linux-raid.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH v1 0/8] raid6: support read-modify-write
@ 2014-08-10 11:56 stockhausen
  2014-08-19  7:16 ` NeilBrown
  0 siblings, 1 reply; 3+ messages in thread
From: stockhausen @ 2014-08-10 11:56 UTC (permalink / raw)
  To: linux-raid

[-- Attachment #1: Type: text/plain, Size: 1479 bytes --]

Hello all.

First of all thanks to an older patch from Kumar Sundararajan and 
Dan Williams that helped me to understand RAID6 logic inside md 
better. Everything is based on ideas & discussions that started
with http://marc.info/?l=linux-raid&m=136624783417452&w=1

So once again. Implement RMW support for RAID6. This time improve
syndrome calculation too. The logic was split into very small  
parts. Each having a detailed explanation about what is going on.
A few things to note:

1) Patches are based on official 3.16 kernel git.

2) The required optimized syndrome functions were implemented if
possible. Generic & SSE2 are the ones that I could write & test
on my machine. If you want to test/benchmark this patch ensure
that you force select one of the two. Programmers with appropriate
hardware in their hands are encouraged to send the missing 
algorithms.

3) raid6 test program was enhanced to verify algorithm correctness. 
This first release is based on single threaded tests writing blocks 
with random data and size directly onto the /dev/mdX device using 
dd.  Afterwards a md check run verified consistency of the array.
So we are in need for experienced stress testers.

4) With only a VM on my laptop available I have no real benchmark
numbers. They should be on par and sometimes even higher (or less 
CPU consuming) than the old patch because of the syndrome redesign.
Maybe testers can contribute some numbers.

Any kind of feedback appreciated.

Markus


[-- Attachment #2: InterScan_Disclaimer.txt --]
[-- Type: text/plain, Size: 1650 bytes --]

****************************************************************************
Diese E-Mail enthält vertrauliche und/oder rechtlich geschützte
Informationen. Wenn Sie nicht der richtige Adressat sind oder diese E-Mail
irrtümlich erhalten haben, informieren Sie bitte sofort den Absender und
vernichten Sie diese Mail. Das unerlaubte Kopieren sowie die unbefugte
Weitergabe dieser Mail ist nicht gestattet.

Über das Internet versandte E-Mails können unter fremden Namen erstellt oder
manipuliert werden. Deshalb ist diese als E-Mail verschickte Nachricht keine
rechtsverbindliche Willenserklärung.

Collogia
Unternehmensberatung AG
Ubierring 11
D-50678 Köln

Vorstand:
Kadir Akin
Dr. Michael Höhnerbach

Vorsitzender des Aufsichtsrates:
Hans Kristian Langva

Registergericht: Amtsgericht Köln
Registernummer: HRB 52 497

This e-mail may contain confidential and/or privileged information. If you
are not the intended recipient (or have received this e-mail in error)
please notify the sender immediately and destroy this e-mail. Any
unauthorized copying, disclosure or distribution of the material in this
e-mail is strictly forbidden.

e-mails sent over the internet may have been written under a wrong name or
been manipulated. That is why this message sent as an e-mail is not a
legally binding declaration of intention.

Collogia
Unternehmensberatung AG
Ubierring 11
D-50678 Köln

executive board:
Kadir Akin
Dr. Michael Höhnerbach

President of the supervisory board:
Hans Kristian Langva

Registry office: district court Cologne
Register number: HRB 52 497

****************************************************************************

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH v1 0/8] raid6: support read-modify-write
  2014-08-10 11:56 [PATCH v1 0/8] raid6: support read-modify-write stockhausen
@ 2014-08-19  7:16 ` NeilBrown
  2014-08-19 16:40   ` AW: " Markus Stockhausen
  0 siblings, 1 reply; 3+ messages in thread
From: NeilBrown @ 2014-08-19  7:16 UTC (permalink / raw)
  To: stockhausen; +Cc: linux-raid

[-- Attachment #1: Type: text/plain, Size: 2618 bytes --]

On Sun, 10 Aug 2014 11:56:23 +0000 <stockhausen@collogia.de> wrote:

> Hello all.
> 
> First of all thanks to an older patch from Kumar Sundararajan and 
> Dan Williams that helped me to understand RAID6 logic inside md 
> better. Everything is based on ideas & discussions that started
> with http://marc.info/?l=linux-raid&m=136624783417452&w=1
> 
> So once again. Implement RMW support for RAID6. This time improve
> syndrome calculation too. The logic was split into very small  
> parts. Each having a detailed explanation about what is going on.

Thanks a lot for these patches.  The size you split them in is good.
The order is ... not quite right.
It is important for each patch to make sense by itself and for everything to
work sensibly at every point.
Yet your first patch allows PARITY_PREFER_RMW to be set for RAID6 even though
RAID6 doesn't support it yet ... that doesn't make sense.

Also I don't like that fact that you added functions with empty definitions,
even if you later filled out some definitions.
I would prefer to not have the 'hasxor' field, but to expect 'xor_syndrome'
to be NULL for some methods.
Then the RAID6 code would only try RMW is xor_syndrome was non-NULL.
Then you just add a complete xor_syndrome function and it becomes usable.

So I'd probably have that patches:
 1/ add declarations for xor_syndrome and the code to
    report benchmarks on boo, but no actuall xor_syndrome code yet.
 2/ improve test program
 3/ xor_syndrome for generic_int
 4/ add logic for improved rmw_syndrome.
    This always uses rmw for RAID6 if xor_syndrome is defined
 5/ xor_syndrome for SSE2
 6/ config option

Maybe swap the order for 4 and 5.

I want the config option to be last because I don't want it to go upstream,
but everything before it should be able to.

The other thing that is missing from this are performance numbers.
Any numbers at all would be a good start.  If we can collect number from
different use case to get a good overview of the improvement (or not) this
brings, then we might have case for including it upstream.
I realise you know this and made it clear in your point 4.  I just want to
emphasis it again.  I might do some testing myself, but we really need
someone with a larger number of devices to tell us how much better this makes
things work for them.

Thanks for your effort so far.  If you can revise the patches so:
 - config option is last
 - 'hasxor' is not used
I'll put them all in my git tree to make it a little easier for people to
test, and probably do a little myself.

Thanks
NeilBrown

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 828 bytes --]

^ permalink raw reply	[flat|nested] 3+ messages in thread

* AW: [PATCH v1 0/8] raid6: support read-modify-write
  2014-08-19  7:16 ` NeilBrown
@ 2014-08-19 16:40   ` Markus Stockhausen
  0 siblings, 0 replies; 3+ messages in thread
From: Markus Stockhausen @ 2014-08-19 16:40 UTC (permalink / raw)
  To: NeilBrown; +Cc: linux-raid@vger.kernel.org

[-- Attachment #1: Type: text/plain, Size: 531 bytes --]

> Von: NeilBrown [neilb@suse.de]
> Gesendet: Dienstag, 19. August 2014 09:16
> An: Markus Stockhausen
> Cc: linux-raid@vger.kernel.org
> Betreff: Re: [PATCH v1 0/8] raid6: support read-modify-write
> 
> Thanks for your effort so far.  If you can revise the patches so:
>  - config option is last
>  - 'hasxor' is not used
> I'll put them all in my git tree to make it a little easier for people to
> test, and probably do a little myself.

Sent and hopefully up to your expectations.

> Thanks
> NeilBrown

Markus=

[-- Attachment #2: InterScan_Disclaimer.txt --]
[-- Type: text/plain, Size: 1650 bytes --]

****************************************************************************
Diese E-Mail enthält vertrauliche und/oder rechtlich geschützte
Informationen. Wenn Sie nicht der richtige Adressat sind oder diese E-Mail
irrtümlich erhalten haben, informieren Sie bitte sofort den Absender und
vernichten Sie diese Mail. Das unerlaubte Kopieren sowie die unbefugte
Weitergabe dieser Mail ist nicht gestattet.

Über das Internet versandte E-Mails können unter fremden Namen erstellt oder
manipuliert werden. Deshalb ist diese als E-Mail verschickte Nachricht keine
rechtsverbindliche Willenserklärung.

Collogia
Unternehmensberatung AG
Ubierring 11
D-50678 Köln

Vorstand:
Kadir Akin
Dr. Michael Höhnerbach

Vorsitzender des Aufsichtsrates:
Hans Kristian Langva

Registergericht: Amtsgericht Köln
Registernummer: HRB 52 497

This e-mail may contain confidential and/or privileged information. If you
are not the intended recipient (or have received this e-mail in error)
please notify the sender immediately and destroy this e-mail. Any
unauthorized copying, disclosure or distribution of the material in this
e-mail is strictly forbidden.

e-mails sent over the internet may have been written under a wrong name or
been manipulated. That is why this message sent as an e-mail is not a
legally binding declaration of intention.

Collogia
Unternehmensberatung AG
Ubierring 11
D-50678 Köln

executive board:
Kadir Akin
Dr. Michael Höhnerbach

President of the supervisory board:
Hans Kristian Langva

Registry office: district court Cologne
Register number: HRB 52 497

****************************************************************************

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2014-08-19 16:40 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-08-10 11:56 [PATCH v1 0/8] raid6: support read-modify-write stockhausen
2014-08-19  7:16 ` NeilBrown
2014-08-19 16:40   ` AW: " Markus Stockhausen

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).