From: Pavel Butsykin <pbutsykin@virtuozzo.com>
To: Eric Blake <eblake@redhat.com>, "Denis V. Lunev" <den@openvz.org>,
qemu-devel@nongnu.org
Cc: Jeff Cody <jcody@redhat.com>,
Markus Armbruster <armbru@redhat.com>,
John Snow <jsnow@redhat.com>,
Stefan Hajnoczi <stefanha@redhat.com>,
Kevin Wolf <kwolf@redhat.com>
Subject: Re: [Qemu-devel] [PATCH 02/11] block/io: add bdrv_co_write_compressed
Date: Wed, 22 Jun 2016 15:26:07 +0300 [thread overview]
Message-ID: <576A83DF.6060202@virtuozzo.com> (raw)
In-Reply-To: <575EC3E2.1010304@redhat.com>
On 13.06.2016 17:32, Eric Blake wrote:
> On 05/31/2016 03:15 AM, Denis V. Lunev wrote:
>> From: Pavel Butsykin <pbutsykin@virtuozzo.com>
>>
>> This patch just adds the interface to the bdrv_co_pwritev_compressed,
>> which is currently not used but will be useful for safe implementation of the
>> bdrv_co_write_compressed callback in format drivers.
>>
>> Signed-off-by: Pavel Butsykin <pbutsykin@virtuozzo.com>
>> Signed-off-by: Denis V. Lunev <den@openvz.org>
>> CC: Jeff Cody <jcody@redhat.com>
>> CC: Markus Armbruster <armbru@redhat.com>
>> CC: Eric Blake <eblake@redhat.com>
>> CC: John Snow <jsnow@redhat.com>
>> CC: Stefan Hajnoczi <stefanha@redhat.com>
>> CC: Kevin Wolf <kwolf@redhat.com>
>> ---
>> block/io.c | 78 +++++++++++++++++++++++++++++++++++++++++++----
>> include/block/block_int.h | 5 +++
>> qemu-img.c | 2 +-
>> 3 files changed, 78 insertions(+), 7 deletions(-)
>>
>> diff --git a/block/io.c b/block/io.c
>> index c5bb6ae..54cd9a4 100644
>> --- a/block/io.c
>> +++ b/block/io.c
>> @@ -1779,8 +1779,8 @@ int bdrv_is_allocated_above(BlockDriverState *top,
>> return 0;
>> }
>>
>> -int bdrv_pwrite_compressed(BlockDriverState *bs, int64_t offset,
>> - const void *buf, int count)
>> +int coroutine_fn bdrv_co_pwritev_compressed(BlockDriverState *bs,
>> + int64_t offset, unsigned int bytes, QEMUIOVector *qiov)
>
> Why the rename s/count/bytes/? Would it be better to get the name right
> in 1/11?
It made like the bdrv_co_pwritev interface. For the
bdrv_pwrite_compressed() func also need rename s/count/bytes/ in 1/11.
>> {
>> BlockDriver *drv = bs->drv;
>> int ret;
>> @@ -1788,18 +1788,84 @@ int bdrv_pwrite_compressed(BlockDriverState *bs, int64_t offset,
>> if (!drv) {
>> return -ENOMEDIUM;
>> }
>> - if (!drv->bdrv_write_compressed) {
>> +
>> + if (!drv->bdrv_co_write_compressed) {
>> return -ENOTSUP;
>
> This is a (temporary) regression - since none of the drivers have
> .bdrv_co_write_compressed yet, you will always fail. Rather, the
> transition period should support both interfaces at once...
>
>> }
>> - ret = bdrv_check_byte_request(bs, offset, count);
>> +
>> + ret = bdrv_check_byte_request(bs, offset, bytes);
>> if (ret < 0) {
>> return ret;
>> }
>>
>> assert(QLIST_EMPTY(&bs->dirty_bitmaps));
>> + assert(qemu_in_coroutine());
>> +
>> + return drv->bdrv_co_write_compressed(bs, offset >> BDRV_SECTOR_BITS,
>> + bytes >> BDRV_SECTOR_BITS, qiov);
>
> ...and call into either the old or the new interface according to what
> is present.
>
>
>> +int bdrv_pwrite_compressed(BlockDriverState *bs, int64_t offset,
>> + const void *buf, int count)
>> +{
>> + BdrvWriteCompressedCo data;
>> + QEMUIOVector qiov;
>> + BlockDriver *drv = bs->drv;
>> + struct iovec iov = {
>> + .iov_base = (void *)buf,
>> + .iov_len = count,
>> + };
>> + qemu_iovec_init_external(&qiov, &iov, 1);
>>
>> - return drv->bdrv_write_compressed(bs, offset >> BDRV_SECTOR_BITS, buf,
>> - count >> BDRV_SECTOR_BITS);
>> + data = (BdrvWriteCompressedCo) {
>> + .bs = bs,
>> + .offset = offset,
>> + .qiov = &qiov,
>> + .ret = -EINPROGRESS,
>> + };
>> +
>> + if (!drv) {
>> + return -ENOMEDIUM;
>> + }
>> +
>> + if (drv->bdrv_write_compressed) {
>> + int ret = bdrv_check_byte_request(bs, offset, count);
>> + if (ret < 0) {
>> + return ret;
>> + }
>> + assert(QLIST_EMPTY(&bs->dirty_bitmaps));
>> + return drv->bdrv_write_compressed(bs, offset >> BDRV_SECTOR_BITS, buf,
>> + count >> BDRV_SECTOR_BITS);
>> + }
>
> Oh, you're catering to the old code up front, without a coroutine, and
> only the new code gets coroutine treatment. Maybe it's okay after all.
Yes, it is made for a gradual transition from
drv->bdrv_write_compressed to drv->bdrv_co_write_compressed.
>> +
>> + if (qemu_in_coroutine()) {
>> + /* Fast-path if already in coroutine context */
>> + bdrv_write_compressed_co_entry(&data);
>> + } else {
>> + AioContext *aio_context = bdrv_get_aio_context(bs);
>> +
>> + Coroutine *co = qemu_coroutine_create(bdrv_write_compressed_co_entry);
>> + qemu_coroutine_enter(co, &data);
>> + while (data.ret == -EINPROGRESS) {
>> + aio_poll(aio_context, true);
>> + }
>> + }
>> + return data.ret;
>> }
>>
>> int bdrv_save_vmstate(BlockDriverState *bs, const uint8_t *buf,
>> diff --git a/include/block/block_int.h b/include/block/block_int.h
>> index 30a9717..ccba9c9 100644
>> --- a/include/block/block_int.h
>> +++ b/include/block/block_int.h
>> @@ -207,6 +207,9 @@ struct BlockDriver {
>> int (*bdrv_write_compressed)(BlockDriverState *bs, int64_t sector_num,
>> const uint8_t *buf, int nb_sectors);
>>
>> + int coroutine_fn (*bdrv_co_write_compressed)(BlockDriverState *bs,
>
> Might be better to name this bdrv_co_pwrite_compressed if we want to
> make it byte-based...
>
>> + int64_t sector_num, int nb_sectors, QEMUIOVector *qiov);
>
> ...it seems odd that you have to add a new sector-based interface given
> that you are trying to convert to byte-based.
>
When I sent this patch, Kevin had not yet converted the sector-based
interfaces to byte-based for format drivers :)
>> +
>> int (*bdrv_snapshot_create)(BlockDriverState *bs,
>> QEMUSnapshotInfo *sn_info);
>> int (*bdrv_snapshot_goto)(BlockDriverState *bs,
>> @@ -535,6 +538,8 @@ int coroutine_fn bdrv_co_preadv(BlockDriverState *bs,
>> int coroutine_fn bdrv_co_pwritev(BlockDriverState *bs,
>> int64_t offset, unsigned int bytes, QEMUIOVector *qiov,
>> BdrvRequestFlags flags);
>> +int coroutine_fn bdrv_co_pwritev_compressed(BlockDriverState *bs,
>> + int64_t offset, unsigned int bytes, QEMUIOVector *qiov);
>>
>> int get_tmp_filename(char *filename, int size);
>> BlockDriver *bdrv_probe_all(const uint8_t *buf, int buf_size,
>> diff --git a/qemu-img.c b/qemu-img.c
>> index eb744d4..ab54027 100644
>> --- a/qemu-img.c
>> +++ b/qemu-img.c
>> @@ -2024,7 +2024,7 @@ static int img_convert(int argc, char **argv)
>> const char *preallocation =
>> qemu_opt_get(opts, BLOCK_OPT_PREALLOC);
>>
>> - if (!drv->bdrv_write_compressed) {
>> + if (!drv->bdrv_write_compressed && !drv->bdrv_co_write_compressed) {
>> error_report("Compression not supported for this file format");
>> ret = -1;
>> goto out;
>>
>
next prev parent reply other threads:[~2016-06-22 12:26 UTC|newest]
Thread overview: 47+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-05-31 9:15 [Qemu-devel] [PATCH v4 00/11] backup compression Denis V. Lunev
2016-05-31 9:15 ` [Qemu-devel] [PATCH 01/11] block: switch blk_write_compressed() to byte-based interface Denis V. Lunev
2016-06-13 13:01 ` Stefan Hajnoczi
2016-06-13 14:23 ` Eric Blake
2016-06-22 12:25 ` Pavel Butsykin
2016-06-28 10:53 ` Kevin Wolf
2016-06-28 11:32 ` Pavel Butsykin
2016-05-31 9:15 ` [Qemu-devel] [PATCH 02/11] block/io: add bdrv_co_write_compressed Denis V. Lunev
2016-06-13 13:18 ` Stefan Hajnoczi
2016-06-13 14:32 ` Eric Blake
2016-06-22 12:26 ` Pavel Butsykin [this message]
2016-06-28 11:09 ` Kevin Wolf
2016-06-28 11:35 ` Pavel Butsykin
2016-05-31 9:15 ` [Qemu-devel] [PATCH 03/11] qcow2: add qcow2_co_write_compressed Denis V. Lunev
2016-06-13 13:18 ` Stefan Hajnoczi
2016-06-13 20:14 ` Eric Blake
2016-06-22 12:27 ` Pavel Butsykin
2016-06-28 11:17 ` Kevin Wolf
2016-06-28 11:30 ` Kevin Wolf
2016-06-28 15:06 ` Pavel Butsykin
2016-05-31 9:15 ` [Qemu-devel] [PATCH 04/11] vmdk: add vmdk_co_write_compressed Denis V. Lunev
2016-06-13 13:18 ` Stefan Hajnoczi
2016-05-31 9:15 ` [Qemu-devel] [PATCH 05/11] qcow: add qcow_co_write_compressed Denis V. Lunev
2016-06-13 13:18 ` Stefan Hajnoczi
2016-05-31 9:15 ` [Qemu-devel] [PATCH 06/11] block: remove BlockDriver.bdrv_write_compressed Denis V. Lunev
2016-06-13 13:19 ` Stefan Hajnoczi
2016-05-31 9:15 ` [Qemu-devel] [PATCH 07/11] block: optimization blk_pwrite_compressed() Denis V. Lunev
2016-06-13 13:11 ` Stefan Hajnoczi
2016-06-13 20:16 ` Eric Blake
2016-06-15 10:22 ` Stefan Hajnoczi
2016-06-24 15:42 ` Eric Blake
2016-06-28 11:47 ` Kevin Wolf
2016-06-28 12:32 ` Pavel Butsykin
2016-06-28 13:05 ` Kevin Wolf
2016-05-31 9:15 ` [Qemu-devel] [PATCH 08/11] drive-backup: added support for data compression Denis V. Lunev
2016-06-13 13:19 ` Stefan Hajnoczi
2016-06-13 20:18 ` Eric Blake
2016-06-23 9:15 ` Pavel Butsykin
2016-06-24 15:39 ` Eric Blake
2016-06-24 15:50 ` Pavel Butsykin
2016-05-31 9:15 ` [Qemu-devel] [PATCH 09/11] blockdev-backup: " Denis V. Lunev
2016-06-13 13:19 ` Stefan Hajnoczi
2016-06-13 20:19 ` Eric Blake
2016-05-31 9:15 ` [Qemu-devel] [PATCH 10/11] qemu-iotests: test backup compression in 055 Denis V. Lunev
2016-06-13 13:17 ` Stefan Hajnoczi
2016-05-31 9:15 ` [Qemu-devel] [PATCH 11/11] qemu-iotests: add vmdk for " Denis V. Lunev
2016-06-13 13:18 ` Stefan Hajnoczi
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=576A83DF.6060202@virtuozzo.com \
--to=pbutsykin@virtuozzo.com \
--cc=armbru@redhat.com \
--cc=den@openvz.org \
--cc=eblake@redhat.com \
--cc=jcody@redhat.com \
--cc=jsnow@redhat.com \
--cc=kwolf@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=stefanha@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).