Ros-dev

ros-dev@reactos.org

5788 discussions

by Timo Kreuzer

Hi, I am pleased to announce the long awaited (or absolutely surprising) release of PSEH3. This complete from scratch rewrite takes reactos exception handling to a new level. It comprises the following advancements: - Use of enums instead of const variables for internal trylevel counting and other mechanisms. enums can be declared in a nested manner as well and only enums are treated by gcc as full compile time constants, while const variables are treated as dynamic and are only optimized away by the compiler later. This generally results in better optimized code. - No more use of dynamic sized arrays. Those make the compiler emit a slow call to _chkstk. It could have been prevented in PSEH2, if enums had been used instead of consts. The new code uses the same registration frame for the top level and nested levels, which also simplifies the code. - No more use of nested function stack trampolines. These are small code chunks that are being dynamically put on the stack, when the address of a nested function is taken. The address of the nested function then points to this trampoline, which sets up a pointer for nested variable access (nested function can access variables in the parent frame). These were parsed by PSEH2 to calculate the neccessary nested frame pointer. PSEH3 uses a static scope table instead that contains the address of the nested function. this is only possible by first declaring the function and putting the implementation after the address is stored in the static table. This is because nested functions that do not access any variables in the parent frame do not need a stack trampoline and their addresses are thus considered compiler constants. GCC doesn't notice that a nested function uses variables of the parent frame, before the function is actually implemented, so with only a prototype, it assumes a constant address. To get the value of the parent frame pointer, PSEH2 calls the nested function 2 times: The first time with a bogus frame pointer and a parameter that tells the function to return the (assumed) address of the current registration frame. Without a proper parent frame pointer the nested function will return an offset that we can use together with the real address of the registration frame (which we know, when calling the filter of finally functions) to calculate the proper parent frame pointer. The 2nd call passes a different parameter and allows the function to do the actual work and access variables in the parent frame. - Improved performance due to use of "asm goto". This construct allows to get away with having labels, that are never accessed by any C code. The compiler doesn't remove them. So we do not need the old hack, of comparing a "volatile" 0 against 0, just to confuse GCC to not optimize away a label. Even though there is a bug/defficiency/limitation in asm goto, that would usually not allow the combination with static C jump labels, we can still do that due to some tricks that force GCC to do the right thing. More information: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51840 - Improved performance due to faster registration. The new registration code is smaller and saves only esp and ebp. Due to some tricks in PSEH3, GCC is forced to assume all registers are being clobbered when the __except block is run. Filter function and __finally function don't make use of registers from the previous code. So all neccessary register save and restore is now fully done by GCC. - The trick to make GCC assume all registers are clobbered at the beginning of the __except block has one additional huge advantage. Now all variables that are accessed by the __except block will be stored in memory locations, so that modifications in the __try block will be available in the __except block, something that doesn't work with PSEH2. This does not prevent the compiler from doing constant expression elimination! So when you try to do 2 different, constant assignments to a variable in the __try block, expect that this will be reduced to only the later assignment. GCC does not know, that the execution can be interrupted in the middle of the __try block between those 2 assignments. If you need such stuff, you must use a volatile variable. - Less stack usage. PSEH3 stores much less data on the stack. No trampolines, only esp and ebp are saved, registration frames are smaller. - No more need for _SEH2_YIELD(). This macro was used to make sure the registration frames are properly unregistered, when leaving the try or except block using goto or return. This is not neccessary anymore due to the use of __attribute__((cleanup)). This attribute is attached to the registration frame and causes the cleanup function to be called automatically, whenever the variable goes out of scope. This happens when a goto or return is used in the code. - Human readable warnings, when special SEH instriniscs, like _exception_info() or _abnormal_termination() are used outside of their valid context. - Cleaner and better readable code, so you won't get braindead, when trying to figure out what's going on. Header: < 300 lines commented code, instead of 370 lines uncommented code, lib: < 250 lines of commented C code + < 70 lines commented asm code, instead of 290 lines of uncommented C code + 120 lines uncommented asm code) I also promised Amine to take a look into implementing a version that is compatible with clang. Initially I planned to include this into this release, but it turns out that clang massivley lacks features to do so and a completely different approach will probably be used for that. The new code is only compatible with GCC 4.7 and later, and it's also currently disabled (cmake option USE_PSEH3). Tesbot showed no noticable regressions. Have fun. Timo

12 years

Bugzilla outage

by Aleksey Bragin

Hello, I turned off Bugzilla now in favor of finishing the switch to JIRA. Once the work is done, I will announce either return to Bugzilla or introduce JIRA. Hopefully, it'll be the latter option. Best regards, Aleksey.

12 years

Fwd: buildbot failure in ReactOS on CMake_x86_MSVCWin Debug

by caemyr＠myopera.com

This commit breaks msvc compilation. Best regards ----- Original message ----- From: buildbot(a)reactos.org To: caemyr(a)myopera.com Subject: buildbot failure in ReactOS on CMake_x86_MSVCWin Debug Date: Tue, 04 Sep 2012 03:10:09 +0000 The Buildbot has finished a build on builder CMake_x86_MSVCWin Debug while building ReactOS. Full details are available at: http://build.reactos.org/builders/CMake_x86_MSVCWin%20Debug/builds/1908 Buildbot URL: http://build.reactos.org/ Buildslave for this Build: Windows_AMD64_1 Build Reason: scheduler Build Source Stamp: 57233 Blamelist: cgutman BUILD FAILED: failed building bootcd sincerely, -The Buildbot -- With best regards Caemyr

12 years

Re: [ros-dev] buildbot failure in ReactOS on Windows_AMD64_1 VBox-Test

by caemyr＠myopera.com

Currently, we experience random crashes in 2nd stage. I was unable to catch any first exception bar the one with vmware detection code (which is outdated and should be removed). After forcing build again, test run failed at Cleanup, due to previously unresolved issue of repeating whole testrun when crash occurs in APItests. On Mon, Sep 3, 2012, at 10:21 AM, buildbot(a)reactos.org wrote: > The Buildbot has finished a build on builder Windows_AMD64_1 VBox-Test > while building ReactOS. > Full details are available at: > http://build.reactos.org/builders/Windows_AMD64_1%20VBox-Test/builds/6496 > > Buildbot URL: http://build.reactos.org/ > > Buildslave for this Build: Windows_AMD64_2 > > Build Reason: The web-page 'force build' button was pressed by 'osiejka': > > Build Source Stamp: 57227 > Blamelist: > > BUILD FAILED: failed Cleanup > > sincerely, > -The Buildbot > -- With best regards Caemyr

12 years

Fwd: buildbot failure in ReactOS on Windows_AMD64_1 VBox-Test

by caemyr＠myopera.com

This one is caused by rev 57225: Backtrace: http://www.reactos.org/paste/index.php/10472/ Please fix ----- Original message ----- From: buildbot(a)reactos.org To: caemyr(a)myopera.com Subject: buildbot failure in ReactOS on Windows_AMD64_1 VBox-Test Date: Sun, 02 Sep 2012 22:50:06 +0000 The Buildbot has finished a build on builder Windows_AMD64_1 VBox-Test while building ReactOS. Full details are available at: http://build.reactos.org/builders/Windows_AMD64_1%20VBox-Test/builds/6492 Buildbot URL: http://build.reactos.org/ Buildslave for this Build: Windows_AMD64_2 Build Reason: Triggerable(Windows_AMD64_1 VBox-Test Trigger) Build Source Stamp: 57225 Blamelist: ion BUILD FAILED: failed test sincerely, -The Buildbot -- With best regards Caemyr

12 years

Problem with dbgwin revisions from Windows builder

by caemyr＠myopera.com

Hiya R3ddr4g0n reported to me that all recent revisions uploaded to iso.reactos.org are not correct. Instead of given revision, all the builds are reporting to be rev 57150. After a quick check, i downloaded bootcd-57221-dbgwin.iso and checked its SHA1: should be 56d2190610e8a1f650faa3c6a6a8348f89d1fa1b but happened to be: f97ac4aba0e5ceed3b5beae3d7ab2bc2c7600fab - which is indeed the same checksum as for x86_CMake-57150 The possible clue from the upload script: bash: rsync: command not found rsync: connection unexpectedly closed (0 bytes received so far) [sender] rsync error: error in rsync protocol data stream (code 12) at /home/lapo/package/rsync-3.0.8-1/src/rsync-3.0.8/io.c(601) [sender=3.0.8] * Done. * Remotely compressing bootcd-57219-dbgwin... Now, Colin may recall that i was strongly against using fancy tricks for sending the revs instead of simply archiving them locally and uploading afterwards. This way with rsync was supposed to be saving bit of bandwidth. Now we have to reupload at least all since 57150, unless you can use the fancy tricks to spot the same iso files. Great thing we can save a little bit of bandwidth but waste time to fix it afterwards. Accoring to the reporter: [21:45:37] <r3ddr4g0n> for sure from 57115 to 57221 If you still think this methode is saving us anything, please reckon that a single view of testbot log means downloading a few megabytes of text at least. I`m stopping the uploads until the problem is resolved in a way to prevent it from happening again. -- With best regards Caemyr

12 years

Re: [ros-dev] [ros-diffs] [ion] 57215: [NTOSKRNL]: Do not return data in failure cases in NtProtectVirtualMemory. [NTOSKRNL]: No longer support non-ARM3 sections in NtProtectVirtualMemory, as the only OS calls were already ...

by Thomas Faber

Hi! On 02.09.2012 03:17, ion(a)svn.reactos.org wrote: > Modified: trunk/reactos/ntoskrnl/mm/ARM3/virtual.c > URL: http://svn.reactos.org/svn/reactos/trunk/reactos/ntoskrnl/mm/ARM3/virtual.c… > ============================================================================== > --- trunk/reactos/ntoskrnl/mm/ARM3/virtual.c [iso-8859-1] (original) > +++ trunk/reactos/ntoskrnl/mm/ARM3/virtual.c [iso-8859-1] Sun Sep 2 01:17:42 2012 > @@ -3668,15 +3795,50 @@ > } > > // > - // Assert on the things we don't yet support > - // > - ASSERT(ZeroBits == 0); > - ASSERT((AllocationType& MEM_LARGE_PAGES) == 0); > - ASSERT((AllocationType& MEM_PHYSICAL) == 0); > - ASSERT((AllocationType& MEM_WRITE_WATCH) == 0); > - ASSERT((AllocationType& MEM_TOP_DOWN) == 0); > - ASSERT((AllocationType& MEM_RESET) == 0); > - ASSERT(Process->VmTopDown == 0); > + // Fail on the things we don't yet support > + // > + if (ZeroBits != 0) > + { > + DPRINT1("Zero bits not supported\n"); > + Status = STATUS_INVALID_PARAMETER; > + goto FailPathNoLock; > + } > + if ((AllocationType& MEM_LARGE_PAGES) == 1) > + { > + DPRINT1("MEM_LARGE_PAGES not supported\n"); > + Status = STATUS_INVALID_PARAMETER; > + goto FailPathNoLock; > + } > + if ((AllocationType& MEM_PHYSICAL) == 1) > + { > + DPRINT1("MEM_PHYSICAL not supported\n"); > + Status = STATUS_INVALID_PARAMETER; > + goto FailPathNoLock; > + } > + if ((AllocationType& MEM_WRITE_WATCH) == 1) > + { > + DPRINT1("MEM_WRITE_WATCH not supported\n"); > + Status = STATUS_INVALID_PARAMETER; > + goto FailPathNoLock; > + } > + if ((AllocationType& MEM_TOP_DOWN) == 1) > + { > + DPRINT1("MEM_TOP_DOWN not supported\n"); > + Status = STATUS_INVALID_PARAMETER; > + goto FailPathNoLock; > + } > + if ((AllocationType& MEM_RESET) == 1) > + { > + DPRINT1("MEM_RESET not supported\n"); > + Status = STATUS_INVALID_PARAMETER; > + goto FailPathNoLock; > + } These guys won't work, you need != 0 or == MEM_BLAH -Thomas

12 years

Monthly meeting - August 2012

by Aleksey Bragin

Hello, Let me invite you to the monthly status meeting taking place last Thursday of this month, 30th of August, 19:00 UTC. The meeting will be at irc://fezile.reactos.org (Port 6667, no SSL) in the channel #meeting, if the server works. If not, then an alternative place will be announced here. Note that the IRC service will only be started shortly before the meeting. Your participation passwords will be emailed to you shortly before the meeting starts. If someone still is not getting passwords sent before a meeting - please email Colin or Pierre before the meeting started to get one. In order to save time, let's choose who is going to be the minute taker on the upcoming meeting. As usual, Z98 is welcome to volunteer. The agenda will be posted shortly before the meeting, suggestions are welcome (send them to my email as usual). With the best regards, Aleksey Bragin.

12 years

Re: [ros-dev] [ros-diffs] [jgardou] 57171: [MINGWEX] - mark DllMain as a weak symbol for GCC. - supply a stubbed DllMain for MSVC. - DllMain is optional, and some DLLs don't implement it. That doesn't mean that they have no...

by Timo Kreuzer

Hi, I don't agree with these changes. If we need a DllEntry in the lib, then we should use the one that is provided by mingw-w64 (all this code is 100% mingw-w64). It simply returns true (thats exactly what the one in MS CRT does) and works for gcc and MSVC. Also what's the point of DisableThreadLibraryCalls() here? As you wrote yourself, no DllMain only means "I have nothing more to initialize than the CRT". But this disables the CRT thread initializion, too. There is also no need for weak symbols, we can simply call the stub DllMain, since it doesn't (shouldn't) do anything. Regards, Timo Am 27.08.2012 01:31, schrieb jgardou(a)svn.reactos.org: > Author: jgardou > Date: Sun Aug 26 23:31:49 2012 > New Revision: 57171 > > URL: http://svn.reactos.org/svn/reactos?rev=57171&view=rev > Log: > [MINGWEX] > - mark DllMain as a weak symbol for GCC. > - supply a stubbed DllMain for MSVC. > - DllMain is optional, and some DLLs don't implement it. That doesn't mean that they have no entry point, it means "I have nothing more to initialize than the CRT". > > Added: > trunk/reactos/lib/sdk/crt/startup/mscdllmain.c (with props) > Modified: > trunk/reactos/lib/sdk/crt/msvcrtex.cmake > trunk/reactos/lib/sdk/crt/startup/crtdll.c > > Modified: trunk/reactos/lib/sdk/crt/msvcrtex.cmake > URL: http://svn.reactos.org/svn/reactos/trunk/reactos/lib/sdk/crt/msvcrtex.cmake… > ============================================================================== > --- trunk/reactos/lib/sdk/crt/msvcrtex.cmake [iso-8859-1] (original) > +++ trunk/reactos/lib/sdk/crt/msvcrtex.cmake [iso-8859-1] Sun Aug 26 23:31:49 2012 > @@ -66,7 +66,9 @@ > endif() > > if(MSVC) > - list(APPEND MSVCRTEX_SOURCE startup/mscmain.c) > + list(APPEND MSVCRTEX_SOURCE > + startup/mscmain.c > + startup/mscdllmain.c) > else() > list(APPEND MSVCRTEX_SOURCE startup/gccmain.c) > endif() > > Modified: trunk/reactos/lib/sdk/crt/startup/crtdll.c > URL: http://svn.reactos.org/svn/reactos/trunk/reactos/lib/sdk/crt/startup/crtdll… > ============================================================================== > --- trunk/reactos/lib/sdk/crt/startup/crtdll.c [iso-8859-1] (original) > +++ trunk/reactos/lib/sdk/crt/startup/crtdll.c [iso-8859-1] Sun Aug 26 23:31:49 2012 > @@ -50,7 +50,19 @@ > > extern int mingw_app_type; > > +/* > + * It is possible that a DLL provides no DllMain entry point. > + * Mark it as a weak symbol for GCC. > + * Tests show that at link time, MSVC looks for a function first in the object files provided, and then > + * in the libraries. This means that we must provide a basic implementation in msvcrtex, which will be used > + * if none is found in the object files provided to link.exe. > + * This also means that we can't rely on a DllMain function implemented in a static library when linking a DLL. > + */ > +#ifdef __GNUC__ > +extern WINBOOL WINAPI DllMain (HANDLE hDllHandle, DWORD dwReason, LPVOID lpreserved) __attribute__((weak)); > +#else > extern WINBOOL WINAPI DllMain (HANDLE hDllHandle, DWORD dwReason, LPVOID lpreserved); > +#endif > > extern WINBOOL WINAPI DllEntryPoint (HANDLE, DWORD, LPVOID); > > @@ -198,10 +210,12 @@ > } > if (dwReason == DLL_PROCESS_ATTACH) > __main (); > - retcode = DllMain(hDllHandle,dwReason,lpreserved); > + if(DllMain) > + retcode = DllMain(hDllHandle,dwReason,lpreserved); > if (dwReason == DLL_PROCESS_ATTACH && ! retcode) > { > - DllMain (hDllHandle, DLL_PROCESS_DETACH, lpreserved); > + if(DllMain) > + DllMain (hDllHandle, DLL_PROCESS_DETACH, lpreserved); > DllEntryPoint (hDllHandle, DLL_PROCESS_DETACH, lpreserved); > _CRT_INIT (hDllHandle, DLL_PROCESS_DETACH, lpreserved); > } > > Added: trunk/reactos/lib/sdk/crt/startup/mscdllmain.c > URL: http://svn.reactos.org/svn/reactos/trunk/reactos/lib/sdk/crt/startup/mscdll… > ============================================================================== > --- trunk/reactos/lib/sdk/crt/startup/mscdllmain.c (added) > +++ trunk/reactos/lib/sdk/crt/startup/mscdllmain.c [iso-8859-1] Sun Aug 26 23:31:49 2012 > @@ -1,0 +1,11 @@ > +#include <oscalls.h> > +#define _DECL_DLLMAIN > +#include <process.h> > + > +WINBOOL WINAPI DllMain (HANDLE hDllHandle, DWORD dwReason, LPVOID lpreserved) > +{ > + /* If the DLL provides no DllMain, then chances are that it doesn't bother with thread initialization */ > + if(dwReason == DLL_PROCESS_ATTACH) > + DisableThreadLibraryCalls(hDllHandle); > + return TRUE; > +} > > Propchange: trunk/reactos/lib/sdk/crt/startup/mscdllmain.c > ------------------------------------------------------------------------------ > svn:eol-style = native > > >

12 years

Re: [ros-dev] [ros-diffs] [tkreuzer] 57141: [FREELDR] Improve readability of sector calculation

by Timo Kreuzer

Btw, we are using a lot of hardcoded 512 values here for sector size even though there is a field in fthe Volume structure for that. Is that correct? Am 23.08.2012 14:02, schrieb tkreuzer(a)svn.reactos.org: > Author: tkreuzer > Date: Thu Aug 23 12:02:30 2012 > New Revision: 57141 > > URL: http://svn.reactos.org/svn/reactos?rev=57141&view=rev > Log: > [FREELDR] > Improve readability of sector calculation > > Modified: > trunk/reactos/boot/freeldr/freeldr/fs/fat.c > > Modified: trunk/reactos/boot/freeldr/freeldr/fs/fat.c > URL: http://svn.reactos.org/svn/reactos/trunk/reactos/boot/freeldr/freeldr/fs/fa… > ============================================================================== > --- trunk/reactos/boot/freeldr/freeldr/fs/fat.c [iso-8859-1] (original) > +++ trunk/reactos/boot/freeldr/freeldr/fs/fat.c [iso-8859-1] Thu Aug 23 12:02:30 2012 > @@ -1338,8 +1338,7 @@ > // > // Seek to right position > // > - Position.HighPart = SectorNumber >> 23; > - Position.LowPart = SectorNumber << 9; > + Position.QuadPart = (ULONGLONG)SectorNumber * 512; > ret = ArcSeek(Volume->DeviceId, &Position, SeekAbsolute); > if (ret != ESUCCESS) > { > > >

12 years

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