There are many places where it is critical to know if invalidate files
operation is in progress, otherwise deadlock would happen. Currently
this can be done by checking the FUSE_S_FAIL_IMMEDIATELY flag bit from
fuse_file state. However this is inconvenient in case where a fuse_file
is not directl
Attached the original patch file.
From: on behalf of Kui Liu
Date: Wednesday, 16 March 2022 at 8:07 PM
To: "devel@openvz.org"
Subject: [Devel] [PATCH RH9 2/2] fuse: Add a new fuse inode state
FUSE_I_INVAL_FILES
There are many places where it is critical to know if invalidate files
, while
ploop devices are backed
by files stored in vstorage filesystem.
With this background info, please see my comments below.
-Original Message-
From: Konstantin Khorenko
Date: Thursday, 13 October 2022 at 5:23 PM
To: Kui Liu , Alexey Kuznetsov
Cc: Devel , Alexander Atanasov
-Original Message-
From: Alexander Atanasov
Date: Thursday, 13 October 2022 at 11:26 PM
To: Kui Liu , Konstantin Khorenko
, Alexey Kuznetsov
Cc: Devel , Alexander Mikhalitsyn
Subject: Re: [Devel] [PATCH RH9] dm-ploop: port the standby mode feature.
On 13.10.22 16:33, Kui Liu
ementation?
Regards,
Liu Kui
-Original Message-
From: Konstantin Khorenko
Date: Tuesday, 18 October 2022 at 1:57 AM
To: Kui Liu , Alexey Kuznetsov
Cc: Devel , Alexander Mikhalitsyn
, Alexander Atanasov
Subject: Re: [Devel] [PATCH RH9] dm-ploop: port the standby mode feature.
Hi L
Hello,
Regarding the QUEUE_FLAG_STANDBY_EN flag, it is like what you added.
However I find some problems with the code, see my comments below.
-Original Message-
From: Alexander Atanasov
Date: Tuesday, 18 October 2022 at 5:41 PM
To: Kui Liu , Konstantin Khorenko
, Alexey
Hello,
-Original Message-
From: Alexander Atanasov
Date: Tuesday, 18 October 2022 at 7:40 PM
To: Kui Liu , Konstantin Khorenko
, Alexey Kuznetsov
Cc: Devel
Subject: Re: [Devel] [PATCH RH9] dm-ploop: port the standby mode feature.
On 18.10.22 13:57, Kui Liu wrote:
> He
-Original Message-
From: Alexander Atanasov
Date: Tuesday, 18 October 2022 at 9:12 PM
To: Kui Liu , Konstantin Khorenko
, Alexey Kuznetsov
Cc: Devel
Subject: Re: [Devel] [PATCH RH9] dm-ploop: port the standby mode feature.
On 18.10.22 15:47, Kui Liu wrote:
> He
-Original Message-
From: Alexander Atanasov
Date: Wednesday, 19 October 2022 at 4:02 PM
To: Kui Liu , Konstantin Khorenko
, Alexey Kuznetsov
Cc: Devel
Subject: Re: [Devel] [PATCH RH9] dm-ploop: port the standby mode feature.
On 19.10.22 9:58, Kui Liu wrote
-Original Message-
From: Alexander Atanasov
Date: Friday, 21 October 2022 at 6:08 PM
To: Kui Liu , Konstantin Khorenko
, Alexey Kuznetsov
Cc: Devel
Subject: Re: [Devel] [PATCH RH9] dm-ploop: port the standby mode feature.
On 19.10.22 11:25, Kui Liu wrote
-Original Message-
From: Alexander Atanasov
Date: Friday, 21 October 2022 at 7:35 PM
To: Kui Liu , Konstantin Khorenko
, Alexey Kuznetsov
Cc: Devel
Subject: Re: [Devel] [PATCH RH9] dm-ploop: port the standby mode feature.
On 21.10.22 14:28, Kui Liu wrote
-Original Message-
From: Alexander Atanasov
Date: Wednesday, 26 October 2022 at 5:14 PM
To: Kui Liu , Konstantin Khorenko
, Alexey Kuznetsov
Cc: Devel
Subject: Re: [Devel] [PATCH RH9] dm-ploop: port the standby mode feature.
Hi,
On 21.10.22 16:28, Kui Liu wrote
-Original Message-
From: Alexander Atanasov
Date: Wednesday, 26 October 2022 at 9:32 PM
To: Kui Liu , Konstantin Khorenko
, Alexey Kuznetsov
Cc: Devel
Subject: Re: [Devel] [PATCH RH9] dm-ploop: port the standby mode feature.
On 26.10.22 16:15, Kui Liu wrote
-Original Message-
From: Alexander Atanasov
Date: Thursday, 27 October 2022 at 4:28 PM
To: Kui Liu , Konstantin Khorenko
, Alexey Kuznetsov , "Denis V.
Lunev"
Cc: Devel
Subject: Re: [Devel] [PATCH RH9] dm-ploop: port the standby mode feature.
On 26.10.22 17:42, Kui
-Original Message-
From: Alexander Atanasov
Date: Tuesday, 1 November 2022 at 11:57 PM
To: Konstantin Khorenko , "Denis V. Lunev"
, Devel
Cc: Kui Liu , Alexey Kuznetsov
Subject: Re: [PATCH vz9 v2] ploop: port and fix the standby mode feature.
On 1.11.22 17:27,
Hello,
This patch looks good, no further questions from me.
Regards,
LIU KUI
-Original Message-
From: Alexander Atanasov
Date: Friday, 4 November 2022 at 8:49 PM
To: Devel
Cc: Konstantin Khorenko , Alexander Atanasov
, "Denis V . Lunev" , Kui
Liu , Alexey Kuznetso
:31 PM
To: Devel mailto:devel@openvz.org>>, Konstantin Khorenko
mailto:khore...@virtuozzo.com>>, Kui Liu
mailto:kui@acronis.com>>
Subject: [PATCH RH9] fuse: pcs: do not use sendpage on pages which cannot be
sendpaged
Use the same approach as in iscsi iscsi_tcp_segment_map() which m
lexey Kuznetsov mailto:kuz...@acronis.com>>
Date: Wednesday, 19 April 2023 at 12:09 AM
To: Kui Liu mailto:kui@acronis.com>>
Cc: Devel mailto:devel@openvz.org>>, Konstantin Khorenko
mailto:khore...@virtuozzo.com>>
Subject: Re: [Devel] [PATCH RH9] fuse: pcs: do not use sendp
It looks like the new API iov_iter_get_pages is not safe for use
when trying to get a page from a bvec to be sent by kernel_sendpage().
So just revert back our own implentation where we check the page
before making get_page(), if the page can't be sendpage, fall back
to copy mode.
Affects: #PSBM-1
In RoCE network, packet loss and dealy due to congestion can happen
quite often. We need to tolerate such event. So increase retry_count
and rnr_retry_count to 7 to allow NIC to retry operations when an
error happens, instead of returning the error directly which causes
the connection to be aborte
can at least have
stable connection in such RoCE network.
-Original Message-
From: Alexey Kuznetsov mailto:kuz...@acronis.com>>
Date: Thursday, 17 August 2023 at 11:52 PM
To: Kui Liu mailto:kui@acronis.com>>
Cc: Devel mailto:devel@openvz.org>>
Subject: Re: [Devel] [
Yes, I think we need it for vz9, the patch can be applied directly to vz9.
-Original Message-
From: Konstantin Khorenko mailto:khore...@virtuozzo.com>>
Date: Friday, 18 August 2023 at 1:40 AM
To: Kui Liu mailto:kui@acronis.com>>, Devel
mailto:devel@openvz.org&
rdma_connect() now can't be called directly in cm event handler,
as it will cause deadlock. Use rdma_connect_locked() instead
https://pmc.acronis.work/browse/VSTOR-73616
Signed-off-by: Liu Kui
---
fs/fuse/kio/pcs/pcs_rdma_conn.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --gi
Use correct lock to protect the kill_requests operation on all request
queues. Test of the FUSE_S_FAIL_IMMEDIATELY bit needs to be done with
corresponding lock to prevent new requests being added to a queue after
it has been killed.
Adjust the order of queues that kill_request operation is perform
Currently flush_bg_queue() takes an input paramenter fiq, to which requests
in the bg_queue will be dispatched without respect to what's been set to
req->args->fiq. This behaviour totally breaks purpose of the request dispatching
algorithm when there are multiple input queues. So instead we need to
October 2023 at 5:23 PM
To: "Denis V. Lunev" , Alexey Kuznetsov
, Devel , "khore...@virtuozzo.com"
, Kui Liu
Subject: Re: fuse: call proper fiemap_prep as it is required by new kernel api
ok for me
From: Denis V. Lunev
Sent: Sunday, October
Then ok for me.
-Original Message-
From: Alexey Kuznetsov mailto:kuz...@virtuozzo.com>>
Date: Monday, 16 October 2023 at 6:34 PM
To: Kui Liu mailto:kui@acronis.com>>
Cc: Devel mailto:devel@openvz.org>>, Andrey Zaitsev
mailto:andrey.zait...@acronis.com>>, &q
Both size of qhash and limit of each bucket can affect performance of
certain workload significantly. There is no single set of value that'd
be the best for all workload, we may need to choose a value based on
workload, so it'd be better make them configurable.
Here we choose the default value to
In our userspace RDMA implementation, it is required that every
RIO_MSG_RDMA_READ_REQ msg must be acked strictly in order. However
this rule can be broken due to a bug in kio, which though is
triggered by very abnormal hardware behaviour that it can take
very long time(>10s) for a WR to complete.
Sure, will do next time. Thanks for reminding me.
-Original Message-
From: Konstantin Khorenko mailto:khore...@virtuozzo.com>>
Date: Thursday, 23 November 2023 at 7:06 PM
To: Kui Liu mailto:kui@acronis.com>>, Devel
mailto:devel@openvz.org>>
Cc: Alexey Ku
Previously, if cm fails after the rio has been created, the rdma_cm_id
would not be destroyed immediately. However the cm_id->context could
still point to rc->id which would no longer be valid. This dealy create
a window during which cm_id->context holds an illegal pointer. If an RMDA
cm event arri
On 9 Jan 2024, at 9:40 PM, Alexey Kuznetsov wrote:
Unfortunately, existing support of splice in fuse is completely useless,
it has many flaws, each of them is fatal, even taken separately.
- it passes only single splice, which requires of user space
to prepare one more splice to merge header.
> On 18 Jan 2024, at 5:08 PM, Alexey Kuznetsov wrote:
>
>> Is it difficult to separate the verification of inputs, such as fds, size of
>> src buffs against dest buffs etc, from the process of copying data? Similar
>> to what’s been done in ‘copy_out_args()’.
>
> I am not sure I understand w
e can just use stored fds from the
first loop
and do data copy. In the out path, we can just do fdput on all stored fds.
However I’m not sure if there is any risk of holding multiple fds at the same
time, or holding
one fd longer than necessary.
> On Thu, Jan 18, 2024 a
; Andrey Zaitsev
; Konstantin Khorenko ; Kui Liu
Subject: [PATCH VZ9] fs/fuse kio: Use kvcalloc to prevent failure in fragmented
systems
Address large memory allocation failure by replacing kcalloc with
kvcalloc. This allows fallback to vmalloc for large allocations,
as virtually contiguous meomory
35 matches
Mail list logo
