Ros-dev

ros-dev@reactos.org

5790 discussions

Re: [ros-dev] [ros-diffs] [tfaber] 71307: [HAL] - Create a wrapper version of HalpEndSoftwareInterrupt that frees its stack before calling the noreturn version of the next software interrupt handler. Fixes excessive stack u...

by Alex Ionescu

Wait, are you saying that Timo didn't discover this obvious bug for years? Best regards, Alex Ionescu On Mon, May 16, 2016 at 4:06 PM, Timo Kreuzer <timo.kreuzer(a)web.de> wrote: > Wait, are you suggesting that "doing everything the same as MS does" and > "I don't need to care for C standards, I have a huge epeen!" and > "OnlyOnePersonCanWriteHalCode AndItsNotYou".... doesn't work? :O > > > Best regards, > Timo Kreuzer > > > Am 10.05.2016 um 17:03 schrieb tfaber(a)svn.reactos.org: >> >> Author: tfaber >> Date: Tue May 10 15:03:56 2016 >> New Revision: 71307 >> >> URL: http://svn.reactos.org/svn/reactos?rev=71307&view=rev >> Log: >> [HAL] >> - Create a wrapper version of HalpEndSoftwareInterrupt that frees its >> stack before calling the noreturn version of the next software interrupt >> handler. Fixes excessive stack usage when DPCs are queued in quick >> succession. >> CORE-11123 #resolve >> >> Added: >> trunk/reactos/hal/halx86/up/pic.S (with props) >> Modified: >> trunk/reactos/hal/halx86/pic.cmake >> trunk/reactos/hal/halx86/up/pic.c >> >> Modified: trunk/reactos/hal/halx86/pic.cmake >> URL: >> http://svn.reactos.org/svn/reactos/trunk/reactos/hal/halx86/pic.cmake?rev=7… >> >> ============================================================================== >> --- trunk/reactos/hal/halx86/pic.cmake [iso-8859-1] (original) >> +++ trunk/reactos/hal/halx86/pic.cmake [iso-8859-1] Tue May 10 15:03:56 >> 2016 >> @@ -1,7 +1,8 @@ >> list(APPEND HAL_PIC_ASM_SOURCE >> generic/systimer.S >> - generic/trap.S) >> + generic/trap.S >> + up/pic.S) >> list(APPEND HAL_PIC_SOURCE >> generic/profil.c >> >> Added: trunk/reactos/hal/halx86/up/pic.S >> URL: >> http://svn.reactos.org/svn/reactos/trunk/reactos/hal/halx86/up/pic.S?rev=71… >> >> ============================================================================== >> --- trunk/reactos/hal/halx86/up/pic.S (added) >> +++ trunk/reactos/hal/halx86/up/pic.S [iso-8859-1] Tue May 10 15:03:56 >> 2016 >> @@ -0,0 +1,48 @@ >> +/* >> + * FILE: hal/halx86/up/pic.S >> + * COPYRIGHT: See COPYING in the top level directory >> + * PURPOSE: HAL PIC Management and Control Code >> + * PROGRAMMER: Thomas Faber (thomas.faber(a)reactos.org) >> + */ >> + >> +/* INCLUDES >> ******************************************************************/ >> + >> +#include <asm.inc> >> + >> +#include <ks386.inc> >> + >> +EXTERN _HalpEndSoftwareInterrupt2@8:PROC >> + >> +/* GLOBALS >> *******************************************************************/ >> + >> +.data >> +ASSUME CS:NOTHING, DS:NOTHING, ES:NOTHING, FS:NOTHING, GS:NOTHING >> + >> +/* FUNCTIONS >> *****************************************************************/ >> + >> +.code >> +PUBLIC _HalpEndSoftwareInterrupt@8 >> +.PROC _HalpEndSoftwareInterrupt@8 >> + FPO 0, 2, 0, 0, 0, FRAME_FPO >> + >> + /* Call the C function with the same arguments we got */ >> + push [esp+8] >> + push [esp+8] >> + call _HalpEndSoftwareInterrupt2@8 >> + >> + /* Check if we got a pointer back */ >> + test eax, eax >> + jnz CallIntHandler >> + >> + /* No? Just return */ >> + ret 8 >> + >> +CallIntHandler: >> + /* We got a pointer to call. Since it won't return, free up our stack >> + space, or we could end up with some nasty deep recursion */ >> + mov ecx, [esp+8] >> + add esp, 12 >> + jmp eax >> +.ENDP >> + >> +END >> >> Propchange: trunk/reactos/hal/halx86/up/pic.S >> >> ------------------------------------------------------------------------------ >> svn:eol-style = native >> >> Modified: trunk/reactos/hal/halx86/up/pic.c >> URL: >> http://svn.reactos.org/svn/reactos/trunk/reactos/hal/halx86/up/pic.c?rev=71… >> >> ============================================================================== >> --- trunk/reactos/hal/halx86/up/pic.c [iso-8859-1] (original) >> +++ trunk/reactos/hal/halx86/up/pic.c [iso-8859-1] Tue May 10 15:03:56 >> 2016 >> @@ -11,6 +11,11 @@ >> #include <hal.h> >> #define NDEBUG >> #include <debug.h> >> + >> +VOID >> +NTAPI >> +HalpEndSoftwareInterrupt(IN KIRQL OldIrql, >> + IN PKTRAP_FRAME TrapFrame); >> /* GLOBALS >> ********************************************************************/ >> @@ -263,7 +268,7 @@ >> * so it will always preempt until we reach PROFILE_LEVEL. >> */ >> 0b00000000000000000001011111110000, /* IRQL 20 */ >> - 0b00000000000000000001001111110000, /* IRQL 20 */ >> + 0b00000000000000000001001111110000, /* IRQL 21 */ >> 0b00000000000000000001000111110000, /* IRQL 22 */ >> 0b00000000000000000001000011110000, /* IRQL 23 */ >> 0b00000000000000000001000001110000, /* IRQL 24 */ >> @@ -732,15 +737,17 @@ >> KeGetPcr()->IRR &= ~(1 << Irql); >> } >> -VOID >> -NTAPI >> -HalpEndSoftwareInterrupt(IN KIRQL OldIrql, >> - IN PKTRAP_FRAME TrapFrame) >> +PHAL_SW_INTERRUPT_HANDLER_2ND_ENTRY >> +NTAPI >> +HalpEndSoftwareInterrupt2(IN KIRQL OldIrql, >> + IN PKTRAP_FRAME TrapFrame) >> { >> ULONG PendingIrql, PendingIrqlMask, PendingIrqMask; >> PKPCR Pcr = KeGetPcr(); >> PIC_MASK Mask; >> + UNREFERENCED_PARAMETER(TrapFrame); >> + >> /* Set old IRQL */ >> Pcr->Irql = OldIrql; >> @@ -749,10 +756,10 @@ >> { >> /* Check for pending software interrupts and compare with >> current IRQL */ >> PendingIrqlMask = Pcr->IRR & FindHigherIrqlMask[OldIrql]; >> - if (!PendingIrqlMask) return; >> + if (!PendingIrqlMask) return NULL; >> /* Check for in-service delayed interrupt */ >> - if (Pcr->IrrActive & 0xFFFFFFF0) return; >> + if (Pcr->IrrActive & 0xFFFFFFF0) return NULL; >> /* Check if pending IRQL affects hardware state */ >> BitScanReverse(&PendingIrql, PendingIrqlMask); >> @@ -777,10 +784,11 @@ >> else >> { >> /* No need to loop checking for hardware interrupts */ >> - SWInterruptHandlerTable2[PendingIrql](TrapFrame); >> - UNREACHABLE; >> + return SWInterruptHandlerTable2[PendingIrql]; >> } >> } >> + >> + return NULL; >> } >> /* EDGE INTERRUPT DISMISSAL FUNCTIONS >> *****************************************/ >> >> >> > > > > _______________________________________________ > Ros-dev mailing list > Ros-dev(a)reactos.org > http://www.reactos.org/mailman/listinfo/ros-dev

9 years, 4 months

Re: [ros-dev] [ros-diffs] [hbelusca] 71339: [NTOS][NDK][RTL]: Move the SEM_xxx flags and their corresponding RTL flags around so that they can be accessed in kernel code. I'll need them soon.

by Alex Ionescu

On Mon, May 16, 2016 at 12:40 AM, <hbelusca(a)svn.reactos.org> wrote: > +// > +// Thread Error Mode Flags > +// > +/* Also defined in psdk/winbase.h */ > +#define SEM_FAILCRITICALERRORS 0x0001 > +#define SEM_NOGPFAULTERRORBOX 0x0002 > +#define SEM_NOALIGNMENTFAULTEXCEPT 0x0004 > +#define SEM_NOOPENFILEERRORBOX 0x8000 Hermes, you consistently ignore the formatting rules of other people's code/project. Please correctly format the NDK flags instead of just copy-pasting. Best regards, Alex Ionescu

9 years, 4 months

Re: [ros-dev] [ros-diffs] [hbelusca] 71336: [NTOS]: Remove hardcoded constants; fix a comment; move some code around to reduce the level of code indentation.

by Alex Ionescu

On Mon, May 16, 2016 at 12:23 AM, <hbelusca(a)svn.reactos.org> wrote: > - if ((!ExReadyForErrors) && (NT_ERROR(ErrorStatus))) > + if (!ExReadyForErrors && NT_ERROR(ErrorStatus)) Let's not start this again. Our coding conventions (and good programming practice) is to have every conditional in its own paren blob. Best regards, Alex Ionescu

9 years, 4 months

by Timo Kreuzer

Wait, are you suggesting that "doing everything the same as MS does" and "I don't need to care for C standards, I have a huge epeen!" and "OnlyOnePersonCanWriteHalCode AndItsNotYou".... doesn't work? :O Best regards, Timo Kreuzer Am 10.05.2016 um 17:03 schrieb tfaber(a)svn.reactos.org: > Author: tfaber > Date: Tue May 10 15:03:56 2016 > New Revision: 71307 > > URL: http://svn.reactos.org/svn/reactos?rev=71307&view=rev > Log: > [HAL] > - Create a wrapper version of HalpEndSoftwareInterrupt that frees its stack before calling the noreturn version of the next software interrupt handler. Fixes excessive stack usage when DPCs are queued in quick succession. > CORE-11123 #resolve > > Added: > trunk/reactos/hal/halx86/up/pic.S (with props) > Modified: > trunk/reactos/hal/halx86/pic.cmake > trunk/reactos/hal/halx86/up/pic.c > > Modified: trunk/reactos/hal/halx86/pic.cmake > URL: http://svn.reactos.org/svn/reactos/trunk/reactos/hal/halx86/pic.cmake?rev=7… > ============================================================================== > --- trunk/reactos/hal/halx86/pic.cmake [iso-8859-1] (original) > +++ trunk/reactos/hal/halx86/pic.cmake [iso-8859-1] Tue May 10 15:03:56 2016 > @@ -1,7 +1,8 @@ > > list(APPEND HAL_PIC_ASM_SOURCE > generic/systimer.S > - generic/trap.S) > + generic/trap.S > + up/pic.S) > > list(APPEND HAL_PIC_SOURCE > generic/profil.c > > Added: trunk/reactos/hal/halx86/up/pic.S > URL: http://svn.reactos.org/svn/reactos/trunk/reactos/hal/halx86/up/pic.S?rev=71… > ============================================================================== > --- trunk/reactos/hal/halx86/up/pic.S (added) > +++ trunk/reactos/hal/halx86/up/pic.S [iso-8859-1] Tue May 10 15:03:56 2016 > @@ -0,0 +1,48 @@ > +/* > + * FILE: hal/halx86/up/pic.S > + * COPYRIGHT: See COPYING in the top level directory > + * PURPOSE: HAL PIC Management and Control Code > + * PROGRAMMER: Thomas Faber (thomas.faber(a)reactos.org) > + */ > + > +/* INCLUDES ******************************************************************/ > + > +#include <asm.inc> > + > +#include <ks386.inc> > + > +EXTERN _HalpEndSoftwareInterrupt2@8:PROC > + > +/* GLOBALS *******************************************************************/ > + > +.data > +ASSUME CS:NOTHING, DS:NOTHING, ES:NOTHING, FS:NOTHING, GS:NOTHING > + > +/* FUNCTIONS *****************************************************************/ > + > +.code > +PUBLIC _HalpEndSoftwareInterrupt@8 > +.PROC _HalpEndSoftwareInterrupt@8 > + FPO 0, 2, 0, 0, 0, FRAME_FPO > + > + /* Call the C function with the same arguments we got */ > + push [esp+8] > + push [esp+8] > + call _HalpEndSoftwareInterrupt2@8 > + > + /* Check if we got a pointer back */ > + test eax, eax > + jnz CallIntHandler > + > + /* No? Just return */ > + ret 8 > + > +CallIntHandler: > + /* We got a pointer to call. Since it won't return, free up our stack > + space, or we could end up with some nasty deep recursion */ > + mov ecx, [esp+8] > + add esp, 12 > + jmp eax > +.ENDP > + > +END > > Propchange: trunk/reactos/hal/halx86/up/pic.S > ------------------------------------------------------------------------------ > svn:eol-style = native > > Modified: trunk/reactos/hal/halx86/up/pic.c > URL: http://svn.reactos.org/svn/reactos/trunk/reactos/hal/halx86/up/pic.c?rev=71… > ============================================================================== > --- trunk/reactos/hal/halx86/up/pic.c [iso-8859-1] (original) > +++ trunk/reactos/hal/halx86/up/pic.c [iso-8859-1] Tue May 10 15:03:56 2016 > @@ -11,6 +11,11 @@ > #include <hal.h> > #define NDEBUG > #include <debug.h> > + > +VOID > +NTAPI > +HalpEndSoftwareInterrupt(IN KIRQL OldIrql, > + IN PKTRAP_FRAME TrapFrame); > > /* GLOBALS ********************************************************************/ > > @@ -263,7 +268,7 @@ > * so it will always preempt until we reach PROFILE_LEVEL. > */ > 0b00000000000000000001011111110000, /* IRQL 20 */ > - 0b00000000000000000001001111110000, /* IRQL 20 */ > + 0b00000000000000000001001111110000, /* IRQL 21 */ > 0b00000000000000000001000111110000, /* IRQL 22 */ > 0b00000000000000000001000011110000, /* IRQL 23 */ > 0b00000000000000000001000001110000, /* IRQL 24 */ > @@ -732,15 +737,17 @@ > KeGetPcr()->IRR &= ~(1 << Irql); > } > > -VOID > -NTAPI > -HalpEndSoftwareInterrupt(IN KIRQL OldIrql, > - IN PKTRAP_FRAME TrapFrame) > +PHAL_SW_INTERRUPT_HANDLER_2ND_ENTRY > +NTAPI > +HalpEndSoftwareInterrupt2(IN KIRQL OldIrql, > + IN PKTRAP_FRAME TrapFrame) > { > ULONG PendingIrql, PendingIrqlMask, PendingIrqMask; > PKPCR Pcr = KeGetPcr(); > PIC_MASK Mask; > > + UNREFERENCED_PARAMETER(TrapFrame); > + > /* Set old IRQL */ > Pcr->Irql = OldIrql; > > @@ -749,10 +756,10 @@ > { > /* Check for pending software interrupts and compare with current IRQL */ > PendingIrqlMask = Pcr->IRR & FindHigherIrqlMask[OldIrql]; > - if (!PendingIrqlMask) return; > + if (!PendingIrqlMask) return NULL; > > /* Check for in-service delayed interrupt */ > - if (Pcr->IrrActive & 0xFFFFFFF0) return; > + if (Pcr->IrrActive & 0xFFFFFFF0) return NULL; > > /* Check if pending IRQL affects hardware state */ > BitScanReverse(&PendingIrql, PendingIrqlMask); > @@ -777,10 +784,11 @@ > else > { > /* No need to loop checking for hardware interrupts */ > - SWInterruptHandlerTable2[PendingIrql](TrapFrame); > - UNREACHABLE; > + return SWInterruptHandlerTable2[PendingIrql]; > } > } > + > + return NULL; > } > > /* EDGE INTERRUPT DISMISSAL FUNCTIONS *****************************************/ > > >

9 years, 4 months

Abuse filter bans legit users and ignores spamers

by Alexander Rechitskiy

9 years, 4 months

https://code.reactos.org/ is down

by Alexander Rechitskiy

9 years, 4 months

CAPCHA does not work

by Alexander Rechitskiy

9 years, 4 months

Build Crash ProgMan

by James Tabor

Linux build, on all my systems, Slackware, Fedora,,,,, Building C object base/shell/progman/progman_pch.h.gch FAILED: /usr/local/RosBE/i386/bin/i686-w64-mingw32-gcc -DDBG=1 -DKDBG=1 -DUNICODE -DUSE_COMPILER_EXCEPTIONS -DWINVER=0x502 -D_DLL -D_M_IX86 -D_SEH_ENABLE_TRACE -D_SETUPAPI_VER=0x502 -D_UNICODE -D_USE_32BIT_TIME_T -D_USE_PSEH3=1 -D_WIN32_IE=0x600 -D_WIN32_WINDOWS=0x502 -D_WIN32_WINNT=0x502 -D_X86_ -D__REACTOS__ -D__USE_CRTIMP -D__i386__ -D_inline=__inline -Di386 -Wa,--compress-debug-sections -pipe -fms-extensions -fno-strict-aliasing -nostdinc -mstackrealign -Wold-style-declaration -Wdeclaration-after-statement -fdebug-prefix-map="/home/ros/reactos"=ReactOS -gdwarf-2 -gstrict-dwarf -femit-struct-debug-detailed=none -feliminate-unused-debug-symbols -march=pentium -mtune=i686 -Werror -Wall -Wpointer-arith -Wno-char-subscripts -Wno-multichar -Wno-unused-value -Wno-maybe-uninitialized -O1 -fno-optimize-sibling-calls -fno-omit-frame-pointer -mpreferred-stack-boundary=3 -fno-set-stack-executable -Winvalid-pch -Werror=invalid-pch -Ibase/shell/progman -I../../base/shell/progman -I../../sdk/include -I../../sdk/include/psdk -I../../sdk/include/dxsdk -Isdk/include -Isdk/include/psdk -Isdk/include/dxsdk -Isdk/include/ddk -Isdk/include/reactos -I../../sdk/include/crt -I../../sdk/include/ddk -I../../sdk/include/ndk -I../../sdk/include/reactos -I../../sdk/include/reactos/libs -MMD -MT base/shell/progman/progman_pch.h.gch -MF base/shell/progman/progman_pch.h.gch.d -o base/shell/progman/progman_pch.h.gch -c ../../base/shell/progman/progman.h ../../base/shell/progman/progman.h:35:9: error: #pragma once in main file [-Werror] cc1: all warnings being treated as errors [206/1104] Linking C executable base/shell/cmd/cmd.exe ninja: build stopped: subcommand failed.

9 years, 4 months

Spam attack, continuing... (episode 12345/infinity)

by Hermès BÉLUSCA - MAÏTO

So, guys, ... We are yet again subject to a spam attack on the forums AND in the wiki (since yesterday for the latter), but nothing was done yet !! https://reactos.org/forum/ https://reactos.org/wiki/Special:RecentChanges For the wiki, the abuse filter hasn't blocked everything !! Regards, Hermès

9 years, 4 months

Google Summer of Code

by Zuodian Hu

Dear ReactOS I am zhu48 from the GSoC, on the network driver project. I'm sorry I am sending this email out so late after the start of the official bonding period. I am going into finals week and I had some trouble getting the mailing list to work. Hello everybody! Zuodian Hu

9 years, 4 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