Re: Avoid speculative indirect calls in kernel
Linus Torvalds writes: On Wed, Jan 3, 2018 at 4:12 PM, Thomas Gleixner <tglx@linutronix.de> wrote: [...] used for patching. That would definitely be the right approach. That would definitely be...
View ArticleRe: [RFC PATCH] asm/generic: introduce if_nospec and nospec_barrier
Linus Torvalds writes: (Summary) On Wed, Jan 3, 2018 at 4:15 PM, Dan Williams <dan.j.williams@intel.com> wrote: [...] controlled value is actually consumed. I'm much less worried about these...
View ArticleRe: [RFC PATCH] asm/generic: introduce if_nospec and nospec_barrier
Linus Torvalds writes: (Summary) On Wed, Jan 3, 2018 at 5:51 PM, Dan Williams <dan.j.williams@intel.com> wrote: [...] or so locations that need some 'nospec' handling. I'd love to see that patch,...
View ArticleRe: [PATCH 0/7] IBRS patch series
Linus Torvalds writes: (Summary) On Thu, Jan 4, 2018 at 9:56 AM, Tim Chen <tim.c.chen@linux.intel.com> wrote: [...] be used instead of retpoline[1]. Can somebody explain this part? Can somebody...
View ArticleRe: [PATCH 0/7] IBRS patch series
Linus Torvalds writes: (Summary) wrote: [...] from get empty, you end up vulnerable. That sounds like it could cause mispredicts, but it doesn't sound _exploitable_. That sounds like it could cause...
View ArticleRe: [PATCH 4.4 00/37] 4.4.110-stable review
Linus Torvalds writes: (Summary) in rcu_dynticks_snap() because %rax is 1 and that would give you the invalid page fault and the right faulting address. invalid page fault and the right faulting...
View ArticleRe: [PATCH 4.4 00/37] 4.4.110-stable review
Linus Torvalds writes: (Summary) The only reason you get any kernel stack printout at all is because 'init' dying will make the kernel print that out. dying will make the kernel print that out. The...
View ArticleRe: [RFC PATCH] asm/generic: introduce if_nospec and nospec_barrier
Linus Torvalds writes: (Summary) Which makes me suspect that lfence is a bad idea too. suspect that lfence is a bad idea too. If one common pattern is going to be bounces checking array accesses like...
View ArticleRe: [RFC PATCH] asm/generic: introduce if_nospec and nospec_barrier
Linus Torvalds writes: On Thu, Jan 4, 2018 at 2:20 PM, Linus Torvalds<torvalds@linux-foundation.org> wrote:[...] __u._val; }) That That __u._val = base[_i & _mask]; __u._val = base[_i &...
View ArticleRe: [PATCH 4.4 00/37] 4.4.110-stable review
Linus Torvalds writes: (Summary) wrote:[...] 74 1d Yeah, it's the "lock xadd" as suspected: Yeah, it's the "lock xadd" as suspected: 0: 0f 1f 44 00 00 nopl 0x0(%rax,%rax,1) 5: 55 push %rbp 6: 31 d2 xor...
View ArticleRe: [RFC PATCH] asm/generic: introduce if_nospec and nospec_barrier
Linus Torvalds writes: (Summary) yeah, that's exactly where we want to make sure that the compiler uses a select or 'setb'. uses a select or 'setb'. That's what gcc does for me in testing: That's what...
View ArticleRe: [PATCH 4.4 00/37] 4.4.110-stable review
Linus Torvalds writes: (Summary) And yeah, that seems to be due to NX in the pgd (nothing else is certainly set), with presumably a virtual machine that doesn't support it. it. So I suspect your patch...
View ArticleRe: 4.15-rc6+ hang
Linus Torvalds writes: (Summary) wrote: [...] (VirtualBox version 5.0.40) Any chance of bisecting this? Any chance of bisecting this? I could imagine that all the stuff we now do for page table...
View ArticleRe: [PATCH v3 01/13] x86/retpoline: Add initial retpoline support
Linus Torvalds writes: (Summary) And that one shouldn't have any relocagtion issues, because all the offsets are relative. offsets are relative. We want to use that one for the entry stub anyway, can't...
View ArticleRe: [PATCH v3 01/13] x86/retpoline: Add initial retpoline support
Linus Torvalds writes: (Summary) On Fri, Jan 5, 2018 at 2:00 PM, Woodhouse, David <dwmw@amazon.co.uk> wrote: [...] +.endm Why do these still force a register name? Why do these still force a...
View ArticleRe: [BISECTED] v4.15-rc: Boot regression on x86_64/AMD
Linus Torvalds writes: (Summary) The breakage was supposedly fixed by three patches from Christian: The breakage was supposedly fixed by three patches from Christian: a19e2696135e: "x86/PCI: Only...
View ArticleRe: [PATCH 06/18] x86, barrier: stop speculation for failed access_ok
Linus Torvalds writes: (Summary) Honestly, this seems completely bogus. Honestly, this seems completely bogus. The description is pure garbage afaik. The description is pure garbage afaik. The fact is,...
View ArticleRe: [PATCH 01/18] asm-generic/barrier: add generic nospec helpers
Linus Torvalds writes: (Summary) wrote: [...] +#define nospec_ptr(ptr, lo, hi) \ Do we actually want this horrible interface? Do we actually want this horrible interface? It just causes the compiler -...
View ArticleRe: [PATCH 06/18] x86, barrier: stop speculation for failed access_ok
Linus Torvalds writes: (Summary) On Fri, Jan 5, 2018 at 6:52 PM, Linus Torvalds<torvalds@linux-foundation.org> wrote:[...] access that needs to be non-speculative. I also suspect we should...
View ArticleRe: [PATCH] string: drop __must_check from strscpy() and restore s ...
Linus Torvalds writes: On Fri, Jan 19, 2018 at 8:50 AM, Tejun Heo <tj@kernel.org> wrote: [...] Applied to cgroup/for-4.16. I did still have it queued up, but I was looking at other issues, and...
View Article