On 05/12/2016 04:26 AM, Maxim Patlasov wrote:
ploop_push_backup_io_read() services ioctl(PLOOP_IOC_PUSH_BACKUP_IO)
for direction = PLOOP_READ.
If at least one extent is successfully filled, the -ENOENT error from
ploop_pb_get_pending() must be interpreted as "OK, no more data present".
https://
We calculate SUnreclaim as total kmem usage minus SReclaimable. This is
wrong, because kmem also charges non-slab objects, such as page tables
and kernel stack pages.
Signed-off-by: Vladimir Davydov
---
mm/memcontrol.c | 10 --
1 file changed, 4 insertions(+), 6 deletions(-)
diff --git
This patch adds new sysctl vm.vfs_cache_min_ratio. If the ratio of
reclaimable slabs (i.e. dcache and icache) to total memory usage of a
cgroup is less than the value of this sysctl (2% by default), slabs
won't be reclaimed from this cgroup on memory pressure.
https://jira.sw.ru/browse/PSBM-34161
Required by vcmmd for estimating the size of available memory, which in
turn is used for calculating the lower boundary for memory limit.
Also, these counters will be used by the following patches.
Signed-off-by: Vladimir Davydov
---
mm/memcontrol.c | 22 --
1 file changed,
Kostya,
On 05/12/2016 07:05 AM, Konstantin Khorenko wrote:
On 05/12/2016 04:26 AM, Maxim Patlasov wrote:
ploop_push_backup_io_read() services ioctl(PLOOP_IOC_PUSH_BACKUP_IO)
for direction = PLOOP_READ.
If at least one extent is successfully filled, the -ENOENT error from
ploop_pb_get_pending()
When more than one user keeps ploop block device opened, dying backup tool
doesn't trigger auto destroy.
The patch tries to detect dying backup tool by comparing the pid of
process starting push_backup with the pid of process closing ploop block
device.
The patch doesn't ensures auto destroy in 1