On 23.11.22 г. 20:53 ч., Konstantin Khorenko wrote:
Currently it's possible to mark a task as a victim even in case it has
already cleared its ->mm.
This might lead (and leads) to a situation when oom_unlock() believes
the OOM context will be released by the "victim" do_exit() ->
exit_oom_vict
On 30.11.22 г. 13:22 ч., Pavel Tikhomirov wrote:
Add a sysctl kernel.freeze_cgroup_timeout (default value 30 * HZ).
If one writes FROZEN to freezer.state file and after a timeout of
kernel.freeze_cgroup_timeout one still reads FREEZING from freezer.state
file (meaning that kernel does not succ
On 7.12.22 г. 14:27 ч., nb wrote:
On 30.11.22 г. 13:22 ч., Pavel Tikhomirov wrote:
Add a sysctl kernel.freeze_cgroup_timeout (default value 30 * HZ).
If one writes FROZEN to freezer.state file and after a timeout of
kernel.freeze_cgroup_timeout one still reads FREEZING from freezer.state
On 7.12.22 г. 20:02 ч., Konstantin Khorenko wrote:
BTW, do we need a similar patch for vz7 as well?
Looking at the vz7 kernel code I'd say yes since processes can still be
woken up on a different CPU than the one they went to sleep on. However,
I wonder why this issue hasn't manifested so f
On 19.01.23 г. 17:02 ч., Pavel Tikhomirov wrote:
I believe this is not covering all cases, for instance it would break
adding rules to second level nested cgroup, if second level nested
cgroup is in "default deny" and it's parent is in "default deny" and
none of them have CGRP_VE_ROOT set. In
On 20.01.23 г. 14:39 ч., Alexander Atanasov wrote:
From: Andrey Ryabinin
Forward port feature: mm: per memory cgroup page cache limit.
The original implementation consisted of these commits:
commit 758d52e33a67 ("configs: Enable CONFIG_PAGE_EXTENSION")
commit 741beaa93c89 ("mm: introduce pag
On 31.01.23 г. 19:08 ч., Alexander Atanasov wrote:
On 31.01.23 18:27, Alexander Atanasov wrote:
Hi,
On 31.01.23 14:23, Nikolay Borisov wrote:
Replace all shift rights/shift lefts by divide and multiply operations.
What's wrong with shifts?
I've checked that this doesn't generate worse a
On 31.01.23 г. 18:27 ч., Alexander Atanasov wrote:
Hi,
On 31.01.23 14:23, Nikolay Borisov wrote:
Replace all shift rights/shift lefts by divide and multiply operations.
What's wrong with shifts?
I've checked that this doesn't generate worse assembly and that shifts
are indeed continued t
On 21.02.23 г. 20:19 ч., Konstantin Khorenko wrote:
On 27.01.2023 12:34, Nikolay Borisov wrote:
Instead of duplicating CAP_SYS_ADMIN check for every block-cbt ioctl
apart from BLKCBTGET simply factor out the check in the beginning of
blk_cbt_ioctl.
Signed-off-by: Nikolay Borisov
---
block/
On 7.03.23 г. 5:23 ч., Pavel Tikhomirov wrote:
After this revert user_struct->pipe_bufs (and obviousely other fields =)
) becomes shared between root user of container and root user on host,
which is likely not what we want, as one ct can make all node starve for
pipe buffers now, see PIPE_MI
On 9.03.23 г. 12:37 ч., Nikolay Borisov wrote:
When preforming setuid the general expectation is that the underlying
user_struct will also be changed. However, in our kernels this doesn't
hold true to the presence of ff716deacf0c ("userns: associate user_struct
with the user_namespace").
Runni
On 20.03.23 г. 11:20 ч., Alexander Atanasov wrote:
Currently there is no way to hint the kernel to avoid triggering
page reclaims. This can be useful in networked file systems,
which can deadlock in the synchronous reclaim path and streaming
which can get unexpected jitter when a synchronouse r
On 20.03.23 г. 11:20 ч., Alexander Atanasov wrote:
Currently there is no way to control if page reclaim can be done.
This can be useful in networked file systems, which can deadlock
in the synchronous reclaim path and streaming which can get
unexpected jitter when a synchronouse reclaim is done
13 matches
Mail list logo
