reserve nocache based on RAM size
A message that is not in the archive (sent off-list or not yet received): [email protected]
Quoted 8 lines
From: Tomas Szepe <[email protected]>
Newsgroups: rhat.general.linux-kernel
Date: Sun, 14 Jul 2002 17:38:05 +0200Since there's no official sparc32 maintainer, I'm sending this patch directly to you. It has now been tested in various configurations (released in the default Aurora 0.3 kernel) and appears to be causing no undesired side effects.
Would you mind to send me 3-4 /proc/meminfos and /proc/cpuinfos from your Aurora boxes with this patch, preferably after some uptime?
Also, did you think about a deadlock-free runtime resizing of the nocache memory? I did not even bother with boot-time resizing, because run-time resizing sounds doable and certainly nobler.
-- Pete
Quoted 11 lines
Quoted 8 lines
From: Tomas Szepe <[email protected]>
Newsgroups: rhat.general.linux-kernel
Date: Sun, 14 Jul 2002 17:38:05 +0200Since there's no official sparc32 maintainer, I'm sending this patch directly to you. It has now been tested in various configurations (released in the default Aurora 0.3 kernel) and appears to be causing no undesired side effects.
Would you mind to send me 3-4
/proc/meminfosand/proc/cpuinfosfrom your Aurora boxes with this patch, preferably after some uptime?
At the moment, I've only got one box up:
$ uname -r
2.4.19-pre10
$ uptime
8:39pm up 41 days, 10:33, 5 users, load average: 0.00, 0.00, 0.00
$ cat /proc/meminfo
total: used: free: shared: buffers: cached:
Mem: 157298688 145948672 11350016 0 41795584 48705536
Swap: 234864640 24662016 210202624
MemTotal: 153612 kB
MemFree: 11084 kB
MemShared: 0 kB
Buffers: 40816 kB
Cached: 42640 kB
SwapCached: 4924 kB
Active: 40752 kB
Inactive: 67556 kB
HighTotal: 64828 kB
HighFree: 4472 kB
LowTotal: 88784 kB
LowFree: 6612 kB
SwapTotal: 229360 kB
SwapFree: 205276 kB
$ cat /proc/cpuinfo
cpu : Texas Instruments, Inc. - SuperSparc-(II)
fpu : SuperSparc on-chip FPU
promlib : Version 3 Revision 2
prom : 2.22
type : sun4m
ncpus probed : 2
ncpus active : 2
Cpu0Bogo : 74.75
Cpu1Bogo : 59.80
MMU type : TI Viking/MXCC
contexts : 65536
nocache total : 3145728
nocache used : 998656
CPU0 : online
CPU1 : onlineQuoted 3 lines
Also, did you think about a deadlock-free runtime resizing of the nocache memory? I did not even bother with boot-time resizing, because run-time resizing sounds doable and certainly nobler.
Not yet, sounds like a good idea, though. I'll certainly have a look at it later on.
T.