[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-block] [PATCH 1/1] nbd: implement bdrv_get_info callback

From: Edgar Kaziakhmedov
Subject: Re: [Qemu-block] [PATCH 1/1] nbd: implement bdrv_get_info callback
Date: Fri, 2 Feb 2018 17:06:48 +0300
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.5.2

On 01/26/2018 05:28 PM, Eric Blake wrote:
On 01/26/2018 06:39 AM, Edgar Kaziakhmedov wrote:

So, let me know if I need to make any changes in patch

On 1/18/18 1:09 PM, Paolo Bonzini wrote:
On 18/01/2018 12:51, Edgar Kaziakhmedov wrote:
+static int nbd_get_info(BlockDriverState *bs, BlockDriverInfo *bdi)
+    if (bs->supported_zero_flags & BDRV_REQ_MAY_UNMAP) {
+        bdi->can_write_zeroes_with_unmap = true;
+    }
+    return 0;
Other drivers set the flag always, while NBD only sets it if the server
knows the flag.
Well, other drivers may be able to always implement it (NBD can only
implement it if the server supports WRITE_ZEROES - and I'm even in the
middle of working up an nbdkit patch [1] that makes it easier to write
an NBD server that specifically does not support WRITE_ZEROES to make
code paths like this easier to test)


I think NBD is more correct, so:

Reviewed-by: Paolo Bonzini <address@hidden>
Agreed; I'm fine queueing this on my NBD queue, except I'd first like to
hear Kevin's opinion:

However, it would be nice to remove can_write_zeroes_with_unmap from
BlockDriverInfo, and make bdrv_can_write_zeroes_with_unmap just return
!!(bs->supported_zero_flags & BDRV_REQ_MAY_UNMAP).  Kevin, what do you
Actually, I may even just give a shot at writing this alternative patch,
to make Kevin's decision easier.
But actually qcow2 performs some checks for version inside get_info callback before setting can_write_zeroes_with_unmap flag, so we can't take into account such checks in bdrv_can_write_zeroes_with_unmap subroutine. Therefore, I don't think it is possible to do it like that.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]