Add new code

Signed-off-by: Carlos Maiolino <[email protected]>

3 files changed, 432 insertions, 0 deletions

diff --git a/PGU/CHAP5/argtest.s b/PGU/CHAP5/argtest.s
new file mode 100644
index 0000000..aded606
--- /dev/null
+++ b/PGU/CHAP5/argtest.s
@@ -0,0 +1,20 @@
+.section .data
+
+.section .text
+
+.globl _start
+
+_start:
+	movq %rsp, %rbp
+	movq (%rbp), %r12
+	movq 16(%rbp), %r13
+
+	movq $2, %rax	#SYS_OPEN
+	movq %r13, %rdi
+	movq $0, %rsi
+	syscall
+
+	movq %r12, %rdi
+	movq $60, %rax
+	syscall
+
diff --git a/PGU/CHAP5/toupper.s b/PGU/CHAP5/toupper.s
new file mode 100644
index 0000000..b3f4f13
--- /dev/null
+++ b/PGU/CHAP5/toupper.s
@@ -0,0 +1,206 @@
+# Convert an input file to an output file with all
+# letters converted to uppercase
+
+.section .data
+
+# CONSTANTS
+.equ SYS_OPEN, 2
+.equ SYS_WRITE, 1
+.equ SYS_READ, 0
+.equ SYS_CLOSE, 3
+.equ SYS_EXIT, 60
+
+# We have no access to headers, so, let's define our flags
+.equ O_RDONLY, 0
+.equ O_CREAT_WRONLY_TRUNC, 03101
+
+# Add some other naming to make my life easier
+.equ STDIN, 0
+.equ STDOUT, 1
+.equ STDERR, 2
+
+.equ END_OF_FILE, 0
+
+.equ NUMBER_ARGUMENTS, 2
+
+################################################################
+
+.section .bss
+
+.equ BUFFER_SIZE, 500
+.lcomm BUFFER_DATA, BUFFER_SIZE
+
+################################################################
+
+.section .text
+
+# STACK POSITIONS
+
+.equ ST_SIZE_RESERVE, 16
+.equ ST_FD_IN, -8
+.equ ST_FD_OUT, -16
+.equ ST_ARGC, 0	# Number of arguments
+.equ ST_ARGV_0, 8	# Program's name
+.equ ST_ARGV_1, 16	# Input file name
+.equ ST_ARGV_2, 24	# Output file name
+
+.globl _start
+.global convert_to_upper # Declaration only
+
+### START OF PROGRAM ###
+_start:
+
+	# Save stack pointer
+	movq %rsp, %rbp
+
+	# Alloc space for FDs on the stack
+	subq $ST_SIZE_RESERVE, %rsp
+
+# Labels are here just to make my life easier in splitting the sections
+# I won't be jumping around...
+
+# int open(const char *pathname, int flags, mode_t mode)
+# Argument's order:   %rdi, %rsi, %rdx
+
+open_files:
+open_fd_in:
+	movq $SYS_OPEN, %rax		# Setup syscall number
+	movq ST_ARGV_1(%rbp), %rdi	# input filename into %rdi
+	movq $O_RDONLY, %rsi		# Open in O_RDONLY
+	movq $0666, %rdx		# regular posix perms
+	syscall				# Invoke kernel to call sys_open()
+
+store_fd_in:
+	movq %rax, ST_FD_IN(%rbp)	# Save file's FD
+
+open_fd_out:
+	movq $SYS_OPEN, %rax		# Setup syscall number
+	movq ST_ARGV_2(%rbp), %rdi	# Output file name location to %ebx
+	movq $O_CREAT_WRONLY_TRUNC, %rsi# Setup writing flags
+	movq $0666, %rdx		# POSIX mode
+	syscall				# Invoke kernel (sys_open())
+
+store_fd_out:
+	movq %rax, ST_FD_OUT(%rbp)	# Save output FD in the stack
+
+
+################################################################
+
+# ssize_t read(int fd, void *buf, size_t count);
+# ssize_t write(int fd, const void *buf, size_t count);
+
+read_loop_begin:
+	movq $SYS_READ, %rax
+	movq ST_FD_IN(%rbp), %rdi	# Read from this FD
+	movq $BUFFER_DATA, %rsi		# Copy buffer's address to %rsi, so
+					# read() knows where to put data read
+
+	movq $BUFFER_SIZE, %rdx		# Set buffer size for read()
+	syscall				# Invoke kernel
+
+	# Number of bytes read should be in %eax
+
+	# Leave loop if we are at EOF
+	cmpq $END_OF_FILE, %rax
+	jle end_loop
+
+continue_read_loop:
+	# Call convert function for the current buffer
+	pushq $BUFFER_DATA	# Buffer location - 2nd argument
+	pushq %rax		# Buffer size - 1st argument
+				# (returned from read() syscall)
+
+	call convert_to_upper
+	popq %rax		# get size back
+	addq $8, %rsp		# Cleanup stack used for buffer_data location
+
+write_file:
+	movq %rax, %rdx			# Use number of bytes read to write to
+					# output file
+	movq $SYS_WRITE, %rax
+	movq ST_FD_OUT(%rbp), %rdi	# Write to this FD
+	movq $BUFFER_DATA, %rsi		# This buffer
+	syscall				# Invoke kernel
+
+	jmp read_loop_begin # Keep reading the file
+
+end_loop:
+	movq $SYS_CLOSE, %rax
+	movq ST_FD_OUT(%rbp), %rdi
+	syscall
+
+	movq $SYS_CLOSE, %rax
+	movq ST_FD_IN(%rbp), %rdi
+	syscall
+
+	movq $SYS_EXIT, %rax
+	movq $0, %rdi
+	syscall
+
+
+# convert_to_upper:
+#
+# Convert all ascii characters in $BUFFER
+# to uppercase
+#
+# - 1st argument: Buffer size
+# - 2nd argument: Buffer address
+#
+# REMEMBER: Arguments must be pushed in the reverse order as they are documented
+#
+#	- %eax - beginning of buffer
+#	- %rdi - buffer length
+#	- %rsi - current buffer offset
+#	- %cl  - byte being examined (lower byte of %rcx)
+
+## CONSTANTS
+.equ LOWERCASE_A, 'a' # lower bundary
+.equ LOWERCASE_Z, 'z' # upper boundary
+
+# CONVERSION
+.equ UPPER_CONVERSION, 'A' - 'a'
+
+## STACK STUFF
+.equ ST_BUFFER_LEN, 16	# Buffer length
+.equ ST_BUFFER, 24	# Actual Buffer
+
+
+.type convert_to_upper, @function
+
+convert_to_upper:
+	pushq %rbp	# Setup function's stack frame. Save current %ebp
+	movq %rsp, %rbp # And set it to current stack pointer
+
+	movq ST_BUFFER(%rbp), %rax
+	movq ST_BUFFER_LEN(%rbp), %rdi
+	movq $0, %rsi
+
+	# If a zero length buffer was given, just leave
+	cmpq $0, %rax
+	je end_convert_loop
+
+convert_loop:
+	movb (%rax, %rsi, 1), %cl	# Get current byte
+
+	# Check if current byte is already in low case
+	cmpb $LOWERCASE_A, %cl
+	jl next_byte
+	cmpb $LOWERCASE_Z, %cl
+	jg next_byte
+
+	# Convert if we are still here
+	addb $UPPER_CONVERSION, %cl
+	movb %cl, (%rax, %rsi, 1) # Replace old value with the new one
+
+next_byte:
+	incq %rsi
+	cmpq %rsi, %rdi		# Continue conversion unless we reached
+	jne convert_loop	# the end of the buffer
+
+end_convert_loop:
+	# no return value, just restore stack pointer
+	# and base pointer and return.
+	movq %rbp, %rsp
+	popq %rbp
+	ret
+
diff --git a/PGU/CHAP5/toupper.up b/PGU/CHAP5/toupper.up
new file mode 100644
index 0000000..d4c2367
--- /dev/null
+++ b/PGU/CHAP5/toupper.up
@@ -0,0 +1,206 @@
+# CONVERT AN INPUT FILE TO AN OUTPUT FILE WITH ALL
+# LETTERS CONVERTED TO UPPERCASE
+
+.SECTION .DATA
+
+# CONSTANTS
+.EQU SYS_OPEN, 2
+.EQU SYS_WRITE, 1
+.EQU SYS_READ, 0
+.EQU SYS_CLOSE, 3
+.EQU SYS_EXIT, 60
+
+# WE HAVE NO ACCESS TO HEADERS, SO, LET'S DEFINE OUR FLAGS
+.EQU O_RDONLY, 0
+.EQU O_CREAT_WRONLY_TRUNC, 03101
+
+# ADD SOME OTHER NAMING TO MAKE MY LIFE EASIER
+.EQU STDIN, 0
+.EQU STDOUT, 1
+.EQU STDERR, 2
+
+.EQU END_OF_FILE, 0
+
+.EQU NUMBER_ARGUMENTS, 2
+
+################################################################
+
+.SECTION .BSS
+
+.EQU BUFFER_SIZE, 500
+.LCOMM BUFFER_DATA, BUFFER_SIZE
+
+################################################################
+
+.SECTION .TEXT
+
+# STACK POSITIONS
+
+.EQU ST_SIZE_RESERVE, 16
+.EQU ST_FD_IN, -8
+.EQU ST_FD_OUT, -16
+.EQU ST_ARGC, 0	# NUMBER OF ARGUMENTS
+.EQU ST_ARGV_0, 8	# PROGRAM'S NAME
+.EQU ST_ARGV_1, 16	# INPUT FILE NAME
+.EQU ST_ARGV_2, 24	# OUTPUT FILE NAME
+
+.GLOBL _START
+.GLOBAL CONVERT_TO_UPPER # DECLARATION ONLY
+
+### START OF PROGRAM ###
+_START:
+
+	# SAVE STACK POINTER
+	MOVQ %RSP, %RBP
+
+	# ALLOC SPACE FOR FDS ON THE STACK
+	SUBQ $ST_SIZE_RESERVE, %RSP
+
+# LABELS ARE HERE JUST TO MAKE MY LIFE EASIER IN SPLITTING THE SECTIONS
+# I WON'T BE JUMPING AROUND...
+
+# INT OPEN(CONST CHAR *PATHNAME, INT FLAGS, MODE_T MODE)
+# ARGUMENT'S ORDER:   %RDI, %RSI, %RDX
+
+OPEN_FILES:
+OPEN_FD_IN:
+	MOVQ $SYS_OPEN, %RAX		# SETUP SYSCALL NUMBER
+	MOVQ ST_ARGV_1(%RBP), %RDI	# INPUT FILENAME INTO %RDI
+	MOVQ $O_RDONLY, %RSI		# OPEN IN O_RDONLY
+	MOVQ $0666, %RDX		# REGULAR POSIX PERMS
+	SYSCALL				# INVOKE KERNEL TO CALL SYS_OPEN()
+
+STORE_FD_IN:
+	MOVQ %RAX, ST_FD_IN(%RBP)	# SAVE FILE'S FD
+
+OPEN_FD_OUT:
+	MOVQ $SYS_OPEN, %RAX		# SETUP SYSCALL NUMBER
+	MOVQ ST_ARGV_2(%RBP), %RDI	# OUTPUT FILE NAME LOCATION TO %EBX
+	MOVQ $O_CREAT_WRONLY_TRUNC, %RSI# SETUP WRITING FLAGS
+	MOVQ $0666, %RDX		# POSIX MODE
+	SYSCALL				# INVOKE KERNEL (SYS_OPEN())
+
+STORE_FD_OUT:
+	MOVQ %RAX, ST_FD_OUT(%RBP)	# SAVE OUTPUT FD IN THE STACK
+
+
+################################################################
+
+# SSIZE_T READ(INT FD, VOID *BUF, SIZE_T COUNT);
+# SSIZE_T WRITE(INT FD, CONST VOID *BUF, SIZE_T COUNT);
+
+READ_LOOP_BEGIN:
+	MOVQ $SYS_READ, %RAX
+	MOVQ ST_FD_IN(%RBP), %RDI	# READ FROM THIS FD
+	MOVQ $BUFFER_DATA, %RSI		# COPY BUFFER'S ADDRESS TO %RSI, SO
+					# READ() KNOWS WHERE TO PUT DATA READ
+
+	MOVQ $BUFFER_SIZE, %RDX		# SET BUFFER SIZE FOR READ()
+	SYSCALL				# INVOKE KERNEL
+
+	# NUMBER OF BYTES READ SHOULD BE IN %EAX
+
+	# LEAVE LOOP IF WE ARE AT EOF
+	CMPQ $END_OF_FILE, %RAX
+	JLE END_LOOP
+
+CONTINUE_READ_LOOP:
+	# CALL CONVERT FUNCTION FOR THE CURRENT BUFFER
+	PUSHQ $BUFFER_DATA	# BUFFER LOCATION - 2ND ARGUMENT
+	PUSHQ %RAX		# BUFFER SIZE - 1ST ARGUMENT
+				# (RETURNED FROM READ() SYSCALL)
+
+	CALL CONVERT_TO_UPPER
+	POPQ %RAX		# GET SIZE BACK
+	ADDQ $8, %RSP		# CLEANUP STACK USED FOR BUFFER_DATA LOCATION
+
+WRITE_FILE:
+	MOVQ %RAX, %RDX			# USE NUMBER OF BYTES READ TO WRITE TO
+					# OUTPUT FILE
+	MOVQ $SYS_WRITE, %RAX
+	MOVQ ST_FD_OUT(%RBP), %RDI	# WRITE TO THIS FD
+	MOVQ $BUFFER_DATA, %RSI		# THIS BUFFER
+	SYSCALL				# INVOKE KERNEL
+
+	JMP READ_LOOP_BEGIN # KEEP READING THE FILE
+
+END_LOOP:
+	MOVQ $SYS_CLOSE, %RAX
+	MOVQ ST_FD_OUT(%RBP), %RDI
+	SYSCALL
+
+	MOVQ $SYS_CLOSE, %RAX
+	MOVQ ST_FD_IN(%RBP), %RDI
+	SYSCALL
+
+	MOVQ $SYS_EXIT, %RAX
+	MOVQ $0, %RDI
+	SYSCALL
+
+
+# CONVERT_TO_UPPER:
+#
+# CONVERT ALL ASCII CHARACTERS IN $BUFFER
+# TO UPPERCASE
+#
+# - 1ST ARGUMENT: BUFFER SIZE
+# - 2ND ARGUMENT: BUFFER ADDRESS
+#
+# REMEMBER: ARGUMENTS MUST BE PUSHED IN THE REVERSE ORDER AS THEY ARE DOCUMENTED
+#
+#	- %EAX - BEGINNING OF BUFFER
+#	- %RDI - BUFFER LENGTH
+#	- %RSI - CURRENT BUFFER OFFSET
+#	- %CL  - BYTE BEING EXAMINED (LOWER BYTE OF %RCX)
+
+## CONSTANTS
+.EQU LOWERCASE_A, 'A' # LOWER BUNDARY
+.EQU LOWERCASE_Z, 'Z' # UPPER BOUNDARY
+
+# CONVERSION
+.EQU UPPER_CONVERSION, 'A' - 'A'
+
+## STACK STUFF
+.EQU ST_BUFFER_LEN, 16	# BUFFER LENGTH
+.EQU ST_BUFFER, 24	# ACTUAL BUFFER
+
+
+.TYPE CONVERT_TO_UPPER, @FUNCTION
+
+CONVERT_TO_UPPER:
+	PUSHQ %RBP	# SETUP FUNCTION'S STACK FRAME. SAVE CURRENT %EBP
+	MOVQ %RSP, %RBP # AND SET IT TO CURRENT STACK POINTER
+
+	MOVQ ST_BUFFER(%RBP), %RAX
+	MOVQ ST_BUFFER_LEN(%RBP), %RDI
+	MOVQ $0, %RSI
+
+	# IF A ZERO LENGTH BUFFER WAS GIVEN, JUST LEAVE
+	CMPQ $0, %RAX
+	JE END_CONVERT_LOOP
+
+CONVERT_LOOP:
+	MOVB (%RAX, %RSI, 1), %CL	# GET CURRENT BYTE
+
+	# CHECK IF CURRENT BYTE IS ALREADY IN LOW CASE
+	CMPB $LOWERCASE_A, %CL
+	JL NEXT_BYTE
+	CMPB $LOWERCASE_Z, %CL
+	JG NEXT_BYTE
+
+	# CONVERT IF WE ARE STILL HERE
+	ADDB $UPPER_CONVERSION, %CL
+	MOVB %CL, (%RAX, %RSI, 1) # REPLACE OLD VALUE WITH THE NEW ONE
+
+NEXT_BYTE:
+	INCQ %RSI
+	CMPQ %RSI, %RDI		# CONTINUE CONVERSION UNLESS WE REACHED
+	JNE CONVERT_LOOP	# THE END OF THE BUFFER
+
+END_CONVERT_LOOP:
+	# NO RETURN VALUE, JUST RESTORE STACK POINTER
+	# AND BASE POINTER AND RETURN.
+	MOVQ %RBP, %RSP
+	POPQ %RBP
+	RET
+