1 |
On Fri, Aug 17, 2018 at 8:57 AM Alexander Kapshuk |
2 |
<alexander.kapshuk@×××××.com> wrote: |
3 |
> |
4 |
> On Fri, Aug 17, 2018 at 8:10 AM <tuxic@××××××.de> wrote: |
5 |
> > |
6 |
> > On 08/17 02:53, Adam Carter wrote: |
7 |
> > > On Fri, Aug 17, 2018 at 1:15 PM, <tuxic@××××××.de> wrote: |
8 |
> > > |
9 |
> > > > Hi, |
10 |
> > > > |
11 |
> > > > CPU bugs seem to be more and more common: |
12 |
> > > > https://www.heise.de/security/meldung/Linux-Kernel-und- |
13 |
> > > > Distributionen-schuetzen-vor-Prozessorluecke-Foreshadow-L1TF-4137264.html |
14 |
> > > > https://www.heise.de/security/meldung/Spectre-NG-Foreshadow- |
15 |
> > > > gefaehrdet-Intel-Prozessoren-4137209.html |
16 |
> > > > (sorry, I only know of this german spoken references...) |
17 |
> > > > |
18 |
> > > > With Linux kernel 4.18.1 Linus has introduced a fix (aka workaround) |
19 |
> > > > of the Foreshadow bug. |
20 |
> > > > |
21 |
> > > |
22 |
> > > 4.18, 4.17, 4.14, 4.9, and 4.4 have all had the fixes applied. |
23 |
> > > |
24 |
> > > > |
25 |
> > > > Unfortunately compiling that kernel (as downloaded from |
26 |
> > > > https://mirrors.edge.kernel.org/pub/linux/kernel/v4.x/ ) |
27 |
> > > > |
28 |
> > > > gives me this bug: |
29 |
> > > > |
30 |
> > > |
31 |
> > > gentoo-sources with gcc 7.3 builds fine for me. |
32 |
> > > |
33 |
> > > Intel: grep . /sys/devices/system/cpu/vulnerabilities/* |
34 |
> > > /sys/devices/system/cpu/vulnerabilities/l1tf:Mitigation: PTE Inversion |
35 |
> > > /sys/devices/system/cpu/vulnerabilities/meltdown:Mitigation: PTI |
36 |
> > > /sys/devices/system/cpu/vulnerabilities/spec_store_bypass:Mitigation: |
37 |
> > > Speculative Store Bypass disabled via prctl and seccomp |
38 |
> > > /sys/devices/system/cpu/vulnerabilities/spectre_v1:Mitigation: __user |
39 |
> > > pointer sanitization |
40 |
> > > /sys/devices/system/cpu/vulnerabilities/spectre_v2:Mitigation: Full generic |
41 |
> > > retpoline, IBPB, IBRS_FW |
42 |
> > > |
43 |
> > > AMD: grep . /sys/devices/system/cpu/vulnerabilities/* |
44 |
> > > /sys/devices/system/cpu/vulnerabilities/l1tf:Not affected |
45 |
> > > /sys/devices/system/cpu/vulnerabilities/meltdown:Not affected |
46 |
> > > /sys/devices/system/cpu/vulnerabilities/spec_store_bypass:Mitigation: |
47 |
> > > Speculative Store Bypass disabled via prctl and seccomp |
48 |
> > > /sys/devices/system/cpu/vulnerabilities/spectre_v1:Mitigation: __user |
49 |
> > > pointer sanitization |
50 |
> > > /sys/devices/system/cpu/vulnerabilities/spectre_v2:Mitigation: Full AMD |
51 |
> > > retpoline, IBPB |
52 |
> > |
53 |
> > Hi, |
54 |
> > |
55 |
> > I am happy, that other sources do work for you Adam. |
56 |
> > |
57 |
> > Interesting would be, why the original sources does not compile for |
58 |
> > me. |
59 |
> > Any idea? |
60 |
> > |
61 |
> > |
62 |
> > |
63 |
> |
64 |
> This problem has been reported upstream. See below. |
65 |
> https://lkml.org/lkml/2018/8/15/118 |
66 |
> |
67 |
> In particular: |
68 |
> |
69 |
> Build is successful with |
70 |
> CONFIG_KVM=y |
71 |
> CONFIG_KVM_INTEL=y |
72 |
> CONFIG_KVM_AMD=y |
73 |
> but fails if only |
74 |
> CONFIG_KVM=y |
75 |
> CONFIG_KVM_AMD=y |
76 |
> are selected. |
77 |
|
78 |
And a follow up email with a reference to a patch that addresses the issue: |
79 |
https://lkml.org/lkml/2018/8/15/122 |