blob: b00b34b37c5ee62cd104edb2694a8ce3a8b1a4ef (
plain) (
blame)
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
|
Description: Change mincore() to count "mapped" pages rather than "cached" pages
References:
https://www.openwall.com/lists/oss-security/2019/01/07/1
https://www.theregister.co.uk/2019/01/05/boffins_beat_page_cache/
https://arxiv.org/abs/1901.01161
https://bugzilla.suse.com/show_bug.cgi?id=1120843
https://twitter.com/lavados/status/1081205802046607361
https://lore.kernel.org/lkml/CAHk-=wiqbKEC5jUXr3ax+oUuiRrp=QMv_ZnUfO-SPv=UNJ-OTw@mail.gmail.com/T/#mf09fc35229f38e5b657f299a3a3865c8ca39e7e1
Notes:
carnil> 574823bfab82d9d8fa47f422778043fbb4b4f50e was marked explicitly
carnil> not yet for stable as it is expected to check if there are issues
carnil> with this approach. There are enough tools which depend on mincore()
carnil> to try to tell wheter a file is loaded in cache or not.
carnil> 574823bfab82d9d8fa47f422778043fbb4b4f50e was reverted in
carnil> 30bac164aca750892b93eef350439a0562a68647 again to open way for an
carnil> alternative approach of solving the original information leak issue.
carnil> Further approach https://lore.kernel.org/lkml/20190130124420.1834-1-vbabka@suse.cz/
carnil> Upstream 134fca9063ad ("mm/mincore.c: make mincore() more
carnil> conservative") relates to this.
Bugs:
upstream: released (5.2-rc1) [134fca9063ad4851de767d1768180e5dede9a881]
4.19-upstream-stable: released (4.19.45) [f580a54bbd522f2518fd642f7d4d73ad728e5d58]
4.9-upstream-stable: released (4.9.178) [fef85fb00224cb22f9efec262e9ffe8970e555ee]
3.16-upstream-stable: released (3.16.69) [b96659f18c61120dbf8b4cc36fbc05589bf9dc02]
sid: released (4.19.37-4) [bugfix/all/mm-mincore.c-make-mincore-more-conservative.patch]
4.9-stretch-security: released (4.9.168-1+deb9u3) [bugfix/all/mm-mincore.c-make-mincore-more-conservative.patch]
3.16-jessie-security: released (3.16.68-2) [bugfix/all/mm-mincore.c-make-mincore-more-conservative.patch]
|
