faulty Bochs BIOS file
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
bochs (Ubuntu) |
Confirmed
|
Undecided
|
Unassigned |
Bug Description
On a 22.04 system all attempts to use Bochs with its default BIOS file (i.e. /usr/share/
In the following I'll describe steps one could use to try to reproduce my findings. Please note that steps (3) and (4) might look at first glance somewhat convoluted, but my aim was to ensure a high degree of repeatability.
(1) install the required packages
sudo apt install -y bochs-x bochsbios
(2) get the alternative BIOS file
wget https:/
(3) run a minimal testcase (with the faulty BIOS file), and capture the output in a (slightly changed) log file
( echo 6 ; sleep 1 ; echo c ; sleep 1 ; echo q ) \
| bochs -n 2>&1 | sed 's#^[0-9]*##' > fail.log
(4) repeat the testcase with the alternative BIOS file (and select 'Quit', when a failure message pops up)
( echo 6 ; sleep 1 ; echo c ; sleep 1 ; echo q ) \
| bochs -n romimage:
(5) compare the relevant lines from the two log files
grep -E '_read| rom|e\[' fail.log good.log
In my case the result of step (5) looks like:
fail.log:e[ ] Switching off quick start, because no configuration file was found.
fail.log:i[MEM0 ] rom at 0xfffe0000/131072 ('/usr/
fail.log:i[MEM0 ] rom at 0xc0000/38400 ('/usr/
fail.log:i[BIOS ] Starting rombios32
fail.log:e[CPU0 ] interrupt(): vector must be within IDT table limits, IDT.limit = 0x0
fail.log:e[CPU0 ] interrupt(): vector must be within IDT table limits, IDT.limit = 0x0
fail.log:e[CPU0 ] interrupt(): vector must be within IDT table limits, IDT.limit = 0x0
fail.
fail.log:e[CPU0 ] exception(): 3rd (13) exception with no resolution, shutdown status is 00h, resetting
good.log:e[ ] Switching off quick start, because no configuration file was found.
good.log:i[ ] parsing arg 3, romimage:
good.log:i[MEM0 ] rom at 0xfffe0000/131072 ('BIOS-
good.log:i[MEM0 ] rom at 0xc0000/38400 ('/usr/
good.log:i[BIOS ] Starting rombios32
To summarize: in the first run the process fails due to the faulty BIOS, in the second run one ends up with a message about the missing boot medium (which is entirely expected, due to the way the test was set up).
It seems that other users have run into the same problem, but apparently without lodging a bug report:
https:/
PS: I'm attaching my two log files in the hope they might be useful.
Status changed to 'Confirmed' because the bug affects multiple users.