-
Notifications
You must be signed in to change notification settings - Fork 2
/
errata.yaml
48 lines (40 loc) · 1.8 KB
/
errata.yaml
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
- date: 2020-June-23
errata:
- test: check5
description: Missing clobber in P0
- test: MP.RT.inv+dmb+addr
description: Incorrect handler made it impossible to tell difference between pagefault and new translation
- date: 2020-July-22
errata:
- test: '@all'
description: >
fixed race condition in locking in the harness that could sometimes
cause the tool to deadlock. Whether it caused other issues is unknown.
- test: MP.RT.inv+dmb+addr
description: Incorrect handler made it impossible to tell difference between pagefault and new translation
- test: CoWT.inv CoWT.inv+dmb CoWT.inv+dsb CoWT.inv+dsb-isb
description: >
Fixed sync handlers. It used to set the sync handler for the incorrect (non-existent) thread.
Also the handler had no ERET.
- test: MP.TT.inv+dmb+tpo MP.TT.inv+dmb+dsb-isb
description: Fixed register write that caused final state to be incorrect.
- date: 2020-August-10
errata:
- test: check1 check2
description: >
fixed case where if _both_ loads faulted then the second load's exception handler would also fault
causing the entire harness to crash.
More generally tests need to be more resilient to unexpected faults and have more detection.
- date: 2020-September-28
errata:
- test: '@aliasing'
since: 2020-September-14
description: >
an alias of the first variable always resulted in no aliasing happening
- date: 2020-October-8
errata:
- test: '@pgtable'
description: >
insufficient locking around vmm_ensure_level meant that during per-thread initialization
of the state could mean we could have returned a pointer to an entry that was subsequently
overwritten by another vmm_ensure_level call leaving it dangling and irrelevant.