public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
* show_mem producing invalid output
@ 2015-07-10 17:33 Tom Lendacky
  0 siblings, 0 replies; 2+ messages in thread
From: Tom Lendacky @ 2015-07-10 17:33 UTC (permalink / raw)
  To: Vishnu Pratap Singh; +Cc: Linux Kernel Mailing List

Hi Vishnu,

Commit 49abd8c28046 ("lib/show_mem.c: add cma reserved information")
causes the following output (last four lines shown):

...
261888 pages RAM
0 pages HighMem/MovableOnly
18446744073709545533 pages reserved
8192 pages cma reserved

The very large "pages reserved" number is obviously invalid and is
the result of the subtraction in the printk where "reserved" is less
than "totalcma_pages".

This is from an arm64 system.

Thanks,
Tom

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: show_mem producing invalid output
       [not found] <47.BF.25195.F5743A55@epcpsbgx4.samsung.com>
@ 2015-07-13 13:36 ` Tom Lendacky
  0 siblings, 0 replies; 2+ messages in thread
From: Tom Lendacky @ 2015-07-13 13:36 UTC (permalink / raw)
  To: vishnu.ps; +Cc: Linux Kernel Mailing List

On 07/13/2015 12:06 AM, Vishnu Pratap Singh wrote:
> Hi Tom,

Hi Vishnu,

> 
> Can you please share the complete show_mem output, I wanted to know the reserved
> pages value.

sysrq: SysRq : Show Memory
Mem-Info:
active_anon:1427 inactive_anon:161 isolated_anon:0
 active_file:524 inactive_file:953 isolated_file:0
 unevictable:0 dirty:0 writeback:0 unstable:0
 slab_reclaimable:195 slab_unreclaimable:740
 mapped:517 shmem:178 pagetables:70 bounce:0
 free:255076 free_pcp:71 free_cma:8156
DMA free:3993792kB min:3840kB low:4800kB high:5760kB active_anon:30656kB inactive_anon:5120kB active_file:16320kB inactive_file:27968kB unevictable:0kB isolated(anon):0kB isolato
lowmem_reserve[]: 0 764 764
Normal free:12326912kB min:11776kB low:14720kB high:17664kB active_anon:64832kB inactive_anon:5184kB active_file:17216kB inactive_file:33024kB unevictable:0kB isolated(anon):0kBo
lowmem_reserve[]: 0 0 0
DMA: 16*64kB (UE) 4*128kB (U) 3*256kB (UC) 4*512kB (UEMC) 4*1024kB (MC) 3*2048kB (UM) 4*4096kB (UEMC) 3*8192kB (UEC) 2*16384kB (EC) 7*32768kB (UEMC) 6*65536kB (UEMC) 7*131072kB B
Normal: 38*64kB (UEM) 31*128kB (UEM) 28*256kB (UM) 13*512kB (UM) 4*1024kB (UEM) 3*2048kB (EM) 3*4096kB (EM) 0*8192kB 2*16384kB (UE) 2*32768kB (UE) 2*65536kB (EM) 2*131072kB (UE)B
Node 0 hugepages_total=0 hugepages_free=0 hugepages_surp=0 hugepages_size=524288kB
1655 total pagecache pages
0 pages in swap cache
Swap cache stats: add 0, delete 0, find 0/0
Free swap  = 2047936kB
Total swap = 2047936kB
261888 pages RAM
0 pages HighMem/MovableOnly
18446744073709545533 pages reserved
8192 pages cma reserved

> 
> Currently I see the 8192 pages as cma reserved. we get the pages reserved after
> the subtraction of  (reserved-totalcma_pages).

If CMA pages are supposed to be part of the reserved space, shouldn't
the call to adjust_managed_page_count in init_cma_reserved_pageblock of
mm/page_alloc.c be for a negative amount so that the number of managed
pages is decreased (thus increasing the reserved count)?

Thanks,
Tom

> 
> BR,
> 
> Vishnu
> 
> ------- *Original Message* -------
> 
> *Sender* : Tom Lendacky<thomas.lendacky@amd.com>
> 
> *Date* : Jul 10, 2015 23:03 (GMT+05:30)
> 
> *Title* : show_mem producing invalid output
> 
> Hi Vishnu,
> 
> Commit 49abd8c28046 ("lib/show_mem.c: add cma reserved information")
> causes the following output (last four lines shown):
> 
> ...
> 261888 pages RAM
> 0 pages HighMem/MovableOnly
> 18446744073709545533 pages reserved
> 8192 pages cma reserved
> 
> The very large "pages reserved" number is obviously invalid and is
> the result of the subtraction in the printk where "reserved" is less
> than "totalcma_pages".
> 
> This is from an arm64 system.
> 
> Thanks,
> Tom
> 

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2015-07-13 13:36 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2015-07-10 17:33 show_mem producing invalid output Tom Lendacky
     [not found] <47.BF.25195.F5743A55@epcpsbgx4.samsung.com>
2015-07-13 13:36 ` Tom Lendacky

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox