similar to: [PATCH 1/2] export cpu_tlbstate to modules

Twitter user stintel, in this thread: https://twitter.com/stintel/status/948499157282623488 mentions a possible problem with the new patches and the nvidia driver: "As if the @Intel bug isn't bad enough, #KPTI renders @nvidia driver incompatible due to GPL-only symbol 'cpu_tlbstate'. #epicfail" Also: https://twitter.com/tomasz_gwozdz/status/948590364679655429

This adds an smp_ops for voyager, and hooks things up appropriately. This is the first baby-step to making subarch runtime switchable. Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Cc: Eric W. Biederman <ebiederm@xmission.com> --- arch/i386/kernel/Makefile | 1 arch/i386/kernel/smp.c

This adds an smp_ops for voyager, and hooks things up appropriately. This is the first baby-step to making subarch runtime switchable. Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Cc: Eric W. Biederman <ebiederm@xmission.com> --- arch/i386/kernel/Makefile | 1 arch/i386/kernel/smp.c

On Thu, Jul 18, 2019 at 05:58:32PM -0700, Nadav Amit wrote: > @@ -709,8 +716,9 @@ void native_flush_tlb_others(const struct cpumask *cpumask, > * doing a speculative memory access. > */ > if (info->freed_tables) { > - smp_call_function_many(cpumask, flush_tlb_func_remote, > - (void *)info, 1); > + __smp_call_function_many(cpumask, flush_tlb_func_remote,

On Thu, Jul 18, 2019 at 05:58:32PM -0700, Nadav Amit wrote: > @@ -709,8 +716,9 @@ void native_flush_tlb_others(const struct cpumask *cpumask, > * doing a speculative memory access. > */ > if (info->freed_tables) { > - smp_call_function_many(cpumask, flush_tlb_func_remote, > - (void *)info, 1); > + __smp_call_function_many(cpumask, flush_tlb_func_remote,

To improve TLB shootdown performance, flush the remote and local TLBs concurrently. Introduce flush_tlb_multi() that does so. The current flush_tlb_others() interface is kept, since paravirtual interfaces need to be adapted first before it can be removed. This is left for future work. In such PV environments, TLB flushes are not performed, at this time, concurrently. Add a static key to tell

To improve TLB shootdown performance, flush the remote and local TLBs concurrently. Introduce flush_tlb_multi() that does so. The current flush_tlb_others() interface is kept, since paravirtual interfaces need to be adapted first before it can be removed. This is left for future work. In such PV environments, TLB flushes are not performed, at this time, concurrently. Add a static key to tell

I was also able to get the LDT switching functionality out of the critical path in switch_mm, which reduces the number of function calls, potential TLB misses and code size. Signed-off-by: Zachary Amsden <zach@vmware.com> Index: linux-2.6.14-zach-work/include/asm-i386/desc.h =================================================================== ---

I was also able to get the LDT switching functionality out of the critical path in switch_mm, which reduces the number of function calls, potential TLB misses and code size. Signed-off-by: Zachary Amsden <zach@vmware.com> Index: linux-2.6.14-zach-work/include/asm-i386/desc.h =================================================================== ---

Currently, local and remote TLB flushes are not performed concurrently, which introduces unnecessary overhead - each INVLPG can take 100s of cycles. This patch-set allows TLB flushes to be run concurrently: first request the remote CPUs to initiate the flush, then run it locally, and finally wait for the remote CPUs to finish their work. In addition, there are various small optimizations to avoid

These helpers are only for use with kernel threads, and I will tie them more into the kthread infrastructure going forward. Also move the prototypes to kthread.h - mmu_context.h was a little weird to start with as it otherwise contains very low-level MM bits. Signed-off-by: Christoph Hellwig <hch at lst.de> --- drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h | 1 +

[ Cover-letter is identical to v2, including benchmark results, excluding the change log. ] Currently, local and remote TLB flushes are not performed concurrently, which introduces unnecessary overhead - each INVLPG can take 100s of cycles. This patch-set allows TLB flushes to be run concurrently: first request the remote CPUs to initiate the flush, then run it locally, and finally wait for

vhost net module wants to do copy to/from user from a kernel thread, which needs use_mm (like what fs/aio has). Move that into mm/ and export to modules. Acked-by: Andrew Morton <akpm at linux-foundation.org> Signed-off-by: Michael S. Tsirkin <mst at redhat.com> --- fs/aio.c | 47 +---------------------------------- include/linux/mmu_context.h | 9 ++++++

vhost net module wants to do copy to/from user from a kernel thread, which needs use_mm (like what fs/aio has). Move that into mm/ and export to modules. Acked-by: Andrew Morton <akpm at linux-foundation.org> Signed-off-by: Michael S. Tsirkin <mst at redhat.com> --- fs/aio.c | 47 +---------------------------------- include/linux/mmu_context.h | 9 ++++++

vhost net module wants to do copy to/from user from a kernel thread, which needs use_mm (like what fs/aio has). Move that into mm/ and export to modules. Acked-by: Andrew Morton <akpm at linux-foundation.org> Acked-by: Andrea Arcangeli <aarcange at redhat.com> Signed-off-by: Michael S. Tsirkin <mst at redhat.com> --- fs/aio.c | 47

vhost net module wants to do copy to/from user from a kernel thread, which needs use_mm (like what fs/aio has). Move that into mm/ and export to modules. Acked-by: Andrew Morton <akpm at linux-foundation.org> Acked-by: Andrea Arcangeli <aarcange at redhat.com> Signed-off-by: Michael S. Tsirkin <mst at redhat.com> --- fs/aio.c | 47

Anyone who wants to do copy to/from user from a kernel thread, needs use_mm (like what fs/aio has). Move that into mm/, to make reusing and exporting easier down the line, and make aio use it. Next intended user, besides aio, will be vhost-net. Acked-by: Andrea Arcangeli <aarcange at redhat.com> Signed-off-by: Michael S. Tsirkin <mst at redhat.com> --- fs/aio.c |

Anyone who wants to do copy to/from user from a kernel thread, needs use_mm (like what fs/aio has). Move that into mm/, to make reusing and exporting easier down the line, and make aio use it. Next intended user, besides aio, will be vhost-net. Acked-by: Andrea Arcangeli <aarcange at redhat.com> Signed-off-by: Michael S. Tsirkin <mst at redhat.com> --- fs/aio.c |

Seems to work OK for native and Xen. I had to play a bit with the paravirt-sched-clock patch to deal with the VMI changes. Zach, can you check that it still works? Thanks, J

Seems to work OK for native and Xen. I had to play a bit with the paravirt-sched-clock patch to deal with the VMI changes. Zach, can you check that it still works? Thanks, J