Messages in this thread | | | Date | Fri, 12 Nov 1999 11:20:32 -0500 (EST) | From | William Stearns <> | Subject | 2.3.27 always keeping 15M free? |
| |
Good day, all, I've got a P90 laptop with 40M *1. Somewhere between 2.3.24 and 2.3.27-final the kernel started keeping 15M or more free all the time. On a 40M laptop that makes a big difference. *smile* Here's /proc/meminfo: total: used: free: shared: buffers: cached: Mem: 39653376 23949312 15704064 0 118784 12414976 Swap: 132083712 24813568 107270144 MemTotal: 38724 kB MemFree: 15336 kB MemShared: 0 kB Buffers: 116 kB Cached: 12124 kB HighTotal: 0 kB HighFree: 0 kB LowTotal: 38724 kB LowFree: 15336 kB SwapTotal: 128988 kB SwapFree: 104756 kB
Is there any chance that the kernel is not allocating DMA'able memory for some reason? I've tried "echo 20 39 50 >/proc/sys/vm/freepages", but the system still keeps 15-16M free even if I try to open new programs. Echo'ing "1" into overcommit_memory (which was 0) didn't help either. BTW - shouldn't the default be 1 so the system will overcommit memory? Just wondering.
Here are the memory related boot messages: totalpages: 00002830 Memory: 38668k/41152k available (0k kernel code, 0k reserved, 0k data, 0k init, 0k highmem) Buffer-cache hash table entries: 1024 (order: 0, 4096 bytes) Page-cache hash table entries: 16384 (order: 4, 65536 bytes) Freeing unused kernel memory: 56k freed Adding Swap: 128988k swap-space (priority -1)
The entries from /proc/sys/vm: [root@sparrow vm]# cat bdflush 40 500 64 256 500 3000 500 1884 2 [root@sparrow vm]# cat buffermem 2 10 60 [root@sparrow vm]# cat freepages 20 39 50 [root@sparrow vm]# cat kswapd 512 32 32 [root@sparrow vm]# cat overcommit_memory 1 [root@sparrow vm]# cat page-cluster 4 [root@sparrow vm]# cat pagecache 2 15 75 [root@sparrow vm]# cat pagetable_cache 25 50
The (hopefully) relevant portions of the kernel config: CONFIG_M586TSC=y CONFIG_X86_INVLPG=y CONFIG_NOHIGHMEM=y # CONFIG_HIGHMEM4G is not set # CONFIG_HIGHMEM64G is not set # CONFIG_ACPI is not set
What kinds of things could I do to help track this down? I'll start off by going back to an older kernel and figure out where the problem started. With a 2.5 hour compile, though, it may take a while. *smile* Cheers, - Bill
*1 I'd upgrade it if I could, but that's all it'll take.
--------------------------------------------------------------------------- Failure is not an option. It comes bundled with your Microsoft product. -- Ferenc Mantfeld (Courtesy of Jim Thompson <jim@wayport.net>) -------------------------------------------------------------------------- William Stearns (wstearns@pobox.com). Mason, Buildkernel, named2hosts, and ipfwadm2ipchains are at: http://www.pobox.com/~wstearns/ --------------------------------------------------------------------------
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |