From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
To: Eric Blake <eblake@redhat.com>, qemu-devel@nongnu.org
Cc: kwolf@redhat.com, stefanha@redhat.com, pkrempa@redhat.com,
rjones@redhat.com, qemu-block@nongnu.org,
Max Reitz <mreitz@redhat.com>
Subject: Re: [PATCH v2 4/5] nbd: Add new qemu:allocation-depth metacontext
Date: Thu, 8 Oct 2020 16:04:16 +0300 [thread overview]
Message-ID: <b1d3536f-0fcc-11e9-13a6-db144b8a4368@virtuozzo.com> (raw)
In-Reply-To: <f839054c-70e6-cfe5-0c44-e9a519389d65@redhat.com>
08.10.2020 00:54, Eric Blake wrote:
> On 10/7/20 8:40 AM, Vladimir Sementsov-Ogievskiy wrote:
>> 30.09.2020 15:11, Eric Blake wrote:
>>> 'qemu-img map' provides a way to determine which extents of an image
>>> come from the top layer vs. inherited from a backing chain. This is
>>> useful information worth exposing over NBD. There is a proposal to
>>> add a QMP command block-dirty-bitmap-populate which can create a dirty
>>> bitmap that reflects allocation information, at which point
>>> qemu:dirty-bitmap:NAME can expose that information via the creation of
>>> a temporary bitmap, but we can shorten the effort by adding a new
>>> qemu:allocation-depth context that does the same thing without an
>>> intermediate bitmap (this patch does not eliminate the need for that
>>> proposal, as it will have other uses as well).
>>>
>>> For this patch, I just encoded a tri-state value (unallocated, from
>>> this layer, from any of the backing layers); an obvious extension
>>> would be to provide the actual depth in bits 31-4 while keeping bits
>>> 1-0 as a tri-state (leaving bits 3-2 unused, for ease of reading depth
>>> from a hex number). But this extension would require
>>> bdrv_is_allocated_above to return a depth number.
>>>
>>> Note that this patch does not actually enable any way to request a
>>> server to enable this context; that will come in the next patch.
>>>
>
>>> +In the allocation depth context, bits 0 and 1 form a tri-state value:
>>> +
>>> + bits 0-1 clear: NBD_STATE_DEPTH_UNALLOC, means the extent is
>>> unallocated
>>> + bit 0 set: NBD_STATE_DEPTH_LOCAL, the extent is allocated in this
>>> image
>>
>> Maybe, s/this/the top level/, as, what is "this" for NBD client?
>
> Sure.
>
>
>>> @@ -864,12 +867,21 @@ static bool nbd_meta_qemu_query(NBDClient
>>> *client, NBDExportMetaContexts *meta,
>>>
>>> if (!*query) {
>
> We get here for "qemu:".
>
>>> if (client->opt == NBD_OPT_LIST_META_CONTEXT) {
>>> + meta->allocation_depth = meta->exp->alloc_context;
>>> meta->bitmap = !!meta->exp->export_bitmap;
>>> }
>>> trace_nbd_negotiate_meta_query_parse("empty");
>>> return true;
>>> }
>>>
>>> + if (nbd_strshift(&query, "allocation-depth")) {
>
> We get here for "qemu:allocation-depth", and as you pointed out,
> "qemu:allocation-depth-extended".
>
>>> + trace_nbd_negotiate_meta_query_parse("allocation-depth");
>>> + if (nbd_meta_empty_or_pattern(client, "", query)) {
>>
>> How much it differs from "if (!*query) {", I don't see...
>
> The nbd_meta_empty_or_pattern returns false for
> "qemu:allocation-depth-extended"; it only returns true for
> "qemu:allocation-depth". But, as you pointed out,
>
>>
>>> + meta->allocation_depth = meta->exp->alloc_context;
>>> + }
>>> + return true;
>>> + }
>>
>> why not just
>>
>> if (!strcmp(query, "allocation-depth")) {
>> meta->allocation_depth = meta->exp->alloc_context;
>> return true;
>> }
>>
>> ?
>
> that does seem shorter.
>
>>
>> With your code you also parse something like
>> "allocation-depth-extended".. Is it on purpose?
>
> The string is parsed, but does not affect meta->XXX, which means nothing
> gets advertised to the client. The trace messages might differ, but
> behavior is correct.
>
>>
>>> +
>>> if (nbd_strshift(&query, "dirty-bitmap:")) {
>>> trace_nbd_negotiate_meta_query_parse("dirty-bitmap:");
>>> if (!meta->exp->export_bitmap) {
>>
>>
>> Also, "trace_nbd_negotiate_meta_query_skip("not dirty-bitmap"); " at
>> function end should
>> now be something like trace_nbd_negotiate_meta_query_skip("unknown
>> context in qemu: namespace");
>
> Good idea.
>
>
>>> +/* Get allocation depth from the exported device and send it to the
>>> client */
>>> +static int nbd_co_send_block_alloc(NBDClient *client, uint64_t handle,
>>> + BlockDriverState *bs, uint64_t
>>> offset,
>>> + uint32_t length, bool dont_fragment,
>>> + bool last, uint32_t context_id,
>>> + Error **errp)
>>> +{
>>> + int ret;
>>> + unsigned int nb_extents = dont_fragment ? 1 :
>>> NBD_MAX_BLOCK_STATUS_EXTENTS;
>>> + g_autoptr(NBDExtentArray) ea = nbd_extent_array_new(nb_extents);
>>> +
>>> + ret = blockalloc_to_extents(bs, offset, length, ea);
>>> + if (ret < 0) {
>>> + return nbd_co_send_structured_error(
>>> + client, handle, -ret, "can't get block status", errp);
>>> + }
>>> +
>>> + return nbd_co_send_extents(client, handle, ea, last, context_id,
>>> errp);
>>> +}
>>
>>
>> The function is a copy of nbd_co_send_block_status()..
>>
>> Probably, we can reuse nbd_co_send_block_status(), and just call
>> blockstatus_to_extents()
>> or blockalloc_to_extents() depending on context_id parameter.
>
> Nice idea to reduce the duplication.
>
>>
>> Still, I'm OK with it as is.
>>
it was only about the duplicated function
>
> So is that Reviewed-by:, or do I need to post v3 with my tweaks in
> response to your comments?
>
Honestly it wasn't.. So, I found some things to discuss, and it's possible that I've looked through with less attention. Hm. So, with v3 or without, I've have to lookthrough it again..
My suggested diff:
diff --git a/docs/interop/nbd.txt b/docs/interop/nbd.txt
index 56efec7aee..a0d91ff08b 100644
--- a/docs/interop/nbd.txt
+++ b/docs/interop/nbd.txt
@@ -36,7 +36,8 @@ the image, with a single context named:
In the allocation depth context, bits 0 and 1 form a tri-state value:
bits 0-1 clear: NBD_STATE_DEPTH_UNALLOC, means the extent is unallocated
- bit 0 set: NBD_STATE_DEPTH_LOCAL, the extent is allocated in this image
+ bit 0 set: NBD_STATE_DEPTH_LOCAL, the extent is allocated in the top
+ level image
bit 1 set: NBD_STATE_DEPTH_BACKING, the extent is inherited from a
backing layer
diff --git a/include/block/nbd.h b/include/block/nbd.h
index 06208bc250..2d7165960d 100644
--- a/include/block/nbd.h
+++ b/include/block/nbd.h
@@ -262,7 +262,7 @@ enum {
/* Extent flags for qemu:allocation-depth in NBD_REPLY_TYPE_BLOCK_STATUS */
#define NBD_STATE_DEPTH_UNALLOC (0 << 0)
#define NBD_STATE_DEPTH_LOCAL (1 << 0)
-#define NBD_STATE_DEPTH_BACKING (2 << 0)
+#define NBD_STATE_DEPTH_BACKING (1 << 1)
#define NBD_STATE_DEPTH_MASK (0x3)
static inline bool nbd_reply_type_is_error(int type)
diff --git a/nbd/server.c b/nbd/server.c
index 830b21000b..3761ac180f 100644
--- a/nbd/server.c
+++ b/nbd/server.c
@@ -809,7 +809,7 @@ static bool nbd_meta_empty_or_pattern(NBDClient *client, const char *pattern,
{
if (!*query) {
trace_nbd_negotiate_meta_query_parse("empty");
- return !*pattern || client->opt == NBD_OPT_LIST_META_CONTEXT;
+ return client->opt == NBD_OPT_LIST_META_CONTEXT;
}
if (strcmp(query, pattern) == 0) {
trace_nbd_negotiate_meta_query_parse(pattern);
@@ -874,11 +874,9 @@ static bool nbd_meta_qemu_query(NBDClient *client, NBDExportMetaContexts *meta,
return true;
}
- if (nbd_strshift(&query, "allocation-depth")) {
+ if (!strcmp(query, "allocation-depth")) {
trace_nbd_negotiate_meta_query_parse("allocation-depth");
- if (nbd_meta_empty_or_pattern(client, "", query)) {
- meta->allocation_depth = meta->exp->alloc_context;
- }
+ meta->allocation_depth = meta->exp->alloc_context;
return true;
}
@@ -896,7 +894,7 @@ static bool nbd_meta_qemu_query(NBDClient *client, NBDExportMetaContexts *meta,
return true;
}
- trace_nbd_negotiate_meta_query_skip("not dirty-bitmap");
+ trace_nbd_negotiate_meta_query_skip("unknown context in qemu: namespace");
return true;
}
@@ -2045,7 +2043,11 @@ static int nbd_co_send_extents(NBDClient *client, uint64_t handle,
return nbd_co_send_iov(client, iov, 2, errp);
}
-/* Get block status from the exported device and send it to the client */
+/*
+ * Get block status from the exported device and send it to the client,
+ * handling base:allocation or qemu:allocation-depth meta-context, based
+ * on @context_id value.
+ */
static int nbd_co_send_block_status(NBDClient *client, uint64_t handle,
BlockDriverState *bs, uint64_t offset,
uint32_t length, bool dont_fragment,
@@ -2056,27 +2058,12 @@ static int nbd_co_send_block_status(NBDClient *client, uint64_t handle,
unsigned int nb_extents = dont_fragment ? 1 : NBD_MAX_BLOCK_STATUS_EXTENTS;
g_autoptr(NBDExtentArray) ea = nbd_extent_array_new(nb_extents);
- ret = blockstatus_to_extents(bs, offset, length, ea);
- if (ret < 0) {
- return nbd_co_send_structured_error(
- client, handle, -ret, "can't get block status", errp);
+ if (context_id == NBD_META_ID_BASE_ALLOCATION) {
+ ret = blockstatus_to_extents(bs, offset, length, ea);
+ } else {
+ assert(context_id == NBD_META_ID_ALLOCATION_DEPTH);
+ ret = blockalloc_to_extents(bs, offset, length, ea);
}
-
- return nbd_co_send_extents(client, handle, ea, last, context_id, errp);
-}
-
-/* Get allocation depth from the exported device and send it to the client */
-static int nbd_co_send_block_alloc(NBDClient *client, uint64_t handle,
- BlockDriverState *bs, uint64_t offset,
- uint32_t length, bool dont_fragment,
- bool last, uint32_t context_id,
- Error **errp)
-{
- int ret;
- unsigned int nb_extents = dont_fragment ? 1 : NBD_MAX_BLOCK_STATUS_EXTENTS;
- g_autoptr(NBDExtentArray) ea = nbd_extent_array_new(nb_extents);
-
- ret = blockalloc_to_extents(bs, offset, length, ea);
if (ret < 0) {
return nbd_co_send_structured_error(
client, handle, -ret, "can't get block status", errp);
@@ -2433,13 +2420,13 @@ static coroutine_fn int nbd_handle_request(NBDClient *client,
}
if (client->export_meta.allocation_depth) {
- ret = nbd_co_send_block_alloc(client, request->handle,
- blk_bs(exp->common.blk),
- request->from, request->len,
- dont_fragment,
- !--contexts_remaining,
- NBD_META_ID_ALLOCATION_DEPTH,
- errp);
+ ret = nbd_co_send_block_status(client, request->handle,
+ blk_bs(exp->common.blk),
+ request->from, request->len,
+ dont_fragment,
+ !--contexts_remaining,
+ NBD_META_ID_ALLOCATION_DEPTH,
+ errp);
if (ret < 0) {
return ret;
}
with it:
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
--
Best regards,
Vladimir
next prev parent reply other threads:[~2020-10-08 13:10 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-09-30 12:11 [PATCH v2 0/5] Exposing backing-chain allocation over NBD Eric Blake
2020-09-30 12:11 ` [PATCH v2 1/5] qemu-nbd: Honor SIGINT and SIGHUP Eric Blake
2020-10-07 10:32 ` Vladimir Sementsov-Ogievskiy
2020-10-07 21:13 ` Eric Blake
2020-10-08 8:27 ` Vladimir Sementsov-Ogievskiy
2020-09-30 12:11 ` [PATCH v2 2/5] nbd/server: Reject embedded NUL in NBD strings Eric Blake
2020-10-07 10:39 ` Vladimir Sementsov-Ogievskiy
2020-09-30 12:11 ` [PATCH v2 3/5] nbd: Simplify meta-context parsing Eric Blake
2020-10-07 11:51 ` Vladimir Sementsov-Ogievskiy
2020-10-07 21:28 ` Eric Blake
2020-09-30 12:11 ` [PATCH v2 4/5] nbd: Add new qemu:allocation-depth metacontext Eric Blake
2020-10-07 13:40 ` Vladimir Sementsov-Ogievskiy
2020-10-07 21:54 ` Eric Blake
2020-10-08 13:04 ` Vladimir Sementsov-Ogievskiy [this message]
2020-09-30 12:11 ` [PATCH v2 5/5] nbd: Add 'qemu-nbd -A' to expose allocation depth Eric Blake
2020-10-07 14:03 ` Vladimir Sementsov-Ogievskiy
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=b1d3536f-0fcc-11e9-13a6-db144b8a4368@virtuozzo.com \
--to=vsementsov@virtuozzo.com \
--cc=eblake@redhat.com \
--cc=kwolf@redhat.com \
--cc=mreitz@redhat.com \
--cc=pkrempa@redhat.com \
--cc=qemu-block@nongnu.org \
--cc=qemu-devel@nongnu.org \
--cc=rjones@redhat.com \
--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).