sortix-mirror/kernel/textterminal.cpp

920 lines
23 KiB
C++
Raw Permalink Normal View History

/*
* Copyright (c) 2011, 2012, 2013, 2014, 2015, 2016 Jonas 'Sortie' Termansen.
* Copyright (c) 2023 Juhani 'nortti' Krekelä.
*
* Permission to use, copy, modify, and distribute this software for any
* purpose with or without fee is hereby granted, provided that the above
* copyright notice and this permission notice appear in all copies.
*
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
*
* textterminal.cpp
* Translates a character stream to a 2 dimensional array of characters.
*/
#include <stdio.h>
#include <string.h>
#include <wchar.h>
2013-10-27 00:42:10 +00:00
#include <sortix/vga.h>
#include <sortix/kernel/kernel.h>
#include <sortix/kernel/refcount.h>
#include <sortix/kernel/textbuffer.h>
2013-10-27 00:42:10 +00:00
#include "palette.h"
#include "textterminal.h"
namespace Sortix {
static const uint16_t DEFAULT_VGACOLOR = COLOR8_LIGHT_GREY | COLOR8_BLACK << 4;
static const unsigned int DEFAULT_FOREGROUND = 7;
static const unsigned int DEFAULT_BACKGROUND = 0;
static uint32_t ColorFromRGB(uint8_t r, uint8_t g, uint8_t b)
{
return (uint32_t) b << 0 | (uint32_t) g << 8 | (uint32_t) r << 16;
}
2015-03-19 13:27:28 +00:00
TextTerminal::TextTerminal(TextBufferHandle* textbufhandle)
{
memset(&ps, 0, sizeof(ps));
this->textbufhandle = textbufhandle;
Multithreaded kernel and improvement of signal handling. Pardon the big ass-commit, this took months to develop and debug and the refactoring got so far that a clean merge became impossible. The good news is that this commit does quite a bit of cleaning up and generally improves the kernel quality. This makes the kernel fully pre-emptive and multithreaded. This was done by rewriting the interrupt code, the scheduler, introducing new threading primitives, and rewriting large parts of the kernel. During the past few commits the kernel has had its device drivers thread secured; this commit thread secures large parts of the core kernel. There still remains some parts of the kernel that is _not_ thread secured, but this is not a problem at this point. Each user-space thread has an associated kernel stack that it uses when it goes into kernel mode. This stack is by default 8 KiB since that value works for me and is also used by Linux. Strange things tends to happen on x86 in case of a stack overflow - there is no ideal way to catch such a situation right now. The system call conventions were changed, too. The %edx register is now used to provide the errno value of the call, instead of the kernel writing it into a registered global variable. The system call code has also been updated to better reflect the native calling conventions: not all registers have to be preserved. This makes system calls faster and simplifies the assembly. In the kernel, there is no longer the event.h header or the hacky method of 'resuming system calls' that closely resembles cooperative multitasking. If a system call wants to block, it should just block. The signal handling was also improved significantly. At this point, signals cannot interrupt kernel threads (but can always interrupt user-space threads if enabled), which introduces some problems with how a SIGINT could interrupt a blocking read, for instance. This commit introduces and uses a number of new primitives such as kthread_lock_mutex_signal() that attempts to get the lock but fails if a signal is pending. In this manner, the kernel is safer as kernel threads cannot be shut down inconveniently, but in return for complexity as blocking operations must check they if they should fail. Process exiting has also been refactored significantly. The _exit(2) system call sets the exit code and sends SIGKILL to all the threads in the process. Once all the threads have cleaned themselves up and exited, a worker thread calls the process's LastPrayer() method that unmaps memory, deletes the address space, notifies the parent, etc. This provides a very robust way to terminate processes as even half-constructed processes (during a failing fork for instance) can be gracefully terminated. I have introduced a number of kernel threads to help avoid threading problems and simplify kernel design. For instance, there is now a functional generic kernel worker thread that any kernel thread can schedule jobs for. Interrupt handlers run with interrupts off (hence they cannot call kthread_ functions as it may deadlock the system if another thread holds the lock) therefore they cannot use the standard kernel worker threads. Instead, they use a special purpose interrupt worker thread that works much like the generic one expect that interrupt handlers can safely queue work with interrupts off. Note that this also means that interrupt handlers cannot allocate memory or print to the kernel log/screen as such mechanisms uses locks. I'll introduce a lock free algorithm for such cases later on. The boot process has also changed. The original kernel init thread in kernel.cpp creates a new bootstrap thread and becomes the system idle thread. Note that pid=0 now means the kernel, as there is no longer a system idle process. The bootstrap thread launches all the kernel worker threads and then creates a new process and loads /bin/init into it and then creates a thread in pid=1, which starts the system. The bootstrap thread then quietly waits for pid=1 to exit after which it shuts down/reboots/panics the system. In general, the introduction of race conditions and dead locks have forced me to revise a lot of the design and make sure it was thread secure. Since early parts of the kernel was quite hacky, I had to refactor such code. So it seems that the risk of dead locks forces me to write better code. Note that a real preemptive multithreaded kernel simplifies the construction of blocking system calls. My hope is that this will trigger a clean up of the filesystem code that current is almost beyond repair. Almost all of the kernel was modified during this refactoring. To the extent possible, these changes have been backported to older non-multithreaded kernel, but many changes were tightly coupled and went into this commit. Of interest is the implementation of the kthread_ api based on the design of pthreads; this library allows easy synchronization mechanisms and includes C++-style scoped locks. This commit also introduces new worker threads and tested mechanisms for interrupt handlers to schedule work in a kernel worker thread. A lot of code have been rewritten from scratch and has become a lot more stable and correct. Share and enjoy!
2012-08-01 15:30:34 +00:00
this->termlock = KTHREAD_MUTEX_INITIALIZER;
Reset();
}
2012-09-30 11:09:25 +00:00
TextTerminal::~TextTerminal()
{
}
void TextTerminal::Reset()
{
attr = 0;
next_attr = 0;
vgacolor = DEFAULT_VGACOLOR;
fgcolor = palette[DEFAULT_FOREGROUND];
bgcolor = palette[DEFAULT_BACKGROUND];
column = line = 0;
ansisavedposx = ansisavedposy = 0;
ansimode = NONE;
TextBuffer* textbuf = textbufhandle->Acquire();
TextPos fillfrom(0, 0);
TextPos fillto(textbuf->Width()-1, textbuf->Height()-1);
TextChar fillwith(' ', vgacolor, 0, fgcolor, bgcolor);
textbuf->Fill(fillfrom, fillto, fillwith);
textbuf->SetCursorEnabled(true);
UpdateCursor(textbuf);
textbufhandle->Release(textbuf);
}
size_t TextTerminal::Print(const char* string, size_t stringlen)
{
ScopedLock lock(&termlock);
TextBuffer* textbuf = textbufhandle->Acquire();
for ( size_t i = 0; i < stringlen; i++ )
{
if ( string[i] == '\n' )
PutChar(textbuf, '\r');
PutChar(textbuf, string[i]);
}
UpdateCursor(textbuf);
textbufhandle->Release(textbuf);
return stringlen;
}
size_t TextTerminal::PrintRaw(const char* string, size_t stringlen)
{
Multithreaded kernel and improvement of signal handling. Pardon the big ass-commit, this took months to develop and debug and the refactoring got so far that a clean merge became impossible. The good news is that this commit does quite a bit of cleaning up and generally improves the kernel quality. This makes the kernel fully pre-emptive and multithreaded. This was done by rewriting the interrupt code, the scheduler, introducing new threading primitives, and rewriting large parts of the kernel. During the past few commits the kernel has had its device drivers thread secured; this commit thread secures large parts of the core kernel. There still remains some parts of the kernel that is _not_ thread secured, but this is not a problem at this point. Each user-space thread has an associated kernel stack that it uses when it goes into kernel mode. This stack is by default 8 KiB since that value works for me and is also used by Linux. Strange things tends to happen on x86 in case of a stack overflow - there is no ideal way to catch such a situation right now. The system call conventions were changed, too. The %edx register is now used to provide the errno value of the call, instead of the kernel writing it into a registered global variable. The system call code has also been updated to better reflect the native calling conventions: not all registers have to be preserved. This makes system calls faster and simplifies the assembly. In the kernel, there is no longer the event.h header or the hacky method of 'resuming system calls' that closely resembles cooperative multitasking. If a system call wants to block, it should just block. The signal handling was also improved significantly. At this point, signals cannot interrupt kernel threads (but can always interrupt user-space threads if enabled), which introduces some problems with how a SIGINT could interrupt a blocking read, for instance. This commit introduces and uses a number of new primitives such as kthread_lock_mutex_signal() that attempts to get the lock but fails if a signal is pending. In this manner, the kernel is safer as kernel threads cannot be shut down inconveniently, but in return for complexity as blocking operations must check they if they should fail. Process exiting has also been refactored significantly. The _exit(2) system call sets the exit code and sends SIGKILL to all the threads in the process. Once all the threads have cleaned themselves up and exited, a worker thread calls the process's LastPrayer() method that unmaps memory, deletes the address space, notifies the parent, etc. This provides a very robust way to terminate processes as even half-constructed processes (during a failing fork for instance) can be gracefully terminated. I have introduced a number of kernel threads to help avoid threading problems and simplify kernel design. For instance, there is now a functional generic kernel worker thread that any kernel thread can schedule jobs for. Interrupt handlers run with interrupts off (hence they cannot call kthread_ functions as it may deadlock the system if another thread holds the lock) therefore they cannot use the standard kernel worker threads. Instead, they use a special purpose interrupt worker thread that works much like the generic one expect that interrupt handlers can safely queue work with interrupts off. Note that this also means that interrupt handlers cannot allocate memory or print to the kernel log/screen as such mechanisms uses locks. I'll introduce a lock free algorithm for such cases later on. The boot process has also changed. The original kernel init thread in kernel.cpp creates a new bootstrap thread and becomes the system idle thread. Note that pid=0 now means the kernel, as there is no longer a system idle process. The bootstrap thread launches all the kernel worker threads and then creates a new process and loads /bin/init into it and then creates a thread in pid=1, which starts the system. The bootstrap thread then quietly waits for pid=1 to exit after which it shuts down/reboots/panics the system. In general, the introduction of race conditions and dead locks have forced me to revise a lot of the design and make sure it was thread secure. Since early parts of the kernel was quite hacky, I had to refactor such code. So it seems that the risk of dead locks forces me to write better code. Note that a real preemptive multithreaded kernel simplifies the construction of blocking system calls. My hope is that this will trigger a clean up of the filesystem code that current is almost beyond repair. Almost all of the kernel was modified during this refactoring. To the extent possible, these changes have been backported to older non-multithreaded kernel, but many changes were tightly coupled and went into this commit. Of interest is the implementation of the kthread_ api based on the design of pthreads; this library allows easy synchronization mechanisms and includes C++-style scoped locks. This commit also introduces new worker threads and tested mechanisms for interrupt handlers to schedule work in a kernel worker thread. A lot of code have been rewritten from scratch and has become a lot more stable and correct. Share and enjoy!
2012-08-01 15:30:34 +00:00
ScopedLock lock(&termlock);
TextBuffer* textbuf = textbufhandle->Acquire();
for ( size_t i = 0; i < stringlen; i++ )
PutChar(textbuf, string[i]);
UpdateCursor(textbuf);
textbufhandle->Release(textbuf);
return stringlen;
}
size_t TextTerminal::Width()
{
Multithreaded kernel and improvement of signal handling. Pardon the big ass-commit, this took months to develop and debug and the refactoring got so far that a clean merge became impossible. The good news is that this commit does quite a bit of cleaning up and generally improves the kernel quality. This makes the kernel fully pre-emptive and multithreaded. This was done by rewriting the interrupt code, the scheduler, introducing new threading primitives, and rewriting large parts of the kernel. During the past few commits the kernel has had its device drivers thread secured; this commit thread secures large parts of the core kernel. There still remains some parts of the kernel that is _not_ thread secured, but this is not a problem at this point. Each user-space thread has an associated kernel stack that it uses when it goes into kernel mode. This stack is by default 8 KiB since that value works for me and is also used by Linux. Strange things tends to happen on x86 in case of a stack overflow - there is no ideal way to catch such a situation right now. The system call conventions were changed, too. The %edx register is now used to provide the errno value of the call, instead of the kernel writing it into a registered global variable. The system call code has also been updated to better reflect the native calling conventions: not all registers have to be preserved. This makes system calls faster and simplifies the assembly. In the kernel, there is no longer the event.h header or the hacky method of 'resuming system calls' that closely resembles cooperative multitasking. If a system call wants to block, it should just block. The signal handling was also improved significantly. At this point, signals cannot interrupt kernel threads (but can always interrupt user-space threads if enabled), which introduces some problems with how a SIGINT could interrupt a blocking read, for instance. This commit introduces and uses a number of new primitives such as kthread_lock_mutex_signal() that attempts to get the lock but fails if a signal is pending. In this manner, the kernel is safer as kernel threads cannot be shut down inconveniently, but in return for complexity as blocking operations must check they if they should fail. Process exiting has also been refactored significantly. The _exit(2) system call sets the exit code and sends SIGKILL to all the threads in the process. Once all the threads have cleaned themselves up and exited, a worker thread calls the process's LastPrayer() method that unmaps memory, deletes the address space, notifies the parent, etc. This provides a very robust way to terminate processes as even half-constructed processes (during a failing fork for instance) can be gracefully terminated. I have introduced a number of kernel threads to help avoid threading problems and simplify kernel design. For instance, there is now a functional generic kernel worker thread that any kernel thread can schedule jobs for. Interrupt handlers run with interrupts off (hence they cannot call kthread_ functions as it may deadlock the system if another thread holds the lock) therefore they cannot use the standard kernel worker threads. Instead, they use a special purpose interrupt worker thread that works much like the generic one expect that interrupt handlers can safely queue work with interrupts off. Note that this also means that interrupt handlers cannot allocate memory or print to the kernel log/screen as such mechanisms uses locks. I'll introduce a lock free algorithm for such cases later on. The boot process has also changed. The original kernel init thread in kernel.cpp creates a new bootstrap thread and becomes the system idle thread. Note that pid=0 now means the kernel, as there is no longer a system idle process. The bootstrap thread launches all the kernel worker threads and then creates a new process and loads /bin/init into it and then creates a thread in pid=1, which starts the system. The bootstrap thread then quietly waits for pid=1 to exit after which it shuts down/reboots/panics the system. In general, the introduction of race conditions and dead locks have forced me to revise a lot of the design and make sure it was thread secure. Since early parts of the kernel was quite hacky, I had to refactor such code. So it seems that the risk of dead locks forces me to write better code. Note that a real preemptive multithreaded kernel simplifies the construction of blocking system calls. My hope is that this will trigger a clean up of the filesystem code that current is almost beyond repair. Almost all of the kernel was modified during this refactoring. To the extent possible, these changes have been backported to older non-multithreaded kernel, but many changes were tightly coupled and went into this commit. Of interest is the implementation of the kthread_ api based on the design of pthreads; this library allows easy synchronization mechanisms and includes C++-style scoped locks. This commit also introduces new worker threads and tested mechanisms for interrupt handlers to schedule work in a kernel worker thread. A lot of code have been rewritten from scratch and has become a lot more stable and correct. Share and enjoy!
2012-08-01 15:30:34 +00:00
ScopedLock lock(&termlock);
TextBuffer* textbuf = textbufhandle->Acquire();
size_t width = textbuf->Width();
textbufhandle->Release(textbuf);
return width;
}
size_t TextTerminal::Height()
{
Multithreaded kernel and improvement of signal handling. Pardon the big ass-commit, this took months to develop and debug and the refactoring got so far that a clean merge became impossible. The good news is that this commit does quite a bit of cleaning up and generally improves the kernel quality. This makes the kernel fully pre-emptive and multithreaded. This was done by rewriting the interrupt code, the scheduler, introducing new threading primitives, and rewriting large parts of the kernel. During the past few commits the kernel has had its device drivers thread secured; this commit thread secures large parts of the core kernel. There still remains some parts of the kernel that is _not_ thread secured, but this is not a problem at this point. Each user-space thread has an associated kernel stack that it uses when it goes into kernel mode. This stack is by default 8 KiB since that value works for me and is also used by Linux. Strange things tends to happen on x86 in case of a stack overflow - there is no ideal way to catch such a situation right now. The system call conventions were changed, too. The %edx register is now used to provide the errno value of the call, instead of the kernel writing it into a registered global variable. The system call code has also been updated to better reflect the native calling conventions: not all registers have to be preserved. This makes system calls faster and simplifies the assembly. In the kernel, there is no longer the event.h header or the hacky method of 'resuming system calls' that closely resembles cooperative multitasking. If a system call wants to block, it should just block. The signal handling was also improved significantly. At this point, signals cannot interrupt kernel threads (but can always interrupt user-space threads if enabled), which introduces some problems with how a SIGINT could interrupt a blocking read, for instance. This commit introduces and uses a number of new primitives such as kthread_lock_mutex_signal() that attempts to get the lock but fails if a signal is pending. In this manner, the kernel is safer as kernel threads cannot be shut down inconveniently, but in return for complexity as blocking operations must check they if they should fail. Process exiting has also been refactored significantly. The _exit(2) system call sets the exit code and sends SIGKILL to all the threads in the process. Once all the threads have cleaned themselves up and exited, a worker thread calls the process's LastPrayer() method that unmaps memory, deletes the address space, notifies the parent, etc. This provides a very robust way to terminate processes as even half-constructed processes (during a failing fork for instance) can be gracefully terminated. I have introduced a number of kernel threads to help avoid threading problems and simplify kernel design. For instance, there is now a functional generic kernel worker thread that any kernel thread can schedule jobs for. Interrupt handlers run with interrupts off (hence they cannot call kthread_ functions as it may deadlock the system if another thread holds the lock) therefore they cannot use the standard kernel worker threads. Instead, they use a special purpose interrupt worker thread that works much like the generic one expect that interrupt handlers can safely queue work with interrupts off. Note that this also means that interrupt handlers cannot allocate memory or print to the kernel log/screen as such mechanisms uses locks. I'll introduce a lock free algorithm for such cases later on. The boot process has also changed. The original kernel init thread in kernel.cpp creates a new bootstrap thread and becomes the system idle thread. Note that pid=0 now means the kernel, as there is no longer a system idle process. The bootstrap thread launches all the kernel worker threads and then creates a new process and loads /bin/init into it and then creates a thread in pid=1, which starts the system. The bootstrap thread then quietly waits for pid=1 to exit after which it shuts down/reboots/panics the system. In general, the introduction of race conditions and dead locks have forced me to revise a lot of the design and make sure it was thread secure. Since early parts of the kernel was quite hacky, I had to refactor such code. So it seems that the risk of dead locks forces me to write better code. Note that a real preemptive multithreaded kernel simplifies the construction of blocking system calls. My hope is that this will trigger a clean up of the filesystem code that current is almost beyond repair. Almost all of the kernel was modified during this refactoring. To the extent possible, these changes have been backported to older non-multithreaded kernel, but many changes were tightly coupled and went into this commit. Of interest is the implementation of the kthread_ api based on the design of pthreads; this library allows easy synchronization mechanisms and includes C++-style scoped locks. This commit also introduces new worker threads and tested mechanisms for interrupt handlers to schedule work in a kernel worker thread. A lot of code have been rewritten from scratch and has become a lot more stable and correct. Share and enjoy!
2012-08-01 15:30:34 +00:00
ScopedLock lock(&termlock);
TextBuffer* textbuf = textbufhandle->Acquire();
size_t height = textbuf->Height();
textbufhandle->Release(textbuf);
return height;
}
void TextTerminal::GetCursor(size_t* column, size_t* row)
2013-12-20 20:55:05 +00:00
{
ScopedLock lock(&termlock);
*column = this->column;
*row = this->line;
}
bool TextTerminal::Sync()
{
// Reading something from the textbuffer may cause it to block while
// finishing rendering, effectively synchronizing with it.
ScopedLock lock(&termlock);
TextBuffer* textbuf = textbufhandle->Acquire();
textbuf->GetCursorPos();
textbufhandle->Release(textbuf);
return true;
}
bool TextTerminal::Invalidate()
{
ScopedLock lock(&termlock);
TextBuffer* textbuf = textbufhandle->Acquire();
textbuf->Invalidate();
textbufhandle->Release(textbuf);
return true;
}
2016-09-03 21:46:10 +00:00
void TextTerminal::BeginReplace()
{
kthread_mutex_lock(&termlock);
textbufhandle->BeginReplace();
}
void TextTerminal::CancelReplace()
{
textbufhandle->CancelReplace();
kthread_mutex_unlock(&termlock);
}
void TextTerminal::FinishReplace(TextBuffer* new_textbuf)
{
textbufhandle->FinishReplace(new_textbuf);
TextBuffer* textbuf = textbufhandle->Acquire();
size_t new_width = textbuf->Width();
size_t new_height = textbuf->Height();
textbufhandle->Release(textbuf);
if ( new_width < column )
column = new_width;
if ( new_height <= line )
line = new_height ? new_height - 1 : 0;
kthread_mutex_unlock(&termlock);
}
2013-11-11 21:21:56 +00:00
bool TextTerminal::EmergencyIsImpaired()
{
// This is during a kernel emergency where preemption has been disabled and
// this is the only thread running.
if ( !kthread_mutex_trylock(&termlock) )
return true;
kthread_mutex_unlock(&termlock);
if ( textbufhandle->EmergencyIsImpaired() )
return true;
TextBuffer* textbuf = textbufhandle->EmergencyAcquire();
bool textbuf_was_impaired = textbuf->EmergencyIsImpaired();
textbufhandle->EmergencyRelease(textbuf);
if ( textbuf_was_impaired )
return true;
return false;
}
bool TextTerminal::EmergencyRecoup()
{
// This is during a kernel emergency where preemption has been disabled and
// this is the only thread running.
if ( !kthread_mutex_trylock(&termlock) )
return false;
kthread_mutex_unlock(&termlock);
if ( textbufhandle->EmergencyIsImpaired() &&
!textbufhandle->EmergencyRecoup() )
return false;
TextBuffer* textbuf = textbufhandle->EmergencyAcquire();
bool textbuf_failure = textbuf->EmergencyIsImpaired() &&
!textbuf->EmergencyRecoup();
textbufhandle->EmergencyRelease(textbuf);
if ( !textbuf_failure )
return false;
return true;
}
void TextTerminal::EmergencyReset()
{
// This is during a kernel emergency where preemption has been disabled and
// this is the only thread running.
textbufhandle->EmergencyReset();
TextBuffer* textbuf = textbufhandle->EmergencyAcquire();
textbuf->EmergencyReset();
textbufhandle->EmergencyRelease(textbuf);
this->termlock = KTHREAD_MUTEX_INITIALIZER;
Reset();
}
size_t TextTerminal::EmergencyPrint(const char* string, size_t stringlen)
{
// This is during a kernel emergency where preemption has been disabled and
// this is the only thread running. Another thread may have been interrupted
// while it held the terminal lock. The best case is if the terminal lock is
// currently unused, which would mean everything is safe.
TextBuffer* textbuf = textbufhandle->EmergencyAcquire();
for ( size_t i = 0; i < stringlen; i++ )
{
if ( string[i] == '\n' )
PutChar(textbuf, '\r');
PutChar(textbuf, string[i]);
}
UpdateCursor(textbuf);
textbufhandle->EmergencyRelease(textbuf);
return stringlen;
}
size_t TextTerminal::EmergencyPrintRaw(const char* string, size_t stringlen)
{
// This is during a kernel emergency where preemption has been disabled and
// this is the only thread running. Another thread may have been interrupted
// while it held the terminal lock. The best case is if the terminal lock is
// currently unused, which would mean everything is safe.
2013-11-11 21:21:56 +00:00
TextBuffer* textbuf = textbufhandle->EmergencyAcquire();
for ( size_t i = 0; i < stringlen; i++ )
PutChar(textbuf, string[i]);
UpdateCursor(textbuf);
textbufhandle->EmergencyRelease(textbuf);
return stringlen;
}
size_t TextTerminal::EmergencyWidth()
2013-11-11 21:21:56 +00:00
{
// This is during a kernel emergency where preemption has been disabled and
// this is the only thread running. Another thread may have been interrupted
// while it held the terminal lock. The best case is if the terminal lock is
// currently unused, which would mean everything is safe.
TextBuffer* textbuf = textbufhandle->EmergencyAcquire();
size_t width = textbuf->Width();
textbufhandle->EmergencyRelease(textbuf);
return width;
}
size_t TextTerminal::EmergencyHeight()
2013-11-11 21:21:56 +00:00
{
// This is during a kernel emergency where preemption has been disabled and
// this is the only thread running. Another thread may have been interrupted
// while it held the terminal lock. The best case is if the terminal lock is
// currently unused, which would mean everything is safe.
TextBuffer* textbuf = textbufhandle->EmergencyAcquire();
size_t height = textbuf->Height();
textbufhandle->EmergencyRelease(textbuf);
return height;
}
void TextTerminal::EmergencyGetCursor(size_t* column, size_t* row)
2013-12-20 20:55:05 +00:00
{
// This is during a kernel emergency where preemption has been disabled and
// this is the only thread running. Another thread may have been interrupted
// while it held the terminal lock. The best case is if the terminal lock is
// currently unused, which would mean everything is safe.
*column = this->column;
*row = this->line;
}
2013-11-11 21:21:56 +00:00
bool TextTerminal::EmergencySync()
{
// This is during a kernel emergency where preemption has been disabled and
// this is the only thread running. There is no need to synchronize the
// text buffer here as there is no background thread rendering the console.
return true;
}
void TextTerminal::PutChar(TextBuffer* textbuf, char c)
{
if ( ansimode )
return PutAnsiEscaped(textbuf, c);
if ( mbsinit(&ps) )
{
switch ( c )
{
case '\a': return;
case '\n': Newline(textbuf); return;
case '\r': column = 0; return;
case '\b': Backspace(textbuf); return;
case '\t': Tab(textbuf); return;
case '\e': AnsiReset(); return;
2015-12-27 20:37:10 +00:00
case 127: return;
default: break;
}
}
wchar_t wc;
size_t result = mbrtowc(&wc, &c, 1, &ps);
if ( result == (size_t) -2 )
return;
if ( result == (size_t) -1 )
{
memset(&ps, 0, sizeof(ps));
wc = L'<EFBFBD>';
}
if ( result == (size_t) 0 )
wc = L' ';
if ( textbuf->Width() <= column )
{
column = 0;
Newline(textbuf);
}
TextPos pos(column++, line);
uint16_t tcvgacolor;
uint16_t tcattr = attr | next_attr;
uint32_t tcfgcolor;
uint32_t tcbgcolor;
if ( !(tcattr & ATTR_INVERSE) )
{
tcvgacolor = vgacolor;
tcfgcolor = fgcolor;
tcbgcolor = bgcolor;
}
else
{
tcvgacolor = (vgacolor >> 4 & 0xF) << 0 | (vgacolor >> 0 & 0xF) << 4;
tcfgcolor = bgcolor;
tcbgcolor = fgcolor;
}
TextChar tc(wc, tcvgacolor, tcattr, tcfgcolor, tcbgcolor);
textbuf->SetChar(pos, tc);
next_attr = 0;
}
void TextTerminal::UpdateCursor(TextBuffer* textbuf)
{
textbuf->SetCursorPos(TextPos(column, line));
}
void TextTerminal::Newline(TextBuffer* textbuf)
{
TextPos pos(column, line);
if ( line < textbuf->Height()-1 )
line++;
else
{
uint32_t fillfg = attr & ATTR_INVERSE ? bgcolor : fgcolor;
uint32_t fillbg = attr & ATTR_INVERSE ? fgcolor : bgcolor;
textbuf->Scroll(1, TextChar(' ', vgacolor, 0, fillfg, fillbg));
line = textbuf->Height()-1;
}
}
#if 0
static TextPos DecrementTextPos(TextBuffer* textbuf, TextPos pos)
{
if ( !pos.x && !pos.y )
return pos;
if ( !pos.x )
return TextPos(textbuf->Width(), pos.y-1);
return TextPos(pos.x-1, pos.y);
}
#endif
void TextTerminal::Backspace(TextBuffer* textbuf)
{
if ( column )
{
column--;
TextPos pos(column, line);
TextChar tc = textbuf->GetChar(pos);
next_attr = tc.attr & (ATTR_BOLD | ATTR_UNDERLINE);
if ( tc.c == L'_' )
next_attr |= ATTR_UNDERLINE;
else if ( tc.c == L' ' )
next_attr &= ~ATTR_BOLD;
else
next_attr |= ATTR_BOLD;
}
}
void TextTerminal::Tab(TextBuffer* textbuf)
{
if ( column == textbuf->Width() )
{
column = 0;
Newline(textbuf);
}
column++;
column = -(-column & ~0x7U);
size_t width = textbuf->Width();
if ( width <= column )
column = width;
}
// TODO: This implementation of the 'Ansi Escape Codes' is incomplete and hacky.
void TextTerminal::AnsiReset()
{
next_attr = 0;
ansiusedparams = 0;
ansiparams[0] = 0;
ignoresequence = false;
ansimode = CSI;
}
void TextTerminal::PutAnsiEscaped(TextBuffer* textbuf, char c)
{
// Check the proper prefixes are used.
if ( ansimode == CSI )
{
if ( c == '[' )
ansimode = COMMAND;
else if ( c == '(' || c == ')' || c == '*' || c == '+' ||
c == '-' || c == '.' || c == '/' )
ansimode = CHARSET;
// TODO: Enter and exit alternatve keypad mode.
else if ( c == '=' || c == '>' )
ansimode = NONE;
else
{
ansimode = NONE;
ansimode = NONE;
}
return;
}
if ( ansimode == CHARSET )
{
ansimode = NONE;
return;
}
// Read part of a parameter.
if ( '0' <= c && c <= '9' )
{
if ( ansiusedparams == 0 )
ansiusedparams++;
unsigned val = c - '0';
ansiparams[ansiusedparams-1] *= 10;
ansiparams[ansiusedparams-1] += val;
}
// Parameter delimiter.
else if ( c == ';' )
{
if ( ansiusedparams == ANSI_NUM_PARAMS )
{
ansimode = NONE;
return;
}
ansiparams[ansiusedparams++] = 0;
}
// Left for future standardization, so discard this sequence.
else if ( c == ':' )
{
ignoresequence = true;
}
else if ( c == '>' )
{
ansimode = GREATERTHAN;
}
// Run a command.
else if ( 64 <= c && c <= 126 )
{
if ( !ignoresequence )
{
if ( ansimode == COMMAND )
RunAnsiCommand(textbuf, c);
else if ( ansimode == GREATERTHAN )
{
// Send Device Attributes
if ( c == 'c' )
{
// TODO: Send an appropriate response through the terminal.
}
else
{
ansimode = NONE;
return;
}
ansimode = NONE;
}
}
else
ansimode = NONE;
}
// Something I don't understand, and ignore intentionally.
else if ( c == '?' )
{
}
// TODO: There are some rare things that should be supported here.
// Ignore unknown input.
else
{
ansimode = NONE;
}
}
void TextTerminal::RunAnsiCommand(TextBuffer* textbuf, char c)
{
const unsigned width = (unsigned) textbuf->Width();
const unsigned height = (unsigned) textbuf->Height();
switch ( c )
{
case 'A': // Cursor up
{
unsigned dist = 0 < ansiusedparams ? ansiparams[0] : 1;
if ( line < dist )
line = 0;
else
line -= dist;
} break;
case 'B': // Cursor down
{
unsigned dist = 0 < ansiusedparams ? ansiparams[0] : 1;
if ( height <= line + dist )
line = height-1;
else
line += dist;
} break;
case 'C': // Cursor forward
{
unsigned dist = 0 < ansiusedparams ? ansiparams[0] : 1;
if ( width <= column + dist )
column = width-1;
else
column += dist;
} break;
case 'D': // Cursor backward
{
unsigned dist = 0 < ansiusedparams ? ansiparams[0] : 1;
if ( column < dist )
column = 0;
else
column -= dist;
} break;
case 'E': // Move to beginning of line N lines down.
{
column = 0;
unsigned dist = 0 < ansiusedparams ? ansiparams[0] : 1;
if ( height <= line + dist )
line = height-1;
else
line += dist;
} break;
case 'F': // Move to beginning of line N lines up.
{
column = 0;
unsigned dist = 0 < ansiusedparams ? ansiparams[0] : 1;
if ( line < dist )
line = 0;
else
line -= dist;
} break;
case 'G': // Move the cursor to column N.
{
unsigned pos = 0 < ansiusedparams ? ansiparams[0]-1 : 0;
if ( width <= pos )
pos = width-1;
column = pos;
} break;
case 'H': // Move the cursor to line Y, column X.
case 'f':
{
unsigned posy = 0 < ansiusedparams ? ansiparams[0]-1 : 0;
unsigned posx = 1 < ansiusedparams ? ansiparams[1]-1 : 0;
if ( width <= posx )
posx = width-1;
if ( height <= posy )
posy = height-1;
column = posx;
line = posy;
} break;
case 'J': // Erase parts of the screen.
{
unsigned mode = 0 < ansiusedparams ? ansiparams[0] : 0;
TextPos from(0, 0);
TextPos to(0, 0);
if ( mode == 0 ) // From cursor to end.
from = TextPos{column, line},
to = TextPos{width-1, height-1};
if ( mode == 1 ) // From start to cursor.
from = TextPos{0, 0},
to = TextPos{column, line};
if ( mode == 2 ) // Everything.
from = TextPos{0, 0},
to = TextPos{width-1, height-1};
uint32_t fillfg = attr & ATTR_INVERSE ? bgcolor : fgcolor;
uint32_t fillbg = attr & ATTR_INVERSE ? fgcolor : bgcolor;
textbuf->Fill(from, to, TextChar(' ', vgacolor, 0, fillfg, fillbg));
} break;
case 'K': // Erase parts of the current line.
{
unsigned mode = 0 < ansiusedparams ? ansiparams[0] : 0;
TextPos from(0, 0);
TextPos to(0, 0);
if ( mode == 0 ) // From cursor to end.
from = TextPos{column, line},
to = TextPos{width-1, line};
if ( mode == 1 ) // From start to cursor.
from = TextPos{0, line},
to = TextPos{column, line};
if ( mode == 2 ) // Everything.
from = TextPos{0, line},
to = TextPos{width-1, line};
uint32_t fillfg = attr & ATTR_INVERSE ? bgcolor : fgcolor;
uint32_t fillbg = attr & ATTR_INVERSE ? fgcolor : bgcolor;
textbuf->Fill(from, to, TextChar(' ', vgacolor, 0, fillfg, fillbg));
} break;
case 'L': // Append lines before current line.
{
column = 0;
unsigned count = 0 < ansiusedparams ? ansiparams[0] : 1;
if ( height - line < count )
count = height - line;
TextPos from(0, line);
TextPos move_to(0, line + count);
unsigned move_lines = height - (line + count);
textbuf->Move(move_to, from, move_lines * width);
if ( 0 < count )
{
TextPos fill_to(width - 1, line + count - 1);
uint32_t fill_fg = attr & ATTR_INVERSE ? bgcolor : fgcolor;
uint32_t fill_bg = attr & ATTR_INVERSE ? fgcolor : bgcolor;
TextChar fill_char(' ', vgacolor, 0, fill_fg, fill_bg);
textbuf->Fill(from, fill_to, fill_char);
}
} break;
case 'M': // Delete lines starting from beginning of current line.
{
column = 0;
unsigned count = 0 < ansiusedparams ? ansiparams[0] : 1;
if ( height - line < count )
count = height - line;
TextPos move_from(0, line + count);
TextPos move_to(0, line);
unsigned move_lines = height - (line + count);
textbuf->Move(move_to, move_from, move_lines * width);
if ( 0 < count )
{
TextPos fill_from(0, height - count);
TextPos fill_to(width - 1, height - 1);
uint32_t fill_fg = attr & ATTR_INVERSE ? bgcolor : fgcolor;
uint32_t fill_bg = attr & ATTR_INVERSE ? fgcolor : bgcolor;
TextChar fill_char(' ', vgacolor, 0, fill_fg, fill_bg);
textbuf->Fill(fill_from, fill_to, fill_char);
}
} break;
// TODO: CSI Ps P Delete Ps Character(s) (default = 1) (DCH).
// (delete those characters and move the rest of the line leftward).
case 'S': // Scroll a line up and place a new line at the buttom.
{
uint32_t fillfg = attr & ATTR_INVERSE ? bgcolor : fgcolor;
uint32_t fillbg = attr & ATTR_INVERSE ? fgcolor : bgcolor;
textbuf->Scroll(1, TextChar(' ', vgacolor, 0, fillfg, fillbg));
line = height-1;
} break;
case 'T': // Scroll a line up and place a new line at the top.
{
uint32_t fillfg = attr & ATTR_INVERSE ? bgcolor : fgcolor;
uint32_t fillbg = attr & ATTR_INVERSE ? fgcolor : bgcolor;
textbuf->Scroll(-1, TextChar(' ', vgacolor, 0, fillfg, fillbg));
line = 0;
} break;
case 'd': // Move the cursor to line N.
{
unsigned posy = 0 < ansiusedparams ? ansiparams[0]-1 : 0;
if ( height <= posy )
posy = height-1;
line = posy;
} break;
case 'm': // Change how the text is rendered.
{
if ( ansiusedparams == 0 )
{
ansiparams[0] = 0;
ansiusedparams++;
}
// Convert from the ANSI color scheme to the VGA color scheme.
const unsigned conversion[8] =
{
#if 0
0, 4, 2, 6, 1, 5, 3, 7
#else
COLOR8_BLACK, COLOR8_RED, COLOR8_GREEN, COLOR8_BROWN,
COLOR8_BLUE, COLOR8_MAGENTA, COLOR8_CYAN, COLOR8_LIGHT_GREY,
#endif
};
for ( size_t i = 0; i < ansiusedparams; i++ )
{
unsigned cmd = ansiparams[i];
// Turn all attributes off.
if ( cmd == 0 )
{
vgacolor = DEFAULT_VGACOLOR;
attr = 0;
fgcolor = palette[DEFAULT_FOREGROUND];
bgcolor = palette[DEFAULT_BACKGROUND];
}
// Boldness.
else if ( cmd == 1 )
attr |= ATTR_BOLD;
// TODO: 2, Faint
// TODO: 3, Italicized
// Underline.
else if ( cmd == 4 )
attr |= ATTR_UNDERLINE;
// TODO: 5, Blink (appears as Bold)
// Inverse.
else if ( cmd == 7 )
attr |= ATTR_INVERSE;
// TODO: 8, Invisible
// TODO: 9, Crossed-out
// TODO: 21, Doubly-underlined
// Normal (neither bold nor faint).
else if ( cmd == 22 )
attr &= ~ATTR_BOLD;
// TODO: 23, Not italicized
// Not underlined.
else if ( cmd == 24 )
attr &= ~ATTR_UNDERLINE;
// TODO: 25, Steady (not blinking)
// Positive (not inverse).
else if ( cmd == 27 )
attr &= ~ATTR_INVERSE;
// TODO: 28, Visible (not hidden)
// Set text color.
else if ( 30 <= cmd && cmd <= 37 )
{
unsigned val = cmd - 30;
vgacolor &= 0xF0;
vgacolor |= conversion[val] << 0;
fgcolor = palette[val];
}
// Set text color.
else if ( cmd == 38 )
{
if ( 5 <= ansiusedparams - i && ansiparams[i+1] == 2 )
{
uint8_t r = ansiparams[i+2];
uint8_t g = ansiparams[i+3];
uint8_t b = ansiparams[i+4];
i += 5 - 1;
fgcolor = ColorFromRGB(r, g, b);
// TODO: Approxpiate vgacolor.
}
else if ( 3 <= ansiusedparams - i && ansiparams[i+1] == 5 )
{
uint8_t index = ansiparams[i+2];
i += 3 - 1;
fgcolor = palette[index];
// TODO: Approxpiate vgacolor.
}
}
// Set default text color.
else if ( cmd == 39 )
{
vgacolor &= 0xF0;
vgacolor |= DEFAULT_VGACOLOR & 0x0F;
fgcolor = palette[DEFAULT_FOREGROUND];
}
// Set background color.
else if ( 40 <= cmd && cmd <= 47 )
{
unsigned val = cmd - 40;
vgacolor &= 0x0F;
vgacolor |= conversion[val] << 4;
bgcolor = palette[val];
}
// Set background color.
else if ( cmd == 48 )
{
if ( 5 <= ansiusedparams - i && ansiparams[i+1] == 2 )
{
uint8_t r = ansiparams[i+2];
uint8_t g = ansiparams[i+3];
uint8_t b = ansiparams[i+4];
i += 5 - 1;
bgcolor = ColorFromRGB(r, g, b);
// TODO: Approxpiate vgacolor.
}
else if ( 3 <= ansiusedparams - i && ansiparams[i+1] == 5 )
{
uint8_t index = ansiparams[i+2];
i += 3 - 1;
bgcolor = palette[index];
// TODO: Approxpiate vgacolor.
}
}
// Set default background color.
else if ( cmd == 49 )
{
vgacolor &= 0x0F;
vgacolor |= DEFAULT_VGACOLOR & 0xF0;
bgcolor = palette[DEFAULT_BACKGROUND];
}
// Set text color.
else if ( 90 <= cmd && cmd <= 97 )
{
unsigned val = cmd - 90 + 8;
vgacolor &= 0xF0;
vgacolor |= (0x8 | conversion[val - 8]) << 0;
fgcolor = palette[val];
}
// Set background color.
else if ( 100 <= cmd && cmd <= 107 )
{
unsigned val = cmd - 100 + 8;
vgacolor &= 0x0F;
vgacolor |= (0x8 | conversion[val - 8]) << 4;
bgcolor = palette[val];
}
else
{
ansimode = NONE;
}
}
} break;
case 'n': // Request special information from terminal.
{
ansimode = NONE;
// TODO: Handle this code.
} break;
case 's': // Save cursor position.
{
ansisavedposx = column;
ansisavedposy = line;
} break;
case 'u': // Restore cursor position.
{
column = ansisavedposx;
line = ansisavedposy;
if ( width <= column )
column = width-1;
if ( height <= line )
line = height-1;
} break;
case 'l': // Hide cursor.
{
// TODO: This is somehow related to the special char '?'.
if ( 0 < ansiusedparams && ansiparams[0] == 25 )
textbuf->SetCursorEnabled(false);
if ( 0 < ansiusedparams && ansiparams[0] == 1049 )
{}; // TODO: Save scrollback.
} break;
case 'h': // Show cursor.
{
// TODO: This is somehow related to the special char '?'.
if ( 0 < ansiusedparams && ansiparams[0] == 25 )
textbuf->SetCursorEnabled(true);
if ( 0 < ansiusedparams && ansiparams[0] == 1049 )
{}; // TODO: Restore scrollback.
} break;
default:
{
ansimode = NONE;
}
// TODO: Handle other cases.
}
ansimode = NONE;
}
} // namespace Sortix