system/docs: update docs and SF constants to match actual kernel #306
dunkaist
Sweetbread
mxlgv
Doczom
Reference in New Issue
Block a user
Delete Branch "docs-review"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
I don't think new aliases are necessary. It just clutter the code. Except:
SF_THREAD_CONTROL@@ -3691,0 +3695,4 @@* ecx = требуемый размер в байтахВозвращаемое значение:* eax = 0 - неудача* eax = указатель на выделенное кольцо"Кольцевой буфер/кольцевую память" maybe?
@@ -4517,3 +4548,3 @@eax - SF_BLITTER (73)========================================================================================= Функция 74, подфункция -1 =========================================== Функция 74, подфункция 255 ========================Is this really necessary?
@@ -5095,2 +5164,3 @@Возвращаемое значение:* eax = 0 если успех, иначе ошибка.* eax = 0 если успех, иначе отрицательный код ошибки:-EFAULT, -EINVAL, -EMFILE, -ENFILE.I'm not sure if this is specifically stated somewhere in the documentation (error codes).
@@ -5098,3 +5168,3 @@---------------------- Константы для регистров: ----------------------eax - SF_FUTEX (77)ebx - ...ebx - 13Why is there no constant?
@@ -3619,3 +3681,3 @@---------------------- Constants for registers: ----------------------eax - SF_SYS_MISC (68)ebx - SSF_MEM_ALLOC_RING (29)ebx - 30Why is there no constant?
@@ -3834,3 +3896,3 @@ebx - SSF_RESUME (5)======================================================================= Fucntion 69, subfunction 6 - read from memory of debugged process. === Function 69, subfunction 6 - read from memory of debugged process. ==Seems extra =
@@ -4452,3 +4514,3 @@---------------------- Constants for registers: ----------------------eax - SF_NETWORK_GET (74)eax - SF_NETWORK_GET (74) / SF_NETWORK_DEVICE (74)I suspect this is a little-used thing, let's not make aliases and just replace it with something that is often used
@@ -5251,0 +5327,4 @@* Subfunctions 4-7 are reserved and currently return -1.* Subfunctions 8, 9 and 12 are not implemented and return -EBADF.================================================================================= Function 77, Subfunction 10, Read from pipe. ============Maybe
-instead of,?@@ -5251,0 +5342,4 @@---------------------- Constants for registers: ----------------------eax - SF_FUTEX (77)ebx - 10Const?
@@ -5251,0 +5377,4 @@---------------------- Constants for registers: ----------------------eax - SF_FUTEX (77)ebx - 13Const?
New commits pushed, approval review dismissed automatically according to repository settings
@@ -5074,3 +5141,4 @@* esi = сколько байт записатьВозвращаемое значение:* eax = количество записанных байт-EBADF при ошибкеAre these constants documented somewhere?
@@ -132,6 +132,11 @@ SF_STYLE_SETTINGS=48SF_APM=49SF_SET_WINDOW_SHAPE=50SF_CREATE_THREAD=51SF_THREAD_CONTROL=51I suggest getting rid of this
SF_CREATE_THREADconstant and not creating aliases.SF_CREATE_THREADis used in some programs.It should be changed everywhere in a separate PR.
I deleted alias
SF_THREAD_CONTROLfor now,It will be decided here: #321
@@ -187,3 +192,1 @@SSF_WAIT_SIGNAL=14 ; wait for signal from another program/driverSSF_LOAD_DRIVER=16SSF_CONTROL_DRIVER=17SSF_WAIT_SIGNAL=14 ; GetEvent wrapper (wait for kernel event)What is
GetEvent?@@ -190,0 +192,4 @@SSF_WAIT_SIGNAL=14 ; GetEvent wrapper (wait for kernel event)SSF_GET_SERVICE=16SSF_CALL_SERVICE=17SSF_LOAD_DRIVER=16 ; alias (deprecated name)What's the problem with old names?
@@ -70,11 +70,11 @@ SF_SYSTEM=18SSF_WINDOW_BEHAVIOR=25 ; window focus relation with other windowsLet's not touch this file. Fix it as part of this task: #317
Please roll back this file!
@@ -5054,4 +5117,3 @@Параметры:* eax = 77 - номер функции* ebx = 10 - номер подфункции* ecx = дескриптор файлаThere is no mistake here. Reverse this change
@@ -5049,20 +5108,28 @@ Remarks:eax - SF_FUTEX (77)ebx - SSF_WAKE (3)============================================================================= Функция 77, подфункция 10, Прочитать из файла в буфер. =======There is no mistake here. Reverse this change.
@@ -3369,3 +3375,3 @@* буфер, на который указывает ecx, содержит следующую информацию:* +0: dword: идентификатор последующих данных сигнала* +4: данные принятого сигнала (20 байт), формат которых* +4: 5 dword: данные принятого сигнала, формат которыхIn my opinion, 20 bytes is more obvious than 5 dword
In all other similar places, dwords are used, not bytes.
So it is better for consistency.
@@ -3395,3 +3405,3 @@---------------------- Константы для регистров: ----------------------eax - SF_SYS_MISC (68)ebx - SSF_LOAD_DRIVER (16)ebx - SSF_LOAD_DRIVER (16) / SSF_GET_SERVICE (16)In my opinion, services can be confused with daemons. The old name reflects the essence more.
@@ -3418,3 +3430,3 @@---------------------- Константы для регистров: ----------------------eax - SF_SYS_MISC (68)ebx - SSF_CONTROL_DRIVER (17)ebx - SSF_CONTROL_DRIVER (17) / SSF_CALL_SERVICE (17)The same: https://git.kolibrios.org/KolibriOS/kolibrios/pulls/306/files#issuecomment-3987
@@ -4522,3 +4553,3 @@Параметры:* eax = 74 - номер функции* bl = -1 - номер подфункции* bl = -1 (или 255) - номер подфункцииThis should be obvious to the developer. Why explain this?
@@ -4946,3 +4978,3 @@---------------------- Константы для регистров: ----------------------eax - SF_NETWORK_SOCKET (75)bl - SSF_GET_PAIR (10)bl - SSF_GET_PAIR (10) / SSF_SOCKET_PAIR (10)If this is needed for "getting" why change it?
@@ -5033,3 +5091,4 @@-2 - контрольное значение не соответствует---------------------- Константы для регистров: ----------------------eax - SF_FUTEX (77)Since in function 77 it is now better to replace not only futexes with
SF_POSIX. And add prefixes for subfunctions likeSF_FUTEX_CREATE, etc.As these constants are used in some programs, that should be changed in a separate PR.
Will be fixed as part of this task: #321
@@ -5051,2 +5110,3 @@============================================================================= Функция 77, подфункция 10, Прочитать из файла в буфер. =======Замечания:* Подфункции 4-7 зарезервированы и сейчас возвращают -1.Add here that for now only pipe file descriptors are supported
@@ -5063,3 +5129,3 @@---------------------- Константы для регистров: ----------------------eax - SF_FUTEX (77)ebx - ...ebx - SSF_PIPE_READ (10)It is assumed that reading will be in the future on file descriptors. It is better to remove the
PIPEprefix if the constant names are not taken.@@ -5078,3 +5147,3 @@---------------------- Константы для регистров: ----------------------eax - SF_FUTEX (77)ebx - ...ebx - SSF_PIPE_WRITE (11)Same: https://git.kolibrios.org/KolibriOS/kolibrios/pulls/306/files#issuecomment-3992
@@ -5025,4 +5083,2 @@* eax = 77 - номер функции* ebx = 2 - номер подфункции* ecx = дескриптор фьютекса* edx = контрольное значениеIt was written correctly. It is a control value and not a pointer. Please roll back
@Burer Control value. NOT a pointer. Please note:
@@ -4997,3 +5055,3 @@* eax = 77 - номер функции* ebx = 0 - номер подфункции* ecx = указатель на контрольное значение фьютекса (dword)* ecx = указатель на контрольное значение фьютексаIt’s true that this is a pointer to dword. Anything larger than dword will probably be ignored.
@@ -5000,2 +5057,3 @@* ecx = указатель на контрольное значение фьютексаВозвращаемое значение:* eax = дескриптор фьютекса, 0 при ошибке* eax = указатель на дескриптор фьютекса, 0 при ошибкеI apologize, it seems I expressed myself incorrectly. Everything here is correct; it is a descriptor and not a pointer (it doesn’t matter to us how this is implemented in the kernel); even just a pointer can be considered a “descriptor”. Please roll back
@@ -5010,3 +5068,3 @@* eax = 77 - номер функции* ebx = 1 - номер подфункции* ecx = дескриптор фьютекса* ecx = указатель на дескриптор фьютексаPlease roll back. See message above
@@ -5026,3 +5084,2 @@* ebx = 2 - номер подфункции* ecx = дескриптор фьютекса* edx = контрольное значение* ecx = указатель на дескриптор фьютексаPlease roll back. See message above
@@ -5041,3 +5100,3 @@* eax = 77 - номер функции* ebx = 3 - номер подфункции* ecx = дескриптор фьютекса* ecx = указатель на дескриптор фьютексаPlease roll back. See message above
@@ -5053,0 +5112,4 @@* Подфункции 4-7 зарезервированы и сейчас возвращают -1.* Подфункции 8, 9 и 12 не реализованы и возвращают -EBADF (-9).============================================================================= Функция 77, подфункция 10, Прочитать из файла. =======Seems missing
====@@ -5071,3 +5139,2 @@* ebx = 11 - номер подфункции* ecx = дескриптор файла* edx = указатель на буфер, откуда брать данные для записи* ecx = указатель на файловый дескрипторThere is no mistake here. Reverse this change
@@ -5197,11 +5281,12 @@ Remarks:Parameters:* eax = 77 - function number* ebx = 1 - subfunction number* ecx = futex handleThere is no mistake here. Reverse this change
@@ -5181,3 +5269,3 @@* eax = 77 - function number* ebx = 0 - subfunction number* ecx = pointer to futex dword* ecx = pointer to futex control valueCorrect: Pointer to futex dword control value.
@@ -5187,3 +5273,1 @@* Use subfunction 1 to destroy the futex.The kernel destroys the futexes automatically when the processterminates.* eax = pointer to futex handle, 0 on errorThere is no mistake here. Reverse this change
@@ -5215,3 +5300,1 @@* ecx = futex handle* edx = control value* esi = timeout in system ticks or 0 for infinity* ecx = pointer to futex handleThere is no mistake here. Reverse this change
@@ -5216,2 +5300,2 @@* edx = control value* esi = timeout in system ticks or 0 for infinity* ecx = pointer to futex handle* edx = pointer to futex control valueThere is no mistake here. Reverse this change
Control value. NOT a pointer. Please note:
@@ -5235,3 +5315,3 @@* eax = 77 - function number* ebx = 3 - subfunction number* ecx = futex handle* ecx = pointer to futex handleThere is no mistake here. Reverse this change
WIP: system/docs: align docs with actual kernel, add missing SF constantsto system/docs: update docs and SF constants to match actual kernelNew commits pushed, approval review dismissed automatically according to repository settings
@dunkaist @Doczom @Sweetbread Please review this again if possible...