Re: [Qemu-devel] Regression: block: Add .bdrv_co_pwrite_zeroes()

2016-07-21 Thread Eric Blake
On 07/21/2016 03:08 AM, Paolo Bonzini wrote: > > > On 21/07/2016 01:35, Eric Blake wrote: >> Also, while the device is advertising that the optimal discard alignment >> is 15M, that does not tell me the minimum granularity that it can >> actually discard. Can you determine that value? That is,

Re: [Qemu-devel] Regression: block: Add .bdrv_co_pwrite_zeroes()

2016-07-21 Thread Paolo Bonzini
On 21/07/2016 01:35, Eric Blake wrote: > Also, while the device is advertising that the optimal discard alignment > is 15M, that does not tell me the minimum granularity that it can > actually discard. Can you determine that value? That is, if I try to > discard only 1M, does that actually resu

Re: [Qemu-devel] Regression: block: Add .bdrv_co_pwrite_zeroes()

2016-07-21 Thread Paolo Bonzini
On 21/07/2016 09:01, Peter Lieven wrote: > > maximum unmap lba count:30720 > maximum unmap block descriptor count:2 > optimal unmap granularity:30720 > ugavalid:1 > unmap granularity alignment:0 > maximum write same length:30720 Uhm, that's weird. The optimal unmap granularity should really be

Re: [Qemu-devel] Regression: block: Add .bdrv_co_pwrite_zeroes()

2016-07-21 Thread Peter Lieven
Hi Eric, Am 21.07.2016 um 01:35 schrieb Eric Blake: On 07/04/2016 07:49 AM, Peter Lieven wrote: Hi, the above commit: commit d05aa8bb4a8b6aa9a915ec5074fb12ae632d2323 Author: Eric Blake Date: Wed Jun 1 15:10:03 2016 -0600 block: Add .bdrv_co_pwrite_zeroes() introduces a regression (a

Re: [Qemu-devel] Regression: block: Add .bdrv_co_pwrite_zeroes()

2016-07-20 Thread wangweiwei
Sorry, reply is wrong. 在 2016年07月21日 09:38, wangweiwei 写道: 在 2016年07月20日 19:35, Eric Blake 写道: On 07/04/2016 07:49 AM, Peter Lieven wrote: Hi, the above commit: commit d05aa8bb4a8b6aa9a915ec5074fb12ae632d2323 Author: Eric Blake Date: Wed Jun 1 15:10:03 2016 -0600 block: Add .bdrv_co

Re: [Qemu-devel] Regression: block: Add .bdrv_co_pwrite_zeroes()

2016-07-20 Thread wangweiwei
在 2016年07月20日 19:35, Eric Blake 写道: On 07/04/2016 07:49 AM, Peter Lieven wrote: Hi, the above commit: commit d05aa8bb4a8b6aa9a915ec5074fb12ae632d2323 Author: Eric Blake Date: Wed Jun 1 15:10:03 2016 -0600 block: Add .bdrv_co_pwrite_zeroes() introduces a regression (at least for me).

Re: [Qemu-devel] Regression: block: Add .bdrv_co_pwrite_zeroes()

2016-07-20 Thread Eric Blake
On 07/04/2016 07:49 AM, Peter Lieven wrote: > Hi, > > the above commit: > > commit d05aa8bb4a8b6aa9a915ec5074fb12ae632d2323 > Author: Eric Blake > Date: Wed Jun 1 15:10:03 2016 -0600 > > block: Add .bdrv_co_pwrite_zeroes() > > introduces a regression (at least for me). > > The Limits fr

Re: [Qemu-devel] Regression: block: Add .bdrv_co_pwrite_zeroes()

2016-07-18 Thread Peter Lieven
Am 15.07.2016 um 17:40 schrieb Eric Blake: On 07/15/2016 04:09 AM, Peter Lieven wrote: Am 05.07.2016 um 17:09 schrieb Paolo Bonzini: On 05/07/2016 16:59, Eric Blake wrote: And yes, we could probably switch to (potentially slower) / % * instead of bit operations in block/io.c to accommodate a n

Re: [Qemu-devel] Regression: block: Add .bdrv_co_pwrite_zeroes()

2016-07-15 Thread Eric Blake
On 07/15/2016 04:09 AM, Peter Lieven wrote: > Am 05.07.2016 um 17:09 schrieb Paolo Bonzini: >> >> On 05/07/2016 16:59, Eric Blake wrote: >>> And yes, we could probably switch to (potentially slower) / % * instead >>> of bit operations in block/io.c to accommodate a non-power-of-2 optimal >>> size,

Re: [Qemu-devel] Regression: block: Add .bdrv_co_pwrite_zeroes()

2016-07-15 Thread Peter Lieven
Am 05.07.2016 um 17:09 schrieb Paolo Bonzini: > > On 05/07/2016 16:59, Eric Blake wrote: >> And yes, we could probably switch to (potentially slower) / % * instead >> of bit operations in block/io.c to accommodate a non-power-of-2 optimal >> size, but it would require a careful audit to make sure w

Re: [Qemu-devel] Regression: block: Add .bdrv_co_pwrite_zeroes()

2016-07-05 Thread Paolo Bonzini
On 05/07/2016 16:59, Eric Blake wrote: > And yes, we could probably switch to (potentially slower) / % * instead > of bit operations in block/io.c to accommodate a non-power-of-2 optimal > size, but it would require a careful audit to make sure we don't have > even more bit-wise operations lurkin

Re: [Qemu-devel] Regression: block: Add .bdrv_co_pwrite_zeroes()

2016-07-05 Thread Eric Blake
On 07/05/2016 07:37 AM, Paolo Bonzini wrote: > > > On 05/07/2016 03:53, Eric Blake wrote: I think we cannot assert that that these alignments are a power of 2. >> Perhaps that means we should just fix our code to round things down to >> the nearest power of 2 (8MB) for the opt_transfer_len a

Re: [Qemu-devel] Regression: block: Add .bdrv_co_pwrite_zeroes()

2016-07-05 Thread Peter Lieven
Am 05.07.2016 um 15:37 schrieb Paolo Bonzini: On 05/07/2016 03:53, Eric Blake wrote: I think we cannot assert that that these alignments are a power of 2. Perhaps that means we should just fix our code to round things down to the nearest power of 2 (8MB) for the opt_transfer_len and opt_discar

Re: [Qemu-devel] Regression: block: Add .bdrv_co_pwrite_zeroes()

2016-07-05 Thread Paolo Bonzini
On 05/07/2016 15:03, Eric Blake wrote: > bs->bl.pwrite_zeroes_alignment = > -iscsilun->bl.opt_unmap_gran * iscsilun->block_size; > +pow2floor(iscsilun->bl.opt_unmap_gran * iscsilun->block_size); > } else { > bs->bl.pwrite_zeroes_alignment = iscsilun-

Re: [Qemu-devel] Regression: block: Add .bdrv_co_pwrite_zeroes()

2016-07-05 Thread Paolo Bonzini
On 05/07/2016 03:53, Eric Blake wrote: >> > I think we cannot assert that that these alignments are a power of 2. > Perhaps that means we should just fix our code to round things down to > the nearest power of 2 (8MB) for the opt_transfer_len and > opt_discard_alignment values. Can you post a st

Re: [Qemu-devel] Regression: block: Add .bdrv_co_pwrite_zeroes()

2016-07-05 Thread Eric Blake
On 07/05/2016 01:30 AM, Peter Lieven wrote: > Am 05.07.2016 um 03:53 schrieb Eric Blake: >> On 07/04/2016 07:49 AM, Peter Lieven wrote: >>> Hi, >>> >>> the above commit: >>> >>> commit d05aa8bb4a8b6aa9a915ec5074fb12ae632d2323 >>> Author: Eric Blake >>> Date: Wed Jun 1 15:10:03 2016 -0600 >>> >>>

Re: [Qemu-devel] Regression: block: Add .bdrv_co_pwrite_zeroes()

2016-07-05 Thread Peter Lieven
Am 05.07.2016 um 03:53 schrieb Eric Blake: On 07/04/2016 07:49 AM, Peter Lieven wrote: Hi, the above commit: commit d05aa8bb4a8b6aa9a915ec5074fb12ae632d2323 Author: Eric Blake Date: Wed Jun 1 15:10:03 2016 -0600 block: Add .bdrv_co_pwrite_zeroes() introduces a regression (at least fo

Re: [Qemu-devel] Regression: block: Add .bdrv_co_pwrite_zeroes()

2016-07-04 Thread Eric Blake
On 07/04/2016 07:49 AM, Peter Lieven wrote: > Hi, > > the above commit: > > commit d05aa8bb4a8b6aa9a915ec5074fb12ae632d2323 > Author: Eric Blake > Date: Wed Jun 1 15:10:03 2016 -0600 > > block: Add .bdrv_co_pwrite_zeroes() > > introduces a regression (at least for me). > > The Limits fr

[Qemu-devel] Regression: block: Add .bdrv_co_pwrite_zeroes()

2016-07-04 Thread Peter Lieven
Hi, the above commit: commit d05aa8bb4a8b6aa9a915ec5074fb12ae632d2323 Author: Eric Blake Date: Wed Jun 1 15:10:03 2016 -0600 block: Add .bdrv_co_pwrite_zeroes() introduces a regression (at least for me). The Limits from the iSCSI Block Limits VPD have no requirement of being a power o