All of lore.kernel.org
 help / color / mirror / Atom feed
* [f2fs-dev] [question] ZBC/ZAC Offline/read-only zone support
@ 2020-04-14  4:23 Tim Walker
  2020-04-14 11:50 ` Chao Yu
  2020-04-14 14:09 ` Jaegeuk Kim
  0 siblings, 2 replies; 6+ messages in thread
From: Tim Walker @ 2020-04-14  4:23 UTC (permalink / raw)
  To: linux-f2fs-devel

Hello-

I am trying to determine how f2fs handles an offline (or read-only) zone when it is backed by a host-managed SMR HDD. I haven't been able to find any specific code path that places an offline zone into a separate list, or a path that checks the zone condition. Do we support a zone going read-only or offline? Thanks for the info.

Best regards,
Tim Walker

_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [f2fs-dev] [question] ZBC/ZAC Offline/read-only zone support
  2020-04-14  4:23 [f2fs-dev] [question] ZBC/ZAC Offline/read-only zone support Tim Walker
@ 2020-04-14 11:50 ` Chao Yu
  2020-04-15  2:54   ` Tim Walker
  2020-04-14 14:09 ` Jaegeuk Kim
  1 sibling, 1 reply; 6+ messages in thread
From: Chao Yu @ 2020-04-14 11:50 UTC (permalink / raw)
  To: Tim Walker, linux-f2fs-devel

Hi,

On 2020/4/14 12:23, Tim Walker wrote:
> Hello-
> 
> I am trying to determine how f2fs handles an offline (or read-only) zone when it is backed by a host-managed SMR HDD. I haven't been able to find any specific code path that places an offline zone into a separate list, or a path that checks the zone condition. Do we support a zone going read-only or offline? Thanks for the info.

So you mean different zones in SMR device can have different status? some
of them can be tagged as readonly? if my understanding is correct, f2fs
needs to handle such readonly zone correctly, e.g. write/discard IO should
be avoided being issued to those zones, right?

Thanks,

> 
> Best regards,
> Tim Walker
> 
> _______________________________________________
> Linux-f2fs-devel mailing list
> Linux-f2fs-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
> .
> 


_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [f2fs-dev] [question] ZBC/ZAC Offline/read-only zone support
  2020-04-14  4:23 [f2fs-dev] [question] ZBC/ZAC Offline/read-only zone support Tim Walker
  2020-04-14 11:50 ` Chao Yu
@ 2020-04-14 14:09 ` Jaegeuk Kim
  2020-04-15  2:49   ` Tim Walker
  1 sibling, 1 reply; 6+ messages in thread
From: Jaegeuk Kim @ 2020-04-14 14:09 UTC (permalink / raw)
  To: Tim Walker; +Cc: linux-f2fs-devel

On 04/14, Tim Walker wrote:
> Hello-
> 
> I am trying to determine how f2fs handles an offline (or read-only) zone when it is backed by a host-managed SMR HDD. I haven't been able to find any specific code path that places an offline zone into a separate list, or a path that checks the zone condition. Do we support a zone going read-only or offline? Thanks for the info.

Which scenario do we need to see ro or offline zones? Can we expect some
when mounting the filesystem? Or, does setting a zone to readonly happen
in runtime? If we detect them at mount time like former one, we need to
treat the discarded blocks in the zone with unusable space, and set
allocation bits to avoid GC.

> 
> Best regards,
> Tim Walker
> 
> _______________________________________________
> Linux-f2fs-devel mailing list
> Linux-f2fs-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel


_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [f2fs-dev] [question] ZBC/ZAC Offline/read-only zone support
  2020-04-14 14:09 ` Jaegeuk Kim
@ 2020-04-15  2:49   ` Tim Walker
  0 siblings, 0 replies; 6+ messages in thread
From: Tim Walker @ 2020-04-15  2:49 UTC (permalink / raw)
  To: Jaegeuk Kim; +Cc: linux-f2fs-devel












From: Jaegeuk Kim <jaegeuk@kernel.org>

Sent: Tuesday, April 14, 2020 8:09 AM

To: Tim Walker <timtwalker@outlook.com>

Cc: linux-f2fs-devel@lists.sourceforge.net <linux-f2fs-devel@lists.sourceforge.net>

Subject: Re: [f2fs-dev] [question] ZBC/ZAC Offline/read-only zone support

 


On 04/14, Tim Walker wrote:

> Hello-

> 

> I am trying to determine how f2fs handles an offline (or read-only) zone when it is backed by a host-managed SMR HDD. I haven't been able to find any specific code path that places an offline zone into a separate list, or a path that checks the zone condition.
 Do we support a zone going read-only or offline? Thanks for the info.



Which scenario do we need to see ro or offline zones? Can we expect some

when mounting the filesystem? Or, does setting a zone to readonly happen

in runtime? If we detect them at mount time like former one, we need to

treat the discarded blocks in the zone with unusable space, and set

allocation bits to avoid GC.



> 

> Best regards,

> Tim Walker

> 

> _______________________________________________

> Linux-f2fs-devel mailing list

> Linux-f2fs-devel@lists.sourceforge.net

Hello-

A zoned block device HDD zone can transition to Offline or Read_Only at any time, according to the spec. I don't know if any drive manufacturer transitions zones to Offline or Read_Only today. I was trying to see what support f2fs has for zones that end up in that state, since these zone states have been in the spec from the beginning. 

In my mind f2fs would discover that a zone was Offline or Read_Only when f2fs reads the zone status (write pointers) at mount time, any time f2fs resets a write pointer during GC, and any other sequence of events that causes a report zones call. In addition, if a zone were to asynchronously go Read_Only or Offline f2fs would presumably discover it when it tried to write a Read_Only zone or any access to an Offline zone. The error codes are descriptive.

If f2fs were to discover a Read_Only zone it could just mark it to never get GC, since it can't be reclaimed. Thanks.

Best regards,
-Tim Walker

_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [f2fs-dev] [question] ZBC/ZAC Offline/read-only zone support
  2020-04-14 11:50 ` Chao Yu
@ 2020-04-15  2:54   ` Tim Walker
  2020-04-15  7:15     ` Chao Yu
  0 siblings, 1 reply; 6+ messages in thread
From: Tim Walker @ 2020-04-15  2:54 UTC (permalink / raw)
  To: Chao Yu, linux-f2fs-devel


From: Chao Yu <yuchao0@huawei.com>

Sent: Tuesday, April 14, 2020 5:50 AM

To: Tim Walker <timtwalker@outlook.com>; linux-f2fs-devel@lists.sourceforge.net <linux-f2fs-devel@lists.sourceforge.net>

Subject: Re: [f2fs-dev] [question] ZBC/ZAC Offline/read-only zone support

 


Hi,



On 2020/4/14 12:23, Tim Walker wrote:

> Hello-

> 

> I am trying to determine how f2fs handles an offline (or read-only) zone when it is backed by a host-managed SMR HDD. I haven't been able to find any specific code path that places an offline zone into a separate list, or a path that checks the zone condition.
 Do we support a zone going read-only or offline? Thanks for the info.



So you mean different zones in SMR device can have different status? some

of them can be tagged as readonly? if my understanding is correct, f2fs

needs to handle such readonly zone correctly, e.g. write/discard IO should

be avoided being issued to those zones, right?



Thanks,



> 

> Best regards,

> Tim Walker

> 

> _______________________________________________

> Linux-f2fs-devel mailing list

> Linux-f2fs-devel@lists.sourceforge.net

> Hi Chao Yu-

Yes the zones can have different status, according to the spec. I don't know how many drive manufactures use those zone states - if any. I was only asking to see if I had missed Offline or Read_Only zone support when I examined the f2fs source code. I was interested in the implementation details if support had been included. Thanks!

Best regards,
-Tim Walker


_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [f2fs-dev] [question] ZBC/ZAC Offline/read-only zone support
  2020-04-15  2:54   ` Tim Walker
@ 2020-04-15  7:15     ` Chao Yu
  0 siblings, 0 replies; 6+ messages in thread
From: Chao Yu @ 2020-04-15  7:15 UTC (permalink / raw)
  To: Tim Walker, linux-f2fs-devel

Hi Tim,

On 2020/4/15 10:54, Tim Walker wrote:
>> Hi Chao Yu-
> 
> Yes the zones can have different status, according to the spec. I don't know how many drive manufactures use those zone states - if any. I was only asking to see if I had missed Offline or Read_Only zone support when I examined the f2fs source code. I was interested in the implementation details if support had been included. Thanks!

Unfortunately, that support on f2fs has not been done yet.

Thanks,

> 
> Best regards,
> -Tim Walker
> .
> 


_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2020-04-15  7:15 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-04-14  4:23 [f2fs-dev] [question] ZBC/ZAC Offline/read-only zone support Tim Walker
2020-04-14 11:50 ` Chao Yu
2020-04-15  2:54   ` Tim Walker
2020-04-15  7:15     ` Chao Yu
2020-04-14 14:09 ` Jaegeuk Kim
2020-04-15  2:49   ` Tim Walker

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.