Ros-dev

ros-dev@reactos.org

1 participants
5791 discussions

Serial port IRQ detection
by Mark Junker 21 Mar '05

21 Mar '05

Hi, is there a requirement for the serial port IRQ detection? This was also part of my old sources but currently I simply don't know how to do it inside the ReactOS serial driver. BTW: How does the "serenum" stuff works? Regards, Mark

2 1

Re: [ros-dev] Update reported version to Windows 2000 (Service Pack?)
by Steven Edwards 21 Mar '05

21 Mar '05

--- art yerkes <ayerkes(a)speakeasy.net> wrote: > We need a convenient way to report a different version on a per application > basis. I think there's some code to do this but I could not determine > whether it works (despite several people telling me how to do it). While > we're at it, can we verify that this code works and post a howto about > reporting a specific windows version to an application? I don't know much about it but it is supposed to be possible. There is a app compatiblity shell extention that Thomas wrote that is supposed to do it under Windows. Anyone know if this really works under ReactOS? Thanks Steven __________________________________ Do you Yahoo!? Make Yahoo! your home page http://www.yahoo.com/r/hs

1 0

Serial port type detection
by Mark Junker 20 Mar '05

20 Mar '05

Hi, I found some old sources from me where I detect the IRQ and the type (8250, 16450, 16550, 16550A) of a COM port. I converted the type detection sources to C. The detection if there is a COM port at all is made by inverting all bits of the LCR twice. All chips > 8250 have a scratch pad and all 16550A+ have a usable FIFO. I attached the diff for reactos/drivers/dd/serial/legacy.c. BTW: I'm sorry about it but my editor always changes TABs to spaces ... Regards, Mark Index: drivers/dd/serial/legacy.c =================================================================== --- drivers/dd/serial/legacy.c (revision 14202) +++ drivers/dd/serial/legacy.c (working copy) @@ -1,72 +1,101 @@ /* $Id: * * COPYRIGHT: See COPYING in the top level directory * PROJECT: ReactOS kernel * FILE: drivers/bus/serial/legacy.c * PURPOSE: Legacy serial port enumeration * * PROGRAMMERS: Hervé Poussineau (poussine(a)freesurf.fr) */ #define NDEBUG #include "serial.h" +#define SER_TYPE_NONE 0 +#define SER_TYPE_8250 1 +#define SER_TYPE_16450 2 +#define SER_TYPE_16550 3 +#define SER_TYPE_16550A 4 + +static BYTE +DetectPortType(PUCHAR BaseAddress) +{ + BYTE Lcr, TestLcr; + BYTE OldScr, Scr5A, ScrA5; + BOOLEAN FifoEnabled; + BYTE NewFifoStatus; + + Lcr = READ_PORT_UCHAR(SER_LCR(BaseAddress)); + WRITE_PORT_UCHAR(BaseAddress, (BYTE) (Lcr ^ 0xFF)); + TestLcr = (BYTE) (READ_PORT_UCHAR(SER_LCR(BaseAddress)) ^ 0xFF); + WRITE_PORT_UCHAR(BaseAddress, Lcr); + + /* Accessing the LCR must work for a usable serial port */ + if (TestLcr!=Lcr) + return SER_TYPE_NONE; + + /* Ensure that all following accesses are done as required */ + READ_PORT_UCHAR(SER_RBR(BaseAddress)); + READ_PORT_UCHAR(SER_IER(BaseAddress)); + READ_PORT_UCHAR(SER_IIR(BaseAddress)); + READ_PORT_UCHAR(SER_LCR(BaseAddress)); + READ_PORT_UCHAR(SER_MCR(BaseAddress)); + READ_PORT_UCHAR(SER_LSR(BaseAddress)); + READ_PORT_UCHAR(SER_MSR(BaseAddress)); + READ_PORT_UCHAR(SER_SCR(BaseAddress)); + + /* Test scratch pad */ + OldScr = READ_PORT_UCHAR(SER_SCR(BaseAddress)); + WRITE_PORT_UCHAR(SER_SCR(BaseAddress), 0x5A); + Scr5A = READ_PORT_UCHAR(SER_SCR(BaseAddress)); + WRITE_PORT_UCHAR(SER_SCR(BaseAddress), 0xA5); + ScrA5 = READ_PORT_UCHAR(SER_SCR(BaseAddress)); + WRITE_PORT_UCHAR(SER_SCR(BaseAddress), OldScr); + + /* When non-functional, we have a 8250 */ + if (Scr5A!=0x5A || ScrA5!=0xA5) + return SER_TYPE_8250; + + /* Test FIFO type */ + FifoEnabled = (READ_PORT_UCHAR(SER_IIR(BaseAddress)) & 0x80)!=0; + WRITE_PORT_UCHAR(SER_FCR(BaseAddress), SR_FCR_ENABLE_FIFO); + NewFifoStatus = READ_PORT_UCHAR(SER_IIR(BaseAddress)) & 0xC0; + if (!FifoEnabled) + WRITE_PORT_UCHAR(SER_FCR(BaseAddress), 0); + switch (NewFifoStatus) { + case 0x00: + return SER_TYPE_16450; + case 0x80: + return SER_TYPE_16550; + } + + /* FIFO is only functional for 16550A+ */ + return SER_TYPE_16550A; +} + static BOOLEAN SerialDoesPortExist(PUCHAR BaseAddress) { - BOOLEAN Found; - BYTE Mcr; - BYTE Msr; - - Found = FALSE; - - /* save Modem Control Register (MCR) */ - Mcr = READ_PORT_UCHAR(SER_MCR(BaseAddress)); - - /* enable loop mode (set Bit 4 of the MCR) */ - WRITE_PORT_UCHAR(SER_MCR(BaseAddress), 0x10); - - /* clear all modem output bits */ - WRITE_PORT_UCHAR(SER_MCR(BaseAddress), 0x10); - - /* read the Modem Status Register */ - Msr = READ_PORT_UCHAR(SER_MSR(BaseAddress)); - - /* - * the upper nibble of the MSR (modem output bits) must be - * equal to the lower nibble of the MCR (modem input bits) - */ - if ((Msr & 0xf0) == 0x00) - { - /* set all modem output bits */ - WRITE_PORT_UCHAR(SER_MCR(BaseAddress), 0x1f); - - /* read the Modem Status Register */ - Msr = READ_PORT_UCHAR(SER_MSR(BaseAddress)); - - /* - * the upper nibble of the MSR (modem output bits) must be - * equal to the lower nibble of the MCR (modem input bits) - */ - if ((Msr & 0xf0) == 0xf0) - { - /* - * setup a resonable state for the port: - * enable fifo and clear recieve/transmit buffers - */ - WRITE_PORT_UCHAR(SER_FCR(BaseAddress), - (SR_FCR_ENABLE_FIFO | SR_FCR_CLEAR_RCVR | SR_FCR_CLEAR_XMIT)); - WRITE_PORT_UCHAR(SER_FCR(BaseAddress), 0); - READ_PORT_UCHAR(SER_RBR(BaseAddress)); - WRITE_PORT_UCHAR(SER_IER(BaseAddress), 0); - Found = TRUE; - } - } - - /* restore MCR */ - WRITE_PORT_UCHAR(SER_MCR(BaseAddress), Mcr); - - return Found; + BYTE Type; + + Type = DetectPortType(BaseAddress); + if (Type==SER_TYPE_NONE) + return FALSE; + + if (Type==SER_TYPE_16550A) { + /* + * setup a resonable state for the port: + * enable fifo and clear recieve/transmit buffers + */ + WRITE_PORT_UCHAR(SER_FCR(BaseAddress), + (SR_FCR_ENABLE_FIFO | SR_FCR_CLEAR_RCVR | SR_FCR_CLEAR_XMIT)); + WRITE_PORT_UCHAR(SER_FCR(BaseAddress), 0); + } + + READ_PORT_UCHAR(SER_RBR(BaseAddress)); + WRITE_PORT_UCHAR(SER_IER(BaseAddress), 0); + + return TRUE; } NTSTATUS

2 2

Update reported version to Windows 2000 (Service Pack?)
by Steven Edwards 20 Mar '05

20 Mar '05

Hi, Most of the dev team seems to be in agreement with bumping the reported version number to Win2k but if anyone has any objections speak now or forever maintain a patch. I need Win2k to be reported as the version number as many newer applications such as Office 2003 will not install unless that is the latest version. If you find any places where a function, resource or anything reports a value as NT4 or something ReactOS dependant please change it to match Windows 2000 behavior. Also I propose we just report Service Pack 4. Thanks Steven __________________________________ Do you Yahoo!? Yahoo! Small Business - Try our new resources site! http://smallbusiness.yahoo.com/resources/

6 7

BootCD not Formatting
by Jason Filby 20 Mar '05

20 Mar '05

Hi all Currently when running a BootCD, if I ask to format the partition, I'm told that the setup could not complete. This is on trunk, not sure about the 0.2.6 build; can anyone confirm? Thanks Jason

2 1

Explorer compile bug
by Wesley Parish 20 Mar '05

20 Mar '05

This is the text of the final few seconds before the compilation ran out of steam. The compiler is the standard MinGW32-gcc plus MinGW-g++, etc, in the MinGW-3.1.0-1.exe package; the computer's a Compaq 4000 Deskpro; the OS is MS Win95, bog-standard as far as I know. Wesley Parish C:/MINGW/include/c++/3.2.3/bits/stl_tree.h:1068: instantiated from `std::_Rb_t ree_iterator<_Val, _Val&, _Val*> std::_Rb_tree<_Key, _Val, _KeyOfValue, _Compare , _Alloc>::insert_unique(std::_Rb_tree_iterator<_Val, _Val&, _Val*>, const _Val& ) [with _Key = String, _Val = std::pair<const String, ICON_ID>, _KeyOfValue = st d::_Select1st<std::pair<const String, ICON_ID> >, _Compare = std::less<String>, _Alloc = std::allocator<std::pair<const String, ICON_ID> >]' C:/MINGW/include/c++/3.2.3/bits/stl_map.h:262: instantiated from `std::_Rb_tre e<_Key, std::pair<const _Key, _Tp>, std::_Select1st<std::pair<const _Key, _Tp> > , _Compare, _Alloc>::iterator std::map<_Key, _Tp, _Compare, _Alloc>::insert(std: :_Rb_tree<_Key, std::pair<const _Key, _Tp>, std::_Select1st<std::pair<const _Key , _Tp> >, _Compare, _Alloc>::iterator, const std::pair<const _Key, _Tp>&) [with _Key = String, _Tp = ICON_ID, _Compare = std::less<String>, _Alloc = std::alloca tor<std::pair<const String, ICON_ID> >]' C:/MINGW/include/c++/3.2.3/bits/stl_map.h:225: instantiated from `_Tp& std::ma p<_Key, _Tp, _Compare, _Alloc>::operator[](const _Key&) [with _Key = String, _Tp = ICON_ID, _Compare = std::less<String>, _Alloc = std::allocator<std::pair<cons t String, ICON_ID> >]' explorer.cpp:398: instantiated from here C:/MINGW/include/c++/3.2.3/bits/stl_pair.h:88: invalid use of undefined type ` struct String' utility/utility.h:737: forward declaration of `struct String' make[1]: *** [explorer.o] Error 1 make: *** [explorer] Error 2 -- Clinersterton beademung, with all of love - RIP James Blish ----- Mau e ki, he aha te mea nui? You ask, what is the most important thing? Maku e ki, he tangata, he tangata, he tangata. I reply, it is people, it is people, it is people.

2 2

Re: [ros-dev] BootCD not Formatting
by Sylvain Petreolle 20 Mar '05

20 Mar '05

same problem here using trunk revision 14208. tested on qemu using blank and already formated image. --- Jason Filby <jason.filby(a)gmail.com> wrote: > Hi all > > Currently when running a BootCD, if I ask to format the partition, I'm > told that the setup could not complete. This is on trunk, not sure > about the 0.2.6 build; can anyone confirm? > > Thanks > Jason Kind regards, Usurp (aka Sylvain Petreolle) humans are like computers, yesterday the BIOS was all - today its just a word

2 1

Crash at the end of Stage2 setup (2 systems)
by Steven Edwards 20 Mar '05

20 Mar '05

Hi, I think the affects both the branch and the trunk. At the end of stage 2 setup I get a crash in ntoskrnl: 319b and 6760c According to my map. 8006760c <_NtAdjustPrivilegesToken@24+0xbc> Is called in a few places and 319b is /* Do the System Call */ call *%eax 80003199: ff d0 call *%eax movl %eax, KTRAP_FRAME_EAX(%ebp) 8000319b: 89 45 44 mov %eax,0x44(%ebp) Thanks Steven __________________________________ Do you Yahoo!? Yahoo! Small Business - Try our new resources site! http://smallbusiness.yahoo.com/resources/

3 2

Please FIX: Wrong prototype and implementation in Rtl
by Oliver Schneider 20 Mar '05

20 Mar '05

.\lib\rtl\largeint.c There's a sutraction performed in your implementation. But actually it's a logical negation (asm -> NEG; C -> ~) in this function of the low and high part: RtlLargeIntegerNegate() Oliver PS: I have no CVS access, so anyone here having it should fix this.

2 2

New SM + CSR exposes LPC bug
by ea 19 Mar '05

19 Mar '05

I created bug 585 for a blocking bug in LPC message queueing. It seams I can't fix it. :( There is something wrong in lpc/queue.c. New SM that loads subsystems reading the registry config and CSR that registers are attached to 585 (code *not* commited yet). If fixing the LPC bug is not possible easily, I'll change SM to use a scond thread to avoid cross calling from the same thread. ea

1 0

Jump to page:

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Ros-dev