linux_dsm_epyc7002/arch/x86/mm
Ingo Molnar 7366ed771f x86/fpu: Simplify FPU handling by embedding the fpstate in task_struct (again)
So 6 years ago we made the FPU fpstate dynamically allocated:

  aa283f4927 ("x86, fpu: lazy allocation of FPU area - v5")
  61c4628b53 ("x86, fpu: split FPU state from task struct - v5")

In hindsight this was a mistake:

   - it complicated context allocation failure handling, such as:

		/* kthread execs. TODO: cleanup this horror. */
		if (WARN_ON(fpstate_alloc_init(fpu)))
			force_sig(SIGKILL, tsk);

   - it caused us to enable irqs in fpu__restore():

                local_irq_enable();
                /*
                 * does a slab alloc which can sleep
                 */
                if (fpstate_alloc_init(fpu)) {
                        /*
                         * ran out of memory!
                         */
                        do_group_exit(SIGKILL);
                        return;
                }
                local_irq_disable();

   - it (slightly) slowed down task creation/destruction by adding
     slab allocation/free pattens.

   - it made access to context contents (slightly) slower by adding
     one more pointer dereference.

The motivation for the dynamic allocation was two-fold:

   - reduce memory consumption by non-FPU tasks

   - allocate and handle only the necessary amount of context for
     various XSAVE processors that have varying hardware frame
     sizes.

These days, with glibc using SSE memcpy by default and GCC optimizing
for SSE/AVX by default, the scope of FPU using apps on an x86 system is
much larger than it was 6 years ago.

For example on a freshly installed Fedora 21 desktop system, with a
recent kernel, all non-kthread tasks have used the FPU shortly after
bootup.

Also, even modern embedded x86 CPUs try to support the latest vector
instruction set - so they'll too often use the larger xstate frame
sizes.

So remove the dynamic allocation complication by embedding the FPU
fpstate in task_struct again. This should make the FPU a lot more
accessible to all sorts of atomic contexts.

We could still optimize for the xstate frame size in the future,
by moving the state structure to the last element of task_struct,
and allocating only a part of that.

This change is kept minimal by still keeping the ctx_alloc()/free()
routines (that now do nothing substantial) - we'll remove them in
the following patches.

Reviewed-by: Borislav Petkov <bp@alien8.de>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2015-05-19 15:47:49 +02:00
..
kmemcheck x86: Replace __get_cpu_var uses 2014-08-26 13:45:49 -04:00
amdtopology.c x86/mm/numa: Simplify some bit mangling 2013-04-10 19:06:26 +02:00
dump_pagetables.c Merge branch 'x86-mm-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2014-12-10 13:59:34 -08:00
extable.c x86, extable: Switch to relative exception table entries 2012-04-20 17:22:34 -07:00
fault.c x86/asm/entry: Change all 'user_mode_vm()' calls to 'user_mode()' 2015-03-23 11:14:17 +01:00
gup.c mm: convert p[te|md]_numa users to p[te|md]_protnone_numa 2015-02-12 18:54:08 -08:00
highmem_32.c mm: accurately calculate zone->managed_pages for highmem zones 2013-07-03 16:07:33 -07:00
hugetlbpage.c mm/hugetlb: pmd_huge() returns true for non-present hugepage 2015-02-11 17:06:01 -08:00
init_32.c x86: remove the Xen-specific _PAGE_IOMAP PTE flag 2014-09-23 13:36:20 +00:00
init_64.c x86/mm: Use early_param_on_off() for direct_gbpages 2015-03-05 08:02:12 +01:00
init.c Merge branch 'x86-mm-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2015-04-13 13:31:32 -07:00
iomap_32.c x86: Use new cache mode type in mm/iomap_32.c 2014-11-16 11:04:25 +01:00
ioremap.c Merge branch 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2015-05-06 10:57:37 -07:00
kasan_init_64.c kasan: enable instrumentation of global variables 2015-02-13 21:21:42 -08:00
kmmio.c x86: Delete non-required instances of include <linux/init.h> 2014-01-06 21:25:18 -08:00
Makefile mm: move memtest under mm 2015-04-14 16:49:06 -07:00
mm_internal.h x86: Enable PAT to use cache mode translation tables 2014-11-16 11:04:26 +01:00
mmap.c mm: expose arch_mmap_rnd when available 2015-04-14 16:49:05 -07:00
mmio-mod.c x86: delete __cpuinit usage from all x86 files 2013-07-14 19:36:56 -04:00
mpx.c x86/fpu: Simplify FPU handling by embedding the fpstate in task_struct (again) 2015-05-19 15:47:49 +02:00
numa_32.c x86: Fix the initialization of physnode_map 2014-02-01 22:15:51 -08:00
numa_64.c x86, mm: kill numa_free_all_bootmem() 2012-11-17 11:59:47 -08:00
numa_emulation.c x86: delete __cpuinit usage from all x86 files 2013-07-14 19:36:56 -04:00
numa_internal.h x86-32, mm: Rip out x86_32 NUMA remapping code 2013-01-31 14:12:30 -08:00
numa.c x86/mm/numa: Fix kernel stack corruption in numa_init()->numa_clear_kernel_node_hotplug() 2015-04-07 16:01:19 +02:00
pageattr-test.c x86: define _PAGE_NUMA by reusing software bits on the PMD and PTE levels 2014-06-04 16:53:55 -07:00
pageattr.c x86/mm: Simplify enabling direct_gbpages 2015-03-05 08:02:12 +01:00
pat_internal.h x86: Use new cache mode type in memtype related functions 2014-11-16 11:04:26 +01:00
pat_rbtree.c x86: Use new cache mode type in memtype related functions 2014-11-16 11:04:26 +01:00
pat.c x86/mm/pat: Ensure different messages in STRICT_DEVMEM and PAT cases 2015-02-19 02:09:49 +01:00
pf_in.c x86: Eliminate various 'set but not used' warnings 2011-05-21 19:10:33 +02:00
pf_in.h x86 mmiotrace: move files into arch/x86/mm/. 2008-05-24 11:25:37 +02:00
pgtable_32.c x86: Remove set_pmd_pfn 2014-09-01 10:15:31 +02:00
pgtable.c x86, mm: support huge KVA mappings on x86 2015-04-14 16:49:04 -07:00
physaddr.c x86, mm: Make DEBUG_VIRTUAL work earlier in boot 2013-01-25 16:33:22 -08:00
physaddr.h x86: split __phys_addr out into separate file 2009-09-10 11:48:55 -07:00
setup_nx.c x86: delete __cpuinit usage from all x86 files 2013-07-14 19:36:56 -04:00
srat.c x86/mm: Avoid duplicated pxm_to_node() calls 2014-02-09 15:32:31 +01:00
testmmiotrace.c x86, kmmio/mmiotrace: Fix double free of kmmio_fault_pages 2010-06-18 11:30:09 +02:00
tlb.c x86: Store a per-cpu shadow copy of CR4 2015-02-04 12:10:42 +01:00