[SPDK] Re: Using of Compression VBDEV

Luse, Paul E <paul.e.luse at intel.com> Mon, 13 Sep 2021 16:40:04 +0000
Newsgroups dev.linux.lists.spdk
Message-ID <CO1PR11MB48362FF09B4A6CB5EDB81CEDADD99@CO1PR11MB4836.namprd11.prod.outlook.com>
Yup yup. Rusian, let us know if you’re interested in working on these….

From: Harris, James R <james.r.harris(a)intel.com>
Date: Monday, September 13, 2021 at 9:36 AM
To: Storage Performance Development Kit <spdk(a)lists.01.org>
Subject: [SPDK] Re: Using of Compression VBDEV
Or add a new spdk_reduce_vol_flush() API.  Then the reduce library can drive the flush operations just like it does for readv/writev.

We also need support for RESET operation.

-Jim


On 9/13/21, 8:11 AM, "Luse, Paul E" <paul.e.luse(a)intel.com> wrote:

    It’s listed in a block of other non-RW bdev IO types as TODO.  I believe at the time it was just a matter of priorities.  We’d need change the reporting of supported types to match the underlying device, update the reduce lib to allow the bdev module to determine which blocks are associated with the incoming range and then just plumb the flush.

    I may be missing something (Jim feel free to jump in).  Is this something you might be interested in working on?

    Thx
    Paul

    From: Luse, Paul E <paul.e.luse(a)intel.com>
    Date: Monday, September 13, 2021 at 7:31 AM
    To: Storage Performance Development Kit <spdk(a)lists.01.org>
    Subject: [SPDK] Re: Using of Compression VBDEV
    I can’t remember off the top of my head why that’s not supported, Jim may recall.  I’ll look through the code today and jog my memory though

    Thx
    Paul

    From: Ruslan Laishev <zator(a)yandex.ru>
    Date: Monday, September 13, 2021 at 6:50 AM
    To: spdk(a)lists.01.org <spdk(a)lists.01.org>
    Subject: [SPDK] Re: Using of Compression VBDEV
    > I/O type 4 is SPDK_BDEV_IO_TYPE_FLUSH
    >
    > Looks like Compress vbdev does not support this I/O type
    >

    Hello!

    Do I'm need to check something in the SPDK (may be configuration option for build?) ? Or this is a problem inspired by the VM's?
    _______________________________________________
    SPDK mailing list -- spdk(a)lists.01.org
    To unsubscribe send an email to spdk-leave(a)lists.01.org
    _______________________________________________
    SPDK mailing list -- spdk(a)lists.01.org
    To unsubscribe send an email to spdk-leave(a)lists.01.org
    _______________________________________________
    SPDK mailing list -- spdk(a)lists.01.org
    To unsubscribe send an email to spdk-leave(a)lists.01.org

_______________________________________________
SPDK mailing list -- spdk(a)lists.01.org
To unsubscribe send an email to spdk-leave(a)lists.01.org