In PCXT mode, with the 5160 BIOS (built by the included script this week), I have problems when rebooting with CTRL-ALT-DEL. Sometimes, I get into a state when only alternating keypresses are recognized. More commonly, I get a 101 POST error after rebooting.
I've tried this both with and without the XT-IDE bios enabled, and I get the same result. I don't have this problem with non-IBM BIOSes.
Is this "normal" right now, or am I hitting an unusual problem?
Thanks,
Michael
This issue seems to occur when a soft reset performs a reset operation on an interrupt or timer module.
Perhaps the soft reset implementation of the chip(8259 or 8253) is inappropriate.
I will investigate this when I have time.
As part of the BIOS diagnostics, there is an IMR test for the 8259A.
In that test, when IMR = 0h and Timer 0 output = H, the issue seems to occur.
In other words, the timing at which CTRL-ALT-DEL is input determines whether 101 occurs or not.
The 8259 test in the normal reset state is before Timer 0 starts counting, so no interrupt is generated even if IMR = 0h.
However, with CTRL-ALT-DEL, the issue occurs because the 8259 test is started while Timer 0 is already running.
I am currently trying to come up with a solution.
However, these solutions may affect other behaviors.
Therefore, this issue may remain for the time being.
For now, I recommend using the OSD or reset switch.
Even when 101 does not occur, I have experienced the 601 error or the keyboard not working.
These are resolved in the following branches.
kitune-san wrote: ↑Sun Mar 12, 2023 11:21 amI am currently trying to come up with a solution.
However, these solutions may affect other behaviors.
Therefore, this issue may remain for the time being.
kitune-san wrote: ↑Sun Mar 12, 2023 11:31 amEven when 101 does not occur, I have experienced the 601 error or the keyboard not working.
These are resolved in the following branches.
In addition, adding the following commits will eliminate the 101 error.
However, keyboard input during diagnostics will cause the keyboard to get stuck.
https://github.com/kitune-san/PCXT_MiST ... 00966f4b0f
Probably difficult to improve further.
kitune-san wrote: ↑Sun Mar 12, 2023 3:24 pmIn addition, adding the following commits will eliminate the 101 error.
However, keyboard input during diagnostics will cause the keyboard to get stuck.https://github.com/kitune-san/PCXT_MiST ... 00966f4b0f
Probably difficult to improve further.
Hmmm. Sometimes the keyboard stops working. (e.g. Area5150) 