13 years, 2 months

ReactOS MSVC Debugging Snapshot

by Neeraj Yadav

Hi, Amine Khaldi suggested me to take this snapshot of ReactOS being debugged in Visual Studio.Here it is

13 years, 2 months

Re: [ros-dev] [ros-diffs] [ion] 52596: [NTDLL]: More attempts at fixing up the loader, this time in the PE side of things.

by Aleksey Bragin

-----Original Message----- From: ion(a)svn.reactos.org Sent: Sunday, July 10, 2011 6:14 AM To: ros-diffs(a)reactos.org Subject: [ros-diffs] [ion] 52596: [NTDLL]: More attempts at fixing up the loader, this time in the PE side of things. Author: ion Date: Sun Jul 10 02:14:29 2011 New Revision: 52596 /* Check if we got at least one */ - if (BoundEntry || ImportEntry) + if ((BoundEntry) || (ImportEntry)) Could you enlighten me why do you add these redunandant braces? WBR, Aleksey.

13 years, 2 months

Snapshot of ReactOS being Debugged in Visual Studio

by Neeraj Yadav

http://img855.imageshack.us/img855/7541/reactosmsvcdebugging.jpg

13 years, 2 months

TestCD regression

by caemyr＠myopera.com

As some of you might already noticed, we have another runtime regression, this time limited to TestCD. What is even more interesting, same as in AC97 driver issue, it was the CMake build which started to show this bug, then followed by rbuild trunk: CMake r52578: http://build.reactos.org/builders/Windows_AMD64_1%20VBox-Test/builds/1962 RBuild r52580: http://build.reactos.org/builders/Windows_AMD64_1%20VBox-Test/builds/1961 RBuild r52581: http://build.reactos.org/builders/Windows_AMD64_1%20VBox-Test/builds/1963 I haven't seen anything hawkward on bootcd's. Those can be still booted to 3rd stage. With best regards Caemyr

13 years, 2 months

June ReactOS Meeting Minutes

by victor martinez

June Meeting Minutes 2011-06-28 19:05 UTC Fezile IRC Server, #meeting Participants ============= - Jan_Blomqvist_Kinander - Igor_Paliychuk - MfldElton - Usurp - Aleksey_Bragin - Thomas_Faber - Giannis_Adamopoulos - Thomas_Faber - Johannes_Anderwald - Timo_Kreuzer - Olaf_Siejka - Jerome_Gardou - Gabriel_Ilardi - Andrew_Green - Javier_Fernandez - Victor_Martinez - Amine_Khaldi - Claudiu_Mihail - Colin_Finck - Rafal_Harabien - Sylvain_Petreolle - Matthias_Kupfer - Danny_Goette - Maciej_Bialas - Ged_Murphy - Ziliang_Guo - Art_Yerkes - Cameron_Gutman - Neeraj_Yadav - Herve_Poussineau - Daniel_Reimer - James_Tabor - Olaf_Siejka - Samuel_Serapion - Pierre_Schweitzer - Alex_Ionescu - Gabriel_Ilardi Proceedings ============ ○ Meeting called to order at 19:05 UTC by Aleksey Bragin. ○ ReactOS Project coordinator Aleksey Bragin listed the points of the proposed agenda for June meeting: 1. Cmake switching 2. Next ReactOS release 3. Website revamp status 4. GSoC status 5. Developers status reports The Minute Taker rol was given to Victor_Martinez for this meeting. ○ Point 1: CMake switching -------------------------------------------- ■ The ReactOS developer Timo gave a quick sum up about CMake status: He reported that the CMake build is more or less ready, currently missing rosapps. He stated that rosapps are not a major problem as they're rarely used, and cmake files for them can be done if needed. He also mentioned a problem, confirmed by Igor_Paliychuk, related to booting in real hardware, in addition to the winmm:mci regression. The problem detected in real hardware is very specific to the hardware combination thus hard to test. ■ Amine mentioned that several comparative tests have been done through buildbot/testman between the rbuild build and the CMake build, showing no measurable differences. He also stated that the CMake port shouldn't be considered done yet, as it still lacks many planned features. He highlighted some of the CMake patches,made by Jerome, to be sent upstream, like the PCH support and its related dependency tracking, and also the benefits of using CMake to compile ros, like the improved performance, the small disk space needed for builds (half of what the rbuild builds require) and the cool new kdbg features.Gabriel_Ilardi expressed his concers about merging more code in the current trunk situation. Timo replied that no code merge is needed,in case of switching to Cmake. ■ Timo, Colin and Aleksey asked about measuring the CMake build performance over the rbuild build in numbers. Jerome performed a quick benchmark: First test: makex ntoskrnl (all clean, toold already built). Cmake: 04:06 , Rbuild: 04:55 (17% faster) Second test: rebuild (only dependencies check). Cmake: 0:28 , Rbuild: 1:04 (56% faster) The benchmark shows Cmake is faster not only for a full compilation but also recompiling (ntoskrnl) modules. ■ The ReactOS Cmake Team suggested to begin using Cmake with current RosBE to compile ReactOS ISOs in order to assure they are as good as Rbuild’s. A "How to compile with Cmake" tutorial can be found in the following wikipage: http://reactos.org/wiki/Cmake Summing up, the Cmake switch will bring better dependency checking, fast rebuild of single modules and much better symbols support in kdbg according to the ReactOS Cmake Team The technical description was left to be discussed in #reactos-dev, if anyone was interested. ○ Point 2 : Next ReactOS release -------------------------------------------------------- ¦ReactOS release was the 2nd point of the June meeting agenda. ■ Gabriel_Ilardi and Victor_Martinez were asked to report current trunk state from Compatibility/Stability point of view. Accordingly to forum threads, Bugzilla new entries, and Own testing,Gabriel and Victor said important regressions have been introduced due the latest changes, and mainly the LDR rewrite. Wrong icons, impossibility to shutdown, AC97 hang, several installers/apps failing are some of those important regressions. ■ Art_Yerkes has been giving a look at the AC97 bug, he was able to gather a lot of information about the regression and asks for collaboration to fix it. ■ Aleksey_Bragin, as the author of the LDR rewrite, stated that he knew the new LDR could introduce regressions but that it is a big step forward. The top priority is fixing those LDR regressions, instead of reverting, as the old LDR was indeed worse. Aleksey remarks that he needs every dev to take a look at the LDR bugs and try to fix them. The LDR regressions should be really easy to fix thanks to the rewrite, he said. ■ Amine_Khaldi suggested to decide when it is time to release performing bugfixing and full regression testing. ■ The idea of performing full regression tests now was inmediatly rejected by Colin_Finck and Olaf_Siejka since bug reports are currently showing that our trunk is in a really bad shape. Due to this, the Release discussion will be on hold until the next meeting. During this month ReactOS devs are suggested to focus in bugfixing mainly the LDR code. As these bugs are critical and able to hide other minors regressions, a special call is being made: ****************Special Call**************** "Please pick a bug from the linked ones in http://reactos.org/bugzilla/show_bug.cgi?id=6346 " ****************Special Call**************** ○ Point 3 : Website Revamp Status ------------------------------------------------------- ¦ Thanks to Colin's help, a new playground with correct permissions and tied to an SVN branch has been created. ■ Niski and aross, main webrevamp developers, have been focused on Bugzilla and Mediawiki integration, compatdb drupal port, and forum. Forum will have new features as "the best answer" and "karma" modules which will improve forum users experience. ■ Niski has completed the work on Bugzilla and Mediawiki while aross is working on compatdb and the new forum. ■ Right now, Amine said, a design/layout work is needed in parallel. ■ Colin_Finck stated that finding a designer among ReactOS developers is more difficult than findind a developer so he offered himself to help with the design/layout. ○ Point 4: Gsoc status ------------------------------------------------------------------- ■ AndrewGreen is currently cleaning a mixture of explorer_new, atl com and browseui code. The resultant code will be commited afterwards. Although he recognizes his project is behind schedule, after apologizing he says he’s confident to meet the goals by midsemester evaluation. Ged_Murphy cpp shell32 code will allow him to use atl com for everything shell related. Colin_Finck and Olaf_Siejka discussed the best way to test the new implementation to check for any possible regressions, due to the fact that Ged was not around the discussion was postponed. Jerome_Gardou asked Andrew about the WINE Shell32 Gsoc project but accordingly to Olaf, Wine has refused to cooperate regarding those. Status: Behind Schedule. ■ Giannis_Adamopoulos was too busy with exams so his project was parked a little. His commits will raise up when the last exam is done. Status: On time ■ Claudiu_Mihail thanked the great help of Cameron_Gutman and Art_Yerkes. The project is set to meet the deadlines for midterm. There are still two problems remaining: browsers don't work(it will be solved by midterms), slowness of the lwIP implementation. He has followed the test approach using several simple test apps and fixing revealed bugs. After finishing his thesis project, which is highly related to his Gsoc project,he will start commiting again. ■ Neeraj_Yadav has divided his sound project in 5 parts. 1.[Finished] Writing a skeleton server which can hold the audio converter modules and the mixer modules ...This server should be capable of insertion in the reactos audio chain [4 weeks] 2.[Now] Writing mixer modules [2 weeks] 3.Writing audio converter [ 4 weeks] 4.Inserting this audio server in the reactos audio chain [1 week] 5.testing/styling/documentation [1 week] He expects to finish the 2nd step in midterm evaluation. Alex_Ionescu and Giannis_Adamopoulos asked for a way to test Neeraj code. If anyone is interested, this is the best way: 1)make audsrv && make audclient 2)run audsrv -n in one console 3)run audclient in other 4)open as many audclients as you want 5)the server should play a mixed sound Status: On time ■ Thomas_Faber is finishing the main framework needed to run the tests. Most of the old tests have been ported and now he is currently working on getting special-purpose drivers and usermode test parts up and running. Lately he has been quite busy with a university project but he expects to commit the framework in these days. Status: On time ■ Timo_Kreuzer has finished the font driver loading and now he is working on the font mapper, which is complex and needs a good design. Status: Ahead of time ○ Point 5: Developers status reports ------------------------------------------------------------------- ■ Alex_Ionescu will help Aleksey commit his scater/gather DMA patch. ■ Alex_Bragin is working on remaining LDR code and afterwards he'll take part in processing bugzilla patches. ■ Amine_Khaldi is focused in cmake and the website revamp. ■ Art_Yerkes made some improvements to symbol handling in the cmake branch and now he’s working with Claudiu. ■ Colin_Finck has been trying the KDUSBCOM idea (KD DLL for debugging over an USB-to-Serial adapter), and he’s moving now to help with the website layout. ■ Daniel_Reimer is playing with Gcc builds and updating RosBE.He added Cmake support a while ago and he is still battling with GCC 4.6 and Wine RC files. ■ Gabriel_Ilardi is focused in mantaining Bugzilla and the forum and he’s performing some testing in his few spare time. ■ Ged_Murphy is helping with shell32 c++ fixup ■ Herve_Poussineau is currently inactive on ReactOS ■ James_Tabor is still working on bug report issues. ■ Jan_Blomqvist_Kinander will move the Fezile server to the concrete bunker creating a LAN behind Fezile to get automated tests on real hardware. ■ Johannes Anderwald will work in HID (Human Interface Devices for USB) support during the summer, a hard task as it is completly undocumented. ■ Maciej_Bialas is working on mediawiki integration for the new revamped site. ■ Matthias_Kupfer has been bugfixing and commiting some patches. ■ Pierre_Schweitzer has no time nor willingness right now. Once time will be back,he'll invest it in his summer student.Once willingness will be back, he'll keep on his current ReactOS work and help with the ReactOS infrastructure ■ Olaf_Siejka is waiting for Aleksey to finish the LDR code so he can team up with him to solve the remaining issues of sysreg4. After finishing the utf-8 resource conversion, he will get more involved in testing. He will also help with the hardware testbot from the buildbot side. ■ Sylvain_Petreolle is working on French translations. ■ Timo_Kreuzer is working on MSVC builds fixing several bugs that prevent ReactOS to reach the desktop. ■ Victor_Martinez has finished his exams recently and now he’s preparing a forum team to test the revision before the LDR rewrite and the one after it, this way it'll be easier to find regressions. Also he will help with the revamp webpage. ○ Point 6: Suggestions --------------------------- ■ Jan_Blomqvist_Kinander suggested to create a ReactOS annual bootcamp with devs, testers and fanatics plus having a look at SAMBA 4 - ROS port ○ Meeting closed at 22:08 UTC by Colin Finck on behalf of Aleksey Bragin. ○ Minutes written by Víctor Martínez. ○ Reviewed by Amine_Khaldi, and Gabriel_Ilardi

13 years, 2 months

Myreactos.com hosting account will be suspended in the next 24 hours.

by victor martinez

Hi! Sadly the hosting that I was providing for free will be suspended in the next 24 hours. I was totally fed up with these iPage guys so decided to move everything out. The website and all the content in the folders will be removed. I will keep the myreactos.com domain for a year or maybe I will remove it too.If anyone is interested please contact me. If anyone is still using the hosting, please back your files up. WBR, Víctor Martínez

13 years, 2 months

Re: [ros-dev] [ros-diffs] [cgutman] 52522: [LDR] - "Just to be sure" is no reason to overwrite a potential DLL load failure status

by Aleksey Bragin

That was there on purpose really, because the error had to be ignored. I will get back to this tomorrow to review, so just "marking" this revision that way. WBR, Aleksey. -----Original Message----- From: cgutman(a)svn.reactos.org Sent: Monday, July 04, 2011 12:16 AM To: ros-diffs(a)reactos.org Subject: [ros-diffs] [cgutman] 52522: [LDR] - "Just to be sure" is no reason to overwrite a potential DLL load failure status Author: cgutman Date: Sun Jul 3 20:16:12 2011 New Revision: 52522 URL: http://svn.reactos.org/svn/reactos?rev=52522&view=rev Log: [LDR] - "Just to be sure" is no reason to overwrite a potential DLL load failure status Modified: trunk/reactos/dll/ntdll/ldr/ldrapi.c Modified: trunk/reactos/dll/ntdll/ldr/ldrapi.c URL: http://svn.reactos.org/svn/reactos/trunk/reactos/dll/ntdll/ldr/ldrapi.c?rev… ============================================================================== --- trunk/reactos/dll/ntdll/ldr/ldrapi.c [iso-8859-1] (original) +++ trunk/reactos/dll/ntdll/ldr/ldrapi.c [iso-8859-1] Sun Jul 3 20:16:12 2011 @@ -345,9 +345,6 @@ DllName, BaseAddress, TRUE); - - /* Set it to success just to be sure */ - Status = STATUS_SUCCESS; /* Restore the old TLD DLL */ LdrpTopLevelDllBeingLoaded = OldTldDll;

13 years, 2 months

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Ros-dev