We have moved to a new Sailfish OS Forum. Please start new discussions there.
1 | initial version | posted 2020-05-20 16:15:24 +0200 |
After experiencing several out-of-memory killing of applications on my xa2 device, i had a quick look on the memory usage.
top did not report a clear culprit, but did report that indeed, few memory was available. Killing all apps, stopping android support and clearing all caches resulted in around 1Gb memory being available, which should be a lot higher (remember, the XA2 has 3GB RAM, whereas the JP1 only has 1, so the “idle” memory should be around 2Gb, not 1).
There is one thing which i do not think normal (highlighted) :
[root@Sailfish vm]# cat /proc/meminfo
MemTotal: 2703220 kB
MemFree: 967720 kB
MemAvailable: 987900 kB
Buffers: 2816 kB
Cached: 71652 kB
SwapCached: 3960 kB
Active: 164060 kB
Inactive: 128316 kB
Active(anon): 117184 kB
Inactive(anon): 116616 kB
Active(file): 46876 kB
Inactive(file): 11700 kB
Unevictable: 4652 kB
Mlocked: 4652 kB
SwapTotal: 524284 kB
SwapFree: 133272 kB
Dirty: 4 kB
Writeback: 0 kB
AnonPages: 218376 kB
Mapped: 91520 kB
Shmem: 13600 kB
Slab: 397612 kB
SReclaimable: 54376 kB
SUnreclaim: 343236 kB
KernelStack: 394208 kB # <-- this is incredibly high
PageTables: 13376 kB
NFS_Unstable: 0 kB
Bounce: 0 kB
WritebackTmp: 0 kB
CommitLimit: 1875892 kB
Committed_AS: 1296816 kB
VmallocTotal: 258867136 kB
VmallocUsed: 0 kB
VmallocChunk: 0 kB
CmaTotal: 163840 kB
CmaFree: 82916 kB
This value is definitely too high. After a reboot, we get :
KernelStack: 25344 kB
Which is still a lot, but more reasonable. For comparison, here is what i got on the JP1 :
KernelStack: 5960 kB
And on an amd64 desktop computer :
KernelStack: 22032 kB
It really looks like some kernel module is leaking, and consuming far too much...
After experiencing several out-of-memory killing of applications on my xa2 device, i had a quick look on the memory usage.
top did not report a clear culprit, but did report that indeed, few memory was available. Killing all apps, stopping android support and clearing all caches resulted in around 1Gb memory being available, which should be a lot higher (remember, the XA2 has 3GB RAM, whereas the JP1 only has 1, so the “idle” memory should be around 2Gb, not 1).
There is one thing which i do not think normal (highlighted) :
[root@Sailfish vm]# cat /proc/meminfo
MemTotal: 2703220 kB
MemFree: 967720 kB
MemAvailable: 987900 kB
Buffers: 2816 kB
Cached: 71652 kB
SwapCached: 3960 kB
Active: 164060 kB
Inactive: 128316 kB
Active(anon): 117184 kB
Inactive(anon): 116616 kB
Active(file): 46876 kB
Inactive(file): 11700 kB
Unevictable: 4652 kB
Mlocked: 4652 kB
SwapTotal: 524284 kB
SwapFree: 133272 kB
Dirty: 4 kB
Writeback: 0 kB
AnonPages: 218376 kB
Mapped: 91520 kB
Shmem: 13600 kB
Slab: 397612 kB
SReclaimable: 54376 kB
SUnreclaim: 343236 kB
KernelStack: 394208 kB # <-- this is incredibly high
PageTables: 13376 kB
NFS_Unstable: 0 kB
Bounce: 0 kB
WritebackTmp: 0 kB
CommitLimit: 1875892 kB
Committed_AS: 1296816 kB
VmallocTotal: 258867136 kB
VmallocUsed: 0 kB
VmallocChunk: 0 kB
CmaTotal: 163840 kB
CmaFree: 82916 kB
This value is definitely too high. After a reboot, we get :
KernelStack: 25344 kB
Which is still a lot, but more reasonable. For comparison, here is what i got on the JP1 :
KernelStack: 5960 kB
And on an amd64 desktop computer :
KernelStack: 22032 kB
It really looks like some kernel module is leaking, and consuming far too much...