Message-ID: <4E8AF633.3030802@tonian.com>
Date: Tue, 04 Oct 2011 14:04:03 +0200
From: Benny Halevy <bhalevy@tonian.com>
To: Boaz Harrosh <bharrosh@panasas.com>
CC: Trond Myklebust <Trond.Myklebust@netapp.com>,
        NFS list <linux-nfs@vger.kernel.org>, open-osd <osd-dev@open-osd.org>,
        "Welch, Brent" <welch@panasas.com>
Subject: Re: [PATCHSET 00/19] objlayout: Move to ORE
References: <4E8ADEDA.4050709@panasas.com>
In-Reply-To: <4E8ADEDA.4050709@panasas.com>
Content-Type: text/plain; charset=ISO-8859-1
Sender: linux-nfs-owner@vger.kernel.org
MIME-Version: 1.0

Boaz, please change my email address in your contacts lists
to bhalevy@tonian.com, thanks...

On 2011-10-04 12:24, Boaz Harrosh wrote:
> 
> Submitted is the move of the objects-layout-driver to the ORE
> (Objects Raid Engine). Which after this patchset will be
> used by both the exofs file system and the objlayoutdriver.
> (ore.ko is its own library since last Kernel)
> 
> This code is intended for the 3.2 Kernel and is already
> collecting dust in linux-next. (Though the latest bits from today)
> 
> End of this week I will post the RAID5 support to ORE and both
> exofs and objlayoutdriver. Also meant for 3.2
> 
> Trond Hi
>   The First 12 patches are to the ore and exofs to make them
>   compatible with objlayoutdriver. The pnfs-obj patches are
>   dependent on the first part been present.
>   We can either submit them all through your tree. Or alternatively
>   You can send your ACK-by: on the last 7 and I can submit them
>   to Linus through my tree. Which ever you prefer.
> 
> Benny Hi
>   I please need your Review-by: to [PATCH 13/19] and [PATCH 14/19]
>   which change code behaviour. The rest of the patches are just conversions
>   which in theory do not add or change any functionality. (Unless there
>   are bugs, but that's not intended)

Sure

> 
>   Also, With these patches panfs-layout-driver is totally broken.
>   Please just remove it once you rebase on these patches. With the RAID5
>   support it is no longer needed. I've started testing with PanFS export
>   through the STD objlayoutdriver, hope to finish this week. (We always
>   have the old versions)

No problem, less is more :)

Benny

> 
>   Also tomorrow I will send the needed patch for pnfsd-exofs branch
>   that works with these changes.
> 
> These are the list off patches:
> 
> [PATCH 01/19] exofs: Rename struct ore_components comps => oc
> [PATCH 02/19] exofs: Remove unused data_map member from exofs_sb_info
> [PATCH 03/19] ore: Make ore_striping_info and ore_calc_stripe_info public
> [PATCH 04/19] ore/exofs: Change the type of the devices array (API change)
> [PATCH 05/19] ore: Only IO one group at a time (API change)
> [PATCH 06/19] ore: cleanup: Embed an ore_striping_info inside ore_io_state
> [PATCH 07/19] ore: Remove check for ios->kern_buff in _prepare_for_striping to later
> [PATCH 08/19] exofs: Support for short read/writes
> [PATCH 09/19] ore: Support for short read/writes
> [PATCH 10/19] ore: Support for partial component table
> [PATCH 11/19] ore/exofs: Define new ore_verify_layout
> [PATCH 12/19] ore/exofs: Change ore_check_io API
> 
> 	Up to here are the changes need to ore and exofs so the ore
> 	can be used by the objlayoutdriver. Any review is welcome.
> 	Same API will be used for RAID4/5/6 support.
> 
> [PATCH 13/19] pnfs-obj: Remove redundant EOF from objlayout_io_state
> [PATCH 14/19] pnfs-obj: Return PNFS_NOT_ATTEMPTED in case of read/write_pagelist
> 
> 	Benny please review these two. They are independent of the ORE
> 	conversion. I think the [PATCH 14/19] might not be enough and
> 	Error handling needs "more", but the needed changes are their
> 	own patch, to come later.
> 
> [PATCH 15/19] pnfs-obj: Get rid of objlayout_{alloc,free}_io_state
> [PATCH 16/19] pnfs-obj: Rename objlayout_io_state => objlayout_io_res
> [PATCH 17/19] pnfs-obj: move to ore 01: ore_layout & ore_components
> [PATCH 18/19] pnfs-obj: move to ore 02: move to ORE
> [PATCH 19/19] pnfs-obj: move to ore 03: Remove old raid engine
> 
> 	These 5 stage the move to the ore. With these patches I'm
> 	able to pass all the tests I passed with the old code.
> 	Only now with more then 500 lines of code less.
> 
> Cheers
> Boaz
> --
> To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

-- 
Benny Halevy
CTO, Tonian Inc.

Tel: +972-54-802-8340
bhalevy@tonian.com