linux/arch
Dave Hansen e9f4e0a9fe x86/mm: Rip out complicated, out-of-date, buggy TLB flushing
I think the flush_tlb_mm_range() code that tries to tune the
flush sizes based on the CPU needs to get ripped out for
several reasons:

1. It is obviously buggy.  It uses mm->total_vm to judge the
   task's footprint in the TLB.  It should certainly be using
   some measure of RSS, *NOT* ->total_vm since only resident
   memory can populate the TLB.
2. Haswell, and several other CPUs are missing from the
   intel_tlb_flushall_shift_set() function.  Thus, it has been
   demonstrated to bitrot quickly in practice.
3. It is plain wrong in my vm:
	[    0.037444] Last level iTLB entries: 4KB 0, 2MB 0, 4MB 0
	[    0.037444] Last level dTLB entries: 4KB 0, 2MB 0, 4MB 0
	[    0.037444] tlb_flushall_shift: 6
   Which leads to it to never use invlpg.
4. The assumptions about TLB refill costs are wrong:
	http://lkml.kernel.org/r/1337782555-8088-3-git-send-email-alex.shi@intel.com
    (more on this in later patches)
5. I can not reproduce the original data: https://lkml.org/lkml/2012/5/17/59
   I believe the sample times were too short.  Running the
   benchmark in a loop yields times that vary quite a bit.

Note that this leaves us with a static ceiling of 1 page.  This
is a conservative, dumb setting, and will be revised in a later
patch.

This also removes the code which attempts to predict whether we
are flushing data or instructions.  We expect instruction flushes
to be relatively rare and not worth tuning for explicitly.

Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Link: http://lkml.kernel.org/r/20140731154055.ABC88E89@viggo.jf.intel.com
Acked-by: Rik van Riel <riel@redhat.com>
Acked-by: Mel Gorman <mgorman@suse.de>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
2014-07-31 08:48:50 -07:00
..
alpha Merge branch 'sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip into next 2014-06-03 14:00:15 -07:00
arc arc: call find_vma with the mmap_sem held 2014-06-04 16:54:00 -07:00
arm The clock framework changes for 3.16 are pretty typical: mostly clock 2014-06-07 20:27:30 -07:00
arm64 arm64: add APM X-Gene SoC RTC DTS entry 2014-06-06 16:08:06 -07:00
avr32 Merge branch 'locking-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip into next 2014-06-03 12:57:53 -07:00
blackfin blackfin/ptrace: call find_vma with the mmap_sem held 2014-06-06 16:08:16 -07:00
c6x DeviceTree for 3.16: 2014-06-04 10:02:38 -07:00
cris cris: update comments for generic idle conversion 2014-06-06 16:08:18 -07:00
frv sys_sgetmask/sys_ssetmask: add CONFIG_SGETMASK_SYSCALL 2014-06-04 16:54:14 -07:00
hexagon Hexagon: Delete stale barrier.h 2014-05-01 10:09:47 -07:00
ia64 ia64: convert use of typedef ctl_table to struct ctl_table 2014-06-06 16:08:15 -07:00
m32r arch,m32r: Convert smp_mb__*() 2014-04-18 14:20:36 +02:00
m68k signals: kill sigfindinword() 2014-06-06 16:08:11 -07:00
metag hugetlb: restrict hugepage_migration_support() to x86_64 2014-06-04 16:53:51 -07:00
microblaze Microblaze patches for 3.16-rc1 2014-06-05 16:15:33 -07:00
mips arch/mips/dec: switch DECstation systems to rtc-cmos 2014-06-06 16:08:07 -07:00
mn10300 sys_sgetmask/sys_ssetmask: add CONFIG_SGETMASK_SYSCALL 2014-06-04 16:54:14 -07:00
openrisc DeviceTree for 3.16: 2014-06-04 10:02:38 -07:00
parisc - Optimised assembly string/memory routines (based on the AArch64 Cortex 2014-06-06 10:43:28 -07:00
powerpc powerpc: update comments for generic idle conversion 2014-06-06 16:08:18 -07:00
s390 Merge branch 'akpm' (patchbomb from Andrew) into next 2014-06-04 16:55:13 -07:00
score arch,score: Convert smp_mb__*() 2014-04-18 14:20:43 +02:00
sh - Optimised assembly string/memory routines (based on the AArch64 Cortex 2014-06-06 10:43:28 -07:00
sparc sys_sgetmask/sys_ssetmask: add CONFIG_SGETMASK_SYSCALL 2014-06-04 16:54:14 -07:00
tile tile: convert use of typedef ctl_table to struct ctl_table 2014-06-06 16:08:15 -07:00
um mm: split 'tlb_flush_mmu()' into tlb flushing and memory freeing parts 2014-04-25 16:05:40 -07:00
unicore32 arch/unicore32/mm/ioremap.c: return NULL on invalid pfn 2014-06-04 16:53:53 -07:00
x86 x86/mm: Rip out complicated, out-of-date, buggy TLB flushing 2014-07-31 08:48:50 -07:00
xtensa - Optimised assembly string/memory routines (based on the AArch64 Cortex 2014-06-06 10:43:28 -07:00
.gitignore
Kconfig uprobes: Kconfig dependency fix 2014-03-18 16:39:33 -04:00