/proc/meminfo

/proc/meminfo

[Jared Hulbert]

Please have mercy on a linux-mm newbie.  I'd like to understand the
output of /proc/meminfo and /proc/<[0-9]+>/maps.  I want to measure 2
things: First, how much memory in a system is used for code or other
readonly file mmaps or what RAM can be saved by using XIP flash.
Second, at the time a system snapshot is taken how much RAM is
absolutely needed (for example, I assume we could dump caches, flush
buffers, and clean up unused memory.)

Where can I find a good reference to what this all output means?  Are
there other sources of information available?

Here are my assumptions:
# cat /proc/meminfo
MemTotal: = Memory managed by Linux kernel. Total RAM - kernel image.
MemFree: = Memory not allocated.  Not the same as memory availiable to allocate.
Buffers: = ?
Cached: = inode cache
SwapCached: = Used swap space
Active: = Pages allocated by kernel and user processes
Inactive: = Pages allocated but read to be purged
HighTotal: = 2Gig limit stuff
HighFree: =  ""
LowTotal: = ""
LowFree: = ""
SwapTotal: = What is the relationship between this and SwapCached?
SwapFree: =  ""
Dirty: = ?
Writeback: = ?
Slab: = ?
CommitLimit: = ?
Commited_AS: = ?
PageTables: = Memory allocated for use as page tables.
VmallocTotal: = Virtual memory space allocated
VmallocUsed: = ?
VmallocChunk: = ?

# cat /proc/1/maps
08048000-0804E000    r-xp    00000000    75:00    637746    /sbin/init
(readonly, executable mmap of file /sbin/init Probably code)
0804E000-0804F000    rw-p    00000000    75:00    637746    /sbin/init
(readwrite, mmap of file /sbin/init Probably initialized variables
etc)
0804F000-08070000    rw-p    0804F000    00:00    0 (I don't know)

1st column = virtual memory map of map
2nd column = r = read; w = write; x = executable; p =  I don't know
3rd column = I don't know
4th column = size of map (but it often doesn't match the size of column 1)
5th column = name of file
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>

Re: /proc/meminfo

[Dave Hansen]

On Mon, 2005-05-16 at 10:25 -0700, Jared Hulbert wrote:
> Please have mercy on a linux-mm newbie.  I'd like to understand the
> output of /proc/meminfo and /proc/<[0-9]+>/maps.  I want to measure 2
> things: First, how much memory in a system is used for code or other
> readonly file mmaps or what RAM can be saved by using XIP flash.
> Second, at the time a system snapshot is taken how much RAM is
> absolutely needed (for example, I assume we could dump caches, flush
> buffers, and clean up unused memory.)
> 
> Where can I find a good reference to what this all output means?  Are
> there other sources of information available?

Documentation/filesystems/proc.txt

-- Dave

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"aart@kvack.org"> aart@kvack.org </a>

/proc/meminfo

[Abu M. Muttalib]

Hi,

Is the information provided by /proc/meminfo reliable?

~Abu.

Re: /proc/meminfo

[Jiri Slaby]

Abu M. Muttalib napsal(a):
> Hi,
> 
> Is the information provided by /proc/meminfo reliable?
Nope, that's the reason, what's is that for... Why?
Do you have any miscues?

regards,
-- 
Jiri Slaby         www.fi.muni.cz/~xslaby
\_.-^-._   jirislaby@gmail.com   _.-^-._/
B67499670407CE62ACC8 22A032CC55C339D47A7E

/proc/meminfo

[Abu M. Muttalib]

Hi,

Does /proc/meminfo provides reliable information?

~Abu.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: /proc/meminfo

[Jes Sorensen]

>>>>> "Abu" == Abu M Muttalib <abum@aftek.com> writes:

Abu> Hi, Does /proc/meminfo provides reliable information?

Yes

Jes

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

/proc/meminfo

[Dan Am]

[-- Attachment #1: Type: text/plain, Size: 454 bytes --]

Hello,
running rhas2.1 with 2.4.18-e.31smp  (stock-redhat) on a rx2600. I noticed 
that the figures from /proc/meminfo and thus top seem to add somewhat 
different to i386:
"Shared" and "buffers" and "free" add up  to only 1.6 GB with 2GB available   
In i386 all three add up to the total available. 
I include two listings. Does that make sense and if so how ? I did search this 
lists archive for hints on "top" and "meminfo" . 

Tia & regards 

Dan 

[-- Attachment #2: meminfoi386 --]
[-- Type: text/plain, Size: 1102 bytes --]

        total:    used:    free:  shared: buffers:  cached:
Mem:  1057861632 678666240 379195392        0 198885376 366485504
Swap: 789618688        0 789618688
MemTotal:      1033068 kB
MemFree:        370308 kB
MemShared:           0 kB
Buffers:        194224 kB
Cached:         357896 kB
SwapCached:          0 kB
Active:         191296 kB
Inactive:       381736 kB
HighTotal:      130992 kB
HighFree:         2044 kB
LowTotal:       902076 kB
LowFree:        368264 kB
SwapTotal:      771112 kB
SwapFree:       771112 kB
BigFree:             0 kB
        total:    used:    free:  shared: buffers:  cached:
Mem:  1057861632 678674432 379187200        0 198893568 366485504
Swap: 789618688        0 789618688
MemTotal:      1033068 kB
MemFree:        370300 kB
MemShared:           0 kB
Buffers:        194232 kB
Cached:         357896 kB
SwapCached:          0 kB
Active:         191320 kB
Inactive:       381720 kB
HighTotal:      130992 kB
HighFree:         2044 kB
LowTotal:       902076 kB
LowFree:        368256 kB
SwapTotal:      771112 kB
SwapFree:       771112 kB
BigFree:             0 kB

[-- Attachment #3: meminfoia64 --]
[-- Type: text/plain, Size: 678 bytes --]

        total:    used:    free:  shared: buffers:  cached:
Mem:  2092859392 2026700800 66158592        0 49397760 1532116992
Swap: 2147450880        0 2147450880
MemTotal:      2043808 kB
MemFree:         64608 kB
MemShared:           0 kB
Buffers:         48240 kB
Cached:        1496208 kB
SwapCached:          0 kB
Active:         975584 kB
Inact_dirty:    749392 kB
Inact_clean:     36448 kB
Inact_target:   352272 kB
HighTotal:           0 kB
HighFree:            0 kB
LowTotal:      2043808 kB
LowFree:         64608 kB
SwapTotal:     2097120 kB
SwapFree:      2097120 kB
Committed_AS: 49680752 kB
HugePages_Total:     0
HugePages_Free:      0
Hugepagesize:    262144 kB

Re: /proc/meminfo

[William Lee Irwin III]

At some point in the past, I wrote:
>> (a) boot-time code dropping memory regions on the floor
>> (b) some misunderstanding about reporting in /proc/meminfo
>> This is also different from your other bugreport.

On Wed, Jun 04, 2003 at 06:51:09AM +0200, Roets, Chris wrote:
> [Chris >] 1.4 G is a lot to loose now ?

Your quoting style leaves something to be desired.

What exactly is this chart supposed to say? Are you asking about the
numbers not adding up?


-- wli

RE: /proc/meminfo

[Roets, Chris (Tru64&Linux support)]

[Chris >] attached jpeg shows the memory statitiscs of a 6 Gb system
[Chris >] Does anybody have an idea when my remaining 1.4 Gb can be ?

Re: /proc/meminfo

[William Lee Irwin III]

On Tue, Jun 03, 2003 at 03:51:47PM +0200, Roets, Chris wrote:
> [Chris >] attached jpeg shows the memory statitiscs of a 6 Gb system
> [Chris >] Does anybody have an idea when my remaining 1.4 Gb can be ?

What JPEG? This is probably either
(a) boot-time code dropping memory regions on the floor
(b) some misunderstanding about reporting in /proc/meminfo

This is also different from your other bugreport.


-- wli

/proc/meminfo

[Roets, Chris (Tru64&Linux support)]

if you look at /proc/meminfo, what values to you need to add to each
other
to get to the total memory ?
ok, used and free (256241664  5042176),
but then how do you come to used :

Active + Inactive_dirty + Inactive_clean + ....
in the case below, I can't find where 57360 Kb are allocated to.
I looked ad slabinfo, but cold get to the total niether
Does anybody have the calculation to get to the total ?
Thanks,
Chris

[root@tanos4 mypdc]# cat /proc/meminfo
        total:    used:    free:  shared: buffers:  cached:
Mem:  261283840 256241664  5042176    81920  9867264 177315840
Swap: 536862720   434176 536428544
MemTotal:       255160 kB
MemFree:          4924 kB
MemShared:          80 kB
Buffers:          9636 kB
Cached:         172736 kB
SwapCached:        424 kB
Active:           9196 kB
Inact_dirty:    132844 kB
Inact_clean:     40836 kB
Inact_target:      884 kB
HighTotal:           0 kB
HighFree:            0 kB
LowTotal:       255160 kB
LowFree:          4924 kB
SwapTotal:      524280 kB
SwapFree:       523856 kB
NrSwapPages:    130964 pages

Chris Roets
HP Customer support
Tru64 and Linux support
HP Belgium
www.hp.be

Chris.Roets@hp.com
Tel +32 2 729.77.44
http://www.roets.tk