6502/cc65
1
0
Fork 0
mirror of https://github.com/cc65/cc65.git synced 2024-12-23 19:29:37 +00:00
Code Issues Projects Releases Wiki Activity
21030c22a0
cc65/libsrc/c64/gettime.s

84 lines
1.9 KiB
ArmAsm
Raw Normal View History

Added a systime implementation by Stefan Haubenthal with a few changes. The routine reads the TOD clock of CIA1 on the C64 and C128. Since systime was a dummy routine common for all CBMs before, this change adds an individual dummy routine for all other CBM systems. CBM510/610 do also have a TOD clock, so a similar function as in the C64 could be used ... git-svn-id: svn://svn.cc65.org/cc65/trunk@3974 b7a2c559-68d2-44c3-8de9-860c34a00d81
2009-07-27 18:44:37 +00:00
;
; Stefan Haubenthal, 27.7.2009
Replaced _systime with clock_gettime. We want to add the capability to not only get the time but also set the time, but there's no "setter" for the "getter" time(). The first ones that come into mind are gettimeofday() and settimeofday(). However, they take a struct timezone argument that doesn't make sense - even the man pages says "The use of the timezone structure is obsolete; the tz argument should normally be specified as NULL." And POSIX says "Applications should use the clock_gettime() function instead of the obsolescent gettimeofday() function." The ...timeofday() functions work with microseconds while the clock_...time() functions work with nanoseconds. Given that we expect our targets to support only 1/10 of seconds the microseconds look preferable at first sight. However, already microseconds require the cc65 data type 'long' so it's not such a relevant difference to nanoseconds. Additionally clock_getres() seems useful. In order to avoid code duplication clock_gettime() takes over the role of the actual time getter from _systime(). So time() now calls clock_gettime() instead of _systime(). For some reason beyond my understanding _systime() was mentioned in time.h. _systime() worked exactly like e.g. _sysremove() and those _sys...() functions are all considered internal. The only reason I could see would be a performance gain of bypassing the time() wrapper. However, all known _systime() implementations internally called mktime(). And mktime() is implemented in C using an iterative algorithm so I really can't see what would be left to gain here. From that perspective I decided to just remove _systime().
2018-08-15 13:59:11 +00:00
; Oliver Schmidt, 14.8.2018
Added a systime implementation by Stefan Haubenthal with a few changes. The routine reads the TOD clock of CIA1 on the C64 and C128. Since systime was a dummy routine common for all CBMs before, this change adds an individual dummy routine for all other CBM systems. CBM510/610 do also have a TOD clock, so a similar function as in the C64 could be used ... git-svn-id: svn://svn.cc65.org/cc65/trunk@3974 b7a2c559-68d2-44c3-8de9-860c34a00d81
2009-07-27 18:44:37 +00:00
;
Adjusted comments to match actual prototypes.
2018-11-06 10:13:23 +00:00
; int __fastcall__ clock_gettime (clockid_t clk_id, struct timespec *tp);
Added a systime implementation by Stefan Haubenthal with a few changes. The routine reads the TOD clock of CIA1 on the C64 and C128. Since systime was a dummy routine common for all CBMs before, this change adds an individual dummy routine for all other CBM systems. CBM510/610 do also have a TOD clock, so a similar function as in the C64 could be used ... git-svn-id: svn://svn.cc65.org/cc65/trunk@3974 b7a2c559-68d2-44c3-8de9-860c34a00d81
2009-07-27 18:44:37 +00:00
;
.include "time.inc"
.include "c64.inc"
Replaced _systime with clock_gettime. We want to add the capability to not only get the time but also set the time, but there's no "setter" for the "getter" time(). The first ones that come into mind are gettimeofday() and settimeofday(). However, they take a struct timezone argument that doesn't make sense - even the man pages says "The use of the timezone structure is obsolete; the tz argument should normally be specified as NULL." And POSIX says "Applications should use the clock_gettime() function instead of the obsolescent gettimeofday() function." The ...timeofday() functions work with microseconds while the clock_...time() functions work with nanoseconds. Given that we expect our targets to support only 1/10 of seconds the microseconds look preferable at first sight. However, already microseconds require the cc65 data type 'long' so it's not such a relevant difference to nanoseconds. Additionally clock_getres() seems useful. In order to avoid code duplication clock_gettime() takes over the role of the actual time getter from _systime(). So time() now calls clock_gettime() instead of _systime(). For some reason beyond my understanding _systime() was mentioned in time.h. _systime() worked exactly like e.g. _sysremove() and those _sys...() functions are all considered internal. The only reason I could see would be a performance gain of bypassing the time() wrapper. However, all known _systime() implementations internally called mktime(). And mktime() is implemented in C using an iterative algorithm so I really can't see what would be left to gain here. From that perspective I decided to just remove _systime().
2018-08-15 13:59:11 +00:00
.importzp sreg, tmp1, tmp2
.import pushax, pusheax, tosmul0ax, steaxspidx, incsp1, return0
Added clock_settime() for some CBMs. The CIA TOD only stores the time but not the date. Therefore the date set by clock_settime() ist just stored inside the C library for retrieval via clock_gettime(). The "very special" handling of 12AM/PM is based on https://groups.google.com/d/msg/comp.sys.cbm/ysVYSX4AMbc/vHrXCWEhCOUJ saying: ========== 24hr: Wr => Rd => Nx -------------------- 0 : 92 => 12 => 01 <= Switch from 00 to 01 (24-hour notation) 1 : 01 => 01 => 02 2 : 02 => 02 => 03 11 : 11 => 11 => 92 12 : 12 => 92 => 81 <= Switch from 12 to 13 (24-hour notation) 13 : 81 => 81 => 82 14 : 82 => 82 => 83 23 : 91 => 91 => 12 1. column ("24hr"): hour to be tested (decimal) 2. column ("Wr"): hour written to TOD register (BCD) 3. column ("Rd"): hour read from TOD register (BCD) immediately after writing the value in column 2 to see the conversion between AM/PM, if any 4. column ("Nx"): next hour (BCD) after the hour switch ========== Thanks Paul!
2018-08-17 23:13:02 +00:00
.import TM, load_tenth
Added a systime implementation by Stefan Haubenthal with a few changes. The routine reads the TOD clock of CIA1 on the C64 and C128. Since systime was a dummy routine common for all CBMs before, this change adds an individual dummy routine for all other CBM systems. CBM510/610 do also have a TOD clock, so a similar function as in the C64 could be used ... git-svn-id: svn://svn.cc65.org/cc65/trunk@3974 b7a2c559-68d2-44c3-8de9-860c34a00d81
2009-07-27 18:44:37 +00:00
The CIA TOD needs a write to the 1/10 sec register to start. Shortened the code by using initialized data. git-svn-id: svn://svn.cc65.org/cc65/trunk@4224 b7a2c559-68d2-44c3-8de9-860c34a00d81
2009-09-24 19:40:36 +00:00
;----------------------------------------------------------------------------
Added a systime implementation by Stefan Haubenthal with a few changes. The routine reads the TOD clock of CIA1 on the C64 and C128. Since systime was a dummy routine common for all CBMs before, this change adds an individual dummy routine for all other CBM systems. CBM510/610 do also have a TOD clock, so a similar function as in the C64 could be used ... git-svn-id: svn://svn.cc65.org/cc65/trunk@3974 b7a2c559-68d2-44c3-8de9-860c34a00d81
2009-07-27 18:44:37 +00:00
.code
Replaced _systime with clock_gettime. We want to add the capability to not only get the time but also set the time, but there's no "setter" for the "getter" time(). The first ones that come into mind are gettimeofday() and settimeofday(). However, they take a struct timezone argument that doesn't make sense - even the man pages says "The use of the timezone structure is obsolete; the tz argument should normally be specified as NULL." And POSIX says "Applications should use the clock_gettime() function instead of the obsolescent gettimeofday() function." The ...timeofday() functions work with microseconds while the clock_...time() functions work with nanoseconds. Given that we expect our targets to support only 1/10 of seconds the microseconds look preferable at first sight. However, already microseconds require the cc65 data type 'long' so it's not such a relevant difference to nanoseconds. Additionally clock_getres() seems useful. In order to avoid code duplication clock_gettime() takes over the role of the actual time getter from _systime(). So time() now calls clock_gettime() instead of _systime(). For some reason beyond my understanding _systime() was mentioned in time.h. _systime() worked exactly like e.g. _sysremove() and those _sys...() functions are all considered internal. The only reason I could see would be a performance gain of bypassing the time() wrapper. However, all known _systime() implementations internally called mktime(). And mktime() is implemented in C using an iterative algorithm so I really can't see what would be left to gain here. From that perspective I decided to just remove _systime().
2018-08-15 13:59:11 +00:00
.proc _clock_gettime
jsr pushax
jsr pushax
Removed (pretty inconsistently used) tab chars from source code base.
2013-05-09 11:56:54 +00:00
lda CIA1_TODHR
sed
Fixed 12 AM/PM handling. Midnight is 12 AM and noon is 12 PM (see https://en.wikipedia.org/wiki/12-hour_clock). Therefore we need to subtract 12 hours in exactly those two hours.
2018-08-17 22:13:38 +00:00
tax ; Save PM flag
and #%01111111
cmp #$12 ; 12 AM/PM
bcc @L1
sbc #$12
@L1: inx ; Get PM flag
bpl @L2
Removed (pretty inconsistently used) tab chars from source code base.
2013-05-09 11:56:54 +00:00
clc
adc #$12
Fixed 12 AM/PM handling. Midnight is 12 AM and noon is 12 PM (see https://en.wikipedia.org/wiki/12-hour_clock). Therefore we need to subtract 12 hours in exactly those two hours.
2018-08-17 22:13:38 +00:00
@L2: cld
jsr BCD2dec
Removed (pretty inconsistently used) tab chars from source code base.
2013-05-09 11:56:54 +00:00
sta TM + tm::tm_hour
lda CIA1_TODMIN
jsr BCD2dec
sta TM + tm::tm_min
lda CIA1_TODSEC
jsr BCD2dec
sta TM + tm::tm_sec
lda #<TM
ldx #>TM
Replaced _systime with clock_gettime. We want to add the capability to not only get the time but also set the time, but there's no "setter" for the "getter" time(). The first ones that come into mind are gettimeofday() and settimeofday(). However, they take a struct timezone argument that doesn't make sense - even the man pages says "The use of the timezone structure is obsolete; the tz argument should normally be specified as NULL." And POSIX says "Applications should use the clock_gettime() function instead of the obsolescent gettimeofday() function." The ...timeofday() functions work with microseconds while the clock_...time() functions work with nanoseconds. Given that we expect our targets to support only 1/10 of seconds the microseconds look preferable at first sight. However, already microseconds require the cc65 data type 'long' so it's not such a relevant difference to nanoseconds. Additionally clock_getres() seems useful. In order to avoid code duplication clock_gettime() takes over the role of the actual time getter from _systime(). So time() now calls clock_gettime() instead of _systime(). For some reason beyond my understanding _systime() was mentioned in time.h. _systime() worked exactly like e.g. _sysremove() and those _sys...() functions are all considered internal. The only reason I could see would be a performance gain of bypassing the time() wrapper. However, all known _systime() implementations internally called mktime(). And mktime() is implemented in C using an iterative algorithm so I really can't see what would be left to gain here. From that perspective I decided to just remove _systime().
2018-08-15 13:59:11 +00:00
jsr _mktime
ldy #timespec::tv_sec
jsr steaxspidx ; Pops address pushed by 2. pushax
Added clock_settime() for some CBMs. The CIA TOD only stores the time but not the date. Therefore the date set by clock_settime() ist just stored inside the C library for retrieval via clock_gettime(). The "very special" handling of 12AM/PM is based on https://groups.google.com/d/msg/comp.sys.cbm/ysVYSX4AMbc/vHrXCWEhCOUJ saying: ========== 24hr: Wr => Rd => Nx -------------------- 0 : 92 => 12 => 01 <= Switch from 00 to 01 (24-hour notation) 1 : 01 => 01 => 02 2 : 02 => 02 => 03 11 : 11 => 11 => 92 12 : 12 => 92 => 81 <= Switch from 12 to 13 (24-hour notation) 13 : 81 => 81 => 82 14 : 82 => 82 => 83 23 : 91 => 91 => 12 1. column ("24hr"): hour to be tested (decimal) 2. column ("Wr"): hour written to TOD register (BCD) 3. column ("Rd"): hour read from TOD register (BCD) immediately after writing the value in column 2 to see the conversion between AM/PM, if any 4. column ("Nx"): next hour (BCD) after the hour switch ========== Thanks Paul!
2018-08-17 23:13:02 +00:00
jsr load_tenth
Replaced _systime with clock_gettime. We want to add the capability to not only get the time but also set the time, but there's no "setter" for the "getter" time(). The first ones that come into mind are gettimeofday() and settimeofday(). However, they take a struct timezone argument that doesn't make sense - even the man pages says "The use of the timezone structure is obsolete; the tz argument should normally be specified as NULL." And POSIX says "Applications should use the clock_gettime() function instead of the obsolescent gettimeofday() function." The ...timeofday() functions work with microseconds while the clock_...time() functions work with nanoseconds. Given that we expect our targets to support only 1/10 of seconds the microseconds look preferable at first sight. However, already microseconds require the cc65 data type 'long' so it's not such a relevant difference to nanoseconds. Additionally clock_getres() seems useful. In order to avoid code duplication clock_gettime() takes over the role of the actual time getter from _systime(). So time() now calls clock_gettime() instead of _systime(). For some reason beyond my understanding _systime() was mentioned in time.h. _systime() worked exactly like e.g. _sysremove() and those _sys...() functions are all considered internal. The only reason I could see would be a performance gain of bypassing the time() wrapper. However, all known _systime() implementations internally called mktime(). And mktime() is implemented in C using an iterative algorithm so I really can't see what would be left to gain here. From that perspective I decided to just remove _systime().
2018-08-15 13:59:11 +00:00
jsr pusheax
lda CIA1_TOD10
ldx #>$0000
jsr tosmul0ax
ldy #timespec::tv_nsec
jsr steaxspidx ; Pops address pushed by 1. pushax
jsr incsp1
jmp return0
Added a systime implementation by Stefan Haubenthal with a few changes. The routine reads the TOD clock of CIA1 on the C64 and C128. Since systime was a dummy routine common for all CBMs before, this change adds an individual dummy routine for all other CBM systems. CBM510/610 do also have a TOD clock, so a similar function as in the C64 could be used ... git-svn-id: svn://svn.cc65.org/cc65/trunk@3974 b7a2c559-68d2-44c3-8de9-860c34a00d81
2009-07-27 18:44:37 +00:00
Replaced _systime with clock_gettime. We want to add the capability to not only get the time but also set the time, but there's no "setter" for the "getter" time(). The first ones that come into mind are gettimeofday() and settimeofday(). However, they take a struct timezone argument that doesn't make sense - even the man pages says "The use of the timezone structure is obsolete; the tz argument should normally be specified as NULL." And POSIX says "Applications should use the clock_gettime() function instead of the obsolescent gettimeofday() function." The ...timeofday() functions work with microseconds while the clock_...time() functions work with nanoseconds. Given that we expect our targets to support only 1/10 of seconds the microseconds look preferable at first sight. However, already microseconds require the cc65 data type 'long' so it's not such a relevant difference to nanoseconds. Additionally clock_getres() seems useful. In order to avoid code duplication clock_gettime() takes over the role of the actual time getter from _systime(). So time() now calls clock_gettime() instead of _systime(). For some reason beyond my understanding _systime() was mentioned in time.h. _systime() worked exactly like e.g. _sysremove() and those _sys...() functions are all considered internal. The only reason I could see would be a performance gain of bypassing the time() wrapper. However, all known _systime() implementations internally called mktime(). And mktime() is implemented in C using an iterative algorithm so I really can't see what would be left to gain here. From that perspective I decided to just remove _systime().
2018-08-15 13:59:11 +00:00
.endproc
;----------------------------------------------------------------------------
Added a systime implementation by Stefan Haubenthal with a few changes. The routine reads the TOD clock of CIA1 on the C64 and C128. Since systime was a dummy routine common for all CBMs before, this change adds an individual dummy routine for all other CBM systems. CBM510/610 do also have a TOD clock, so a similar function as in the C64 could be used ... git-svn-id: svn://svn.cc65.org/cc65/trunk@3974 b7a2c559-68d2-44c3-8de9-860c34a00d81
2009-07-27 18:44:37 +00:00
; dec = (((BCD>>4)*10) + (BCD&0xf))
Replaced _systime with clock_gettime. We want to add the capability to not only get the time but also set the time, but there's no "setter" for the "getter" time(). The first ones that come into mind are gettimeofday() and settimeofday(). However, they take a struct timezone argument that doesn't make sense - even the man pages says "The use of the timezone structure is obsolete; the tz argument should normally be specified as NULL." And POSIX says "Applications should use the clock_gettime() function instead of the obsolescent gettimeofday() function." The ...timeofday() functions work with microseconds while the clock_...time() functions work with nanoseconds. Given that we expect our targets to support only 1/10 of seconds the microseconds look preferable at first sight. However, already microseconds require the cc65 data type 'long' so it's not such a relevant difference to nanoseconds. Additionally clock_getres() seems useful. In order to avoid code duplication clock_gettime() takes over the role of the actual time getter from _systime(). So time() now calls clock_gettime() instead of _systime(). For some reason beyond my understanding _systime() was mentioned in time.h. _systime() worked exactly like e.g. _sysremove() and those _sys...() functions are all considered internal. The only reason I could see would be a performance gain of bypassing the time() wrapper. However, all known _systime() implementations internally called mktime(). And mktime() is implemented in C using an iterative algorithm so I really can't see what would be left to gain here. From that perspective I decided to just remove _systime().
2018-08-15 13:59:11 +00:00
.proc BCD2dec
tax
Removed (pretty inconsistently used) tab chars from source code base.
2013-05-09 11:56:54 +00:00
and #%00001111
sta tmp1
txa
Added a systime implementation by Stefan Haubenthal with a few changes. The routine reads the TOD clock of CIA1 on the C64 and C128. Since systime was a dummy routine common for all CBMs before, this change adds an individual dummy routine for all other CBM systems. CBM510/610 do also have a TOD clock, so a similar function as in the C64 could be used ... git-svn-id: svn://svn.cc65.org/cc65/trunk@3974 b7a2c559-68d2-44c3-8de9-860c34a00d81
2009-07-27 18:44:37 +00:00
and #%11110000 ; *16
lsr ; *8
sta tmp2
lsr
lsr ; *2
adc tmp2 ; = *10
adc tmp1
rts
.endproc
Reference in New Issue Copy Permalink