[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: VM cache policy change
|
From: |
Samuel Thibault |
|
Subject: |
Re: VM cache policy change |
|
Date: |
Tue, 24 Feb 2015 01:04:19 +0100 |
|
User-agent: |
Mutt/1.5.21+34 (58baf7c9f32f) (2010-12-30) |
Samuel Thibault, le Sun 22 Feb 2015 21:24:48 +0100, a écrit :
> Justus Winter, le Sat 21 Feb 2015 16:02:00 +0100, a écrit :
> > c) chance of swapping deadlocks.
>
> I can experiment with the patch series on the buildds.
I tried to start a gcc-5 build, it gets stuck at the tar x stage, with
gnumach printing:
no more room for vm_map_find_entry in 80223e20 (kmem_map_store)
no more room for kmem_realloc in 80223e20 (kmem_map_store)
The memory state is basically this:
pagesize: 4096
size: 1824M
free: 178M
active: 392M
inactive: 1161M
wired: 92.3M
zero filled: 7098M
reactivated: 13.9M
pageins: 1845M
pageouts: 2313M
page faults: 11825814
cow faults: 1064801
memobj hit ratio: 94%
cached memobjs: 79720
cache: 1055M
swap size: 4096M
swap free: 4096M
cache obj slab bufs objs bufs total reclaimable
name flags size size /slab usage count memory memory
vm_object 0010 80 4k 50 101577 101650 8132k 0k
vm_external 0010 12 4k 254 0 0 0k 0k
small_existence_map 0010 16 4k 254 0 0 0k 0k
large_existence_map 0002 1024 8k 8 0 0 0k 0k
vm_map 0010 64 4k 63 91 126 8k 0k
vm_map_entry 0010 60 4k 63 26325 26838 1704k 24k
vm_map_kentry 0015 60 4k 63 1433 1575 100k 0k
vm_map_copy 0010 280 4k 14 0 14 4k 4k
pmap 0010 20 4k 169 91 169 4k 0k
pv_entry 0010 12 4k 254 21450 21590 340k 0k
kmem_slab 0010 32 4k 127 23 127 4k 0k
kalloc_32 0010 32 4k 127 181 254 8k 0k
kalloc_64 0010 64 4k 63 6 63 4k 0k
kalloc_128 0010 128 4k 31 3 31 4k 0k
kalloc_256 0010 256 4k 15 2 15 4k 0k
kalloc_512 0002 512 4k 8 18 24 12k 0k
kalloc_1024 0002 1024 8k 8 32 40 40k 8k
kalloc_2048 0002 2048 16k 8 40 48 96k 0k
kalloc_4096 0002 4096 32k 8 48 56 224k 0k
kalloc_8192 0002 8192 32k 4 1 4 32k 0k
kalloc_16384 0002 16384 32k 2 0 0 0k 0k
kalloc_32768 0002 32768 32k 1 0 0 0k 0k
kalloc_65536 0002 65536 64k 1 0 0 0k 0k
kalloc_131072 0002 131072 128k 1 0 0 0k 0k
vm_fault_state 0010 60 4k 63 1 63 4k 0k
vm_page 0010 52 4k 72 436 504 28k 0k
ipc_space 0010 60 4k 63 94 126 8k 0k
ipc_tree_entry 0010 32 4k 127 0 127 4k 4k
ipc_port 0010 80 4k 50 262239 262350 20988k 0k
ipc_pset 0010 20 4k 169 72 169 4k 0k
ipc_marequest 0010 16 4k 254 0 0 0k 0k
memory_object_proxy 0010 12 4k 254 0 0 0k 0k
task 0010 192 4k 21 92 105 20k 0k
i386_task_iopb 0002 8192 32k 4 2 4 32k 0k
thread 0010 332 4k 12 325 372 124k 0k
pcb 0010 240 4k 16 325 384 96k 0k
i386_fpsave_state 0000 528 8k 15 167 195 104k 0k
io_buf_ptr_inband 0010 128 4k 31 2 31 4k 0k
io_req 0002 2048 16k 8 0 0 0k 0k
mach_device 0010 48 4k 84 6 84 4k 0k
net_rcv_port 0010 292 4k 13 0 0 0k 0k
net_hash_entry 0010 32 4k 127 0 0 0k 0k
dev_pager 0010 36 4k 101 2 101 4k 0k
dev_pager_entry 0010 16 4k 254 4 254 4k 0k
total: 32148k, reclaimable: 40k
Samuel
- VM cache policy change, Justus Winter, 2015/02/21
- [PATCH gnumach 1/3] VM cache policy change, Justus Winter, 2015/02/21
- [PATCH gnumach 3/3] vm: evict clean pages first, Justus Winter, 2015/02/21
- [PATCH gnumach 2/3] vm: keep track of clean pages, Justus Winter, 2015/02/21
- Re: VM cache policy change, Samuel Thibault, 2015/02/22
- Re: VM cache policy change, Richard Braun, 2015/02/22