illumos

2022-01-14

05:20 AM illumos gate Bug #14398: consolidate bhyve tests

With the proposed change installed on a test machine, I can run all of the bhyve tests together with the provided run... Patrick Mooney

2022-01-12

05:52 AM illumos gate Bug #14260: bhyve PICs should use unsigned levels

I also checked that the assertion/deassertion dtrace probes added by this wad were firing as part of the pin states c... Patrick Mooney

05:48 AM illumos gate Bug #14260: bhyve PICs should use unsigned levels

Overflowing (or underflowing) the interrupt levels is considered broken behavior from the emulation, and it is expect... Patrick Mooney

05:46 AM illumos gate Bug #14260: bhyve PICs should use unsigned levels

This was tested on the typical battery of guests, with all booting and running as expected. During guest operation, ... Patrick Mooney

2022-01-11

08:43 PM illumos gate Bug #14398 (Closed): consolidate bhyve tests

With the integration of #14367, there are two bhyve test directories shipped with @system/bhyve/tests@:
- @/opt/bhyv... Patrick Mooney

08:24 PM illumos gate Bug #14367: bhyve gpt mishandles small mappings

With the fix in place, the bhyve tests suites pass:... Patrick Mooney

2022-01-05

09:49 PM illumos gate Bug #14367 (Closed): bhyve gpt mishandles small mappings

While andyf was testing VARS.fd support from upstream FreeBSD, he hit this panic:... Patrick Mooney

2022-01-04

09:27 PM illumos gate Bug #14364 (In Progress): blocked eventfd write fails interruption

If a thread is writing to an eventfd handle which would overflow, it will block until some reader "makes room" for th... Patrick Mooney

2022-01-03

10:22 PM illumos gate Bug #14362 (In Progress): eventfd should present kernel API

If device drivers are going to support using eventfd as a notification mechanism, they will need an API to establish ... Patrick Mooney

06:35 PM illumos gate Bug #14361 (In Progress): eventfd should track minor in id_space

The eventfd implementation is currently tracking device minor IDs directly in a raw vmem, rather than using the idspa... Patrick Mooney