You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

Wolfgang Wander 1363c3cd86 [PATCH] Avoiding mmap fragmentation ... Ingo recently introduced a great speedup for allocating new mmaps using the free_area_cache pointer which boosts the specweb SSL benchmark by 4-5% and causes huge performance increases in thread creation. The downside of this patch is that it does lead to fragmentation in the mmap-ed areas (visible via /proc/self/maps), such that some applications that work fine under 2.4 kernels quickly run out of memory on any 2.6 kernel. The problem is twofold: 1) the free_area_cache is used to continue a search for memory where the last search ended. Before the change new areas were always searched from the base address on. So now new small areas are cluttering holes of all sizes throughout the whole mmap-able region whereas before small holes tended to close holes near the base leaving holes far from the base large and available for larger requests. 2) the free_area_cache also is set to the location of the last munmap-ed area so in scenarios where we allocate e.g. five regions of 1K each, then free regions 4 2 3 in this order the next request for 1K will be placed in the position of the old region 3, whereas before we appended it to the still active region 1, placing it at the location of the old region 2. Before we had 1 free region of 2K, now we only get two free regions of 1K -> fragmentation. The patch addresses thes issues by introducing yet another cache descriptor cached_hole_size that contains the largest known hole size below the current free_area_cache. If a new request comes in the size is compared against the cached_hole_size and if the request can be filled with a hole below free_area_cache the search is started from the base instead. The results look promising: Whereas 2.6.12-rc4 fragments quickly and my (earlier posted) leakme.c test program terminates after 50000+ iterations with 96 distinct and fragmented maps in /proc/self/maps it performs nicely (as expected) with thread creation, Ingo's test_str02 with 20000 threads requires 0.7s system time. Taking out Ingo's patch (un-patch available per request) by basically deleting all mentions of free_area_cache from the kernel and starting the search for new memory always at the respective bases we observe: leakme terminates successfully with 11 distinctive hardly fragmented areas in /proc/self/maps but thread creating is gringdingly slow: 30+s(!) system time for Ingo's test_str02 with 20000 threads. Now - drumroll ;-) the appended patch works fine with leakme: it ends with only 7 distinct areas in /proc/self/maps and also thread creation seems sufficiently fast with 0.71s for 20000 threads. Signed-off-by: Wolfgang Wander <wwc@rentec.com> Credit-to: "Richard Purdie" <rpurdie@rpsys.net> Signed-off-by: Ken Chen <kenneth.w.chen@intel.com> Acked-by: Ingo Molnar <mingo@elte.hu> (partly) Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>		20 years ago
..
acpi	Linux-2.6.12-rc2	20 years ago
cpufreq	Linux-2.6.12-rc2	20 years ago
Makefile	[PATCH] x86_64: Add pmtimer support	20 years ago
aperture.c	[PATCH] revert x86_64-use-the-e820-hole-to-map-the-iommu-agp-aperture	20 years ago
apic.c	[PATCH] x86_64: Collected NMI watchdog fixes.	20 years ago
asm-offsets.c	Linux-2.6.12-rc2	20 years ago
e820.c	[PATCH] x86-64: Handle empty E820 regions correctly	20 years ago
early_printk.c	[PATCH] make lots of things static	20 years ago
entry.S	[PATCH] x86_64: Readd missing tests in entry.S	20 years ago
genapic.c	[PATCH] x86_64 genapic update	20 years ago
genapic_cluster.c	Linux-2.6.12-rc2	20 years ago
genapic_flat.c	Linux-2.6.12-rc2	20 years ago
head.S	[PATCH] x86_64: Dump stack and prevent recursion on early fault	20 years ago
head64.c	[PATCH] x86_64: saved_command_line overflow fix	20 years ago
i387.c	Linux-2.6.12-rc2	20 years ago
i8259.c	[PATCH] Fix u32 vs. pm_message_t in x86-64	20 years ago
init_task.c	Linux-2.6.12-rc2	20 years ago
io_apic.c	[PATCH] x86_64: More fixes for compilation without CONFIG_ACPI	20 years ago
ioport.c	Linux-2.6.12-rc2	20 years ago
irq.c	Linux-2.6.12-rc2	20 years ago
kprobes.c	[PATCH] Kprobes: Incorrect handling of probes on ret/lret instruction	20 years ago
ldt.c	Linux-2.6.12-rc2	20 years ago
mce.c	[PATCH] x86_64: Remove excessive stack allocation in MCE code with large NR_CPUS	20 years ago
mce_intel.c	Linux-2.6.12-rc2	20 years ago
module.c	[PATCH] uml: kludgy compilation fixes for x86-64 subarch modules support	20 years ago
mpparse.c	[PATCH] x86_64: More fixes for compilation without CONFIG_ACPI	20 years ago
msr.c	Linux-2.6.12-rc2	20 years ago
nmi.c	[PATCH] x86_64: Collected NMI watchdog fixes.	20 years ago
pci-dma.c	Linux-2.6.12-rc2	20 years ago
pci-gart.c	Linux-2.6.12-rc2	20 years ago
pci-nommu.c	Linux-2.6.12-rc2	20 years ago
pmtimer.c	[PATCH] x86_64: Add pmtimer support	20 years ago
process.c	[PATCH] x86_64: TASK_SIZE fixes for compatibility mode processes	20 years ago
ptrace.c	[PATCH] x86_64: TASK_SIZE fixes for compatibility mode processes	20 years ago
reboot.c	Linux-2.6.12-rc2	20 years ago
semaphore.c	Linux-2.6.12-rc2	20 years ago
setup.c	[PATCH] x86_64: 386/x86-64 Further AMD dual core fixes	20 years ago
setup64.c	[PATCH] Increase number of e820 entries hard limit from 32 to 128	20 years ago
signal.c	[PATCH] x86_64: signal.c build fix	20 years ago
smp.c	[PATCH] x86_64: Switch SMP bootup over to new CPU hotplug state machine	20 years ago
smpboot.c	[PATCH] x86_64: i386/x86-64: Export cpu_core_map	20 years ago
suspend.c	Linux-2.6.12-rc2	20 years ago
suspend_asm.S	Linux-2.6.12-rc2	20 years ago
sys_x86_64.c	[PATCH] Avoiding mmap fragmentation	20 years ago
syscall.c	Linux-2.6.12-rc2	20 years ago
time.c	[PATCH] x86_64: More fixes for compilation without CONFIG_ACPI	20 years ago
trampoline.S	Linux-2.6.12-rc2	20 years ago
traps.c	[PATCH] x86_64: CONFIG_BUG=n fixes	20 years ago
vmlinux.lds.S	Linux-2.6.12-rc2	20 years ago
vsyscall.c	[PATCH] x86_64: Make vsyscall.c compile without CONFIG_SYSCTL	20 years ago
x8664_ksyms.c	[PATCH] x86_64: CONFIG_BUG=n fixes	20 years ago