Re: [v7, 05/11] x86/retpoline/entry: Convert entry assembler indirect jumps

2018-01-09 Thread Andrei Vagin
On Tue, Jan 09, 2018 at 08:39:21PM -0800, Dave Hansen wrote: > On 01/09/2018 08:30 PM, Andi Kleen wrote: > > On Tue, Jan 09, 2018 at 07:54:08PM -0800, Andrei Vagin wrote: > >> > >> In my test environment, the kernel with this patch crashes. > > > > I posted a patch for this. > > It's called: > >

Re: [v7, 05/11] x86/retpoline/entry: Convert entry assembler indirect jumps

2018-01-09 Thread Dave Hansen
On 01/09/2018 08:30 PM, Andi Kleen wrote: > On Tue, Jan 09, 2018 at 07:54:08PM -0800, Andrei Vagin wrote: >> >> In my test environment, the kernel with this patch crashes. > > I posted a patch for this. It's called: [PATCH v3 2/3] x86/retpoline: Use better sequences for NOSPEC_CALL/JMP right, A

Re: [v7, 05/11] x86/retpoline/entry: Convert entry assembler indirect jumps

2018-01-09 Thread Andi Kleen
On Tue, Jan 09, 2018 at 07:54:08PM -0800, Andrei Vagin wrote: > > Hi, > > In my test environment, the kernel with this patch crashes. I posted a patch for this. -Andi

Re: [v7, 05/11] x86/retpoline/entry: Convert entry assembler indirect jumps

2018-01-09 Thread Andrei Vagin
Hi, In my test environment, the kernel with this patch crashes. https://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git/commit/?h=x86/pti&id=f3433c1010c6af61c9897f0f0447f81b991feac1 My config and a kernel log are attached. [1.985901] Freeing unused kernel memory: 524K [1.987505] ro

[PATCH v7 05/11] x86/retpoline/entry: Convert entry assembler indirect jumps

2018-01-09 Thread David Woodhouse
Convert indirect jumps in core 32/64bit entry assembler code to use non-speculative sequences when CONFIG_RETPOLINE is enabled. Don't use CALL_NOSPEC in entry_SYSCALL_64_fastpath because the return address after the 'call' instruction must be *precisely* at the .Lentry_SYSCALL_64_after_fastpath la