[Top][All Lists]

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

[Qemu-block] "Internal" bdrv-dirty-bitmaps

From: John Snow
Subject: [Qemu-block] "Internal" bdrv-dirty-bitmaps
Date: Mon, 8 Jul 2019 17:53:11 -0400
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101 Thunderbird/60.7.2

Hi Markus;

recently it's come up a few times that it might be nice to hide bitmaps
that have no names from the query results, because these are "internal"
bitmaps used by QEMU that the user cannot manipulate.

Actually, all of the bitmap-finding functions will refuse to touch
bitmaps with no name.

We added names to bitmaps with 0db6e54a8a2, but bitmaps existed prior to
this and already had a "query" mechanism -- meaning that the "internal"
use bitmaps created by mirror have always been user-visible.

commit 21b56835 added the bdrv_query_dirty_bitmaps function, replacing
and older one-per-node bitmap querying utility, which... leads me to
believe that we've always explicitly exposed these bitmaps...

Further back in time (2012), we see b9a9b3a4626, which explicitly
exposed this information -- intentionally showing the bitmap dirty count
for... either migration or mirror, it could have been both.

I guess it's back from
https://lists.gnu.org/archive/html/qemu-devel/2012-09/msg04545.html --
and just assumed we'd want to see this count during operations that used it.

I guess at this point I definitely can't sweep the last 7 years under a
mat. The problem I have with it is that without a name, these bitmaps
are meaningless to an outside viewer. There could be many no-name
bitmaps at once, each correlating to different operations that could
finish and delete them at any time.

Are these useful to still have? If not, is there a reasonable way to
announce their farewell party by the deprecation policy?


reply via email to

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