Re: [libvirt] [PATCH v2 07/20] network: Alter virNetworkObj @class_id to be @classIdMap

On 08/15/2017 11:32 AM, Michal Privoznik wrote: > On 07/26/2017 05:05 PM, John Ferlan wrote: >> Change the variable name to be a bit more descriptive and less confusing >> when used with the data.network.actual->class_id. >> >> Signed-off-by: John Ferlan <jferlan(a)redhat.com&gt; >> --- >> src/conf/virnetworkobj.c | 39 ++++++++++++++++++++------------------- >> src/conf/virnetworkobj.h | 2 +- >> src/network/bridge_driver.c | 10 +++++----- >> 3 files changed, 26 insertions(+), 25 deletions(-) >> >> diff --git a/src/conf/virnetworkobj.c b/src/conf/virnetworkobj.c >> index 533ed26..fb533b9 100644 >> --- a/src/conf/virnetworkobj.c >> +++ b/src/conf/virnetworkobj.c >> @@ -79,13 +79,13 @@ virNetworkObjNew(void) >> if (!(net = virObjectLockableNew(virNetworkObjClass))) >> return NULL; >> >> - if (!(net->class_id = virBitmapNew(CLASS_ID_BITMAP_SIZE))) >> + if (!(net->classIdMap = virBitmapNew(CLASS_ID_BITMAP_SIZE))) >> goto error; > > One thing I just realized: This creates a bitmap capable of holding > 1<<16 bits - IOW it allocates 8KB of ram even though in reality just the > first 10-20 bits are going to be used. Back in the day when I was > implementing this there were no self-inflating bitmaps, but I guess we > do have them now, right? Maybe we can rewrite this (ofc after you merge > these to avoid conflicts) so that the new bitmap is used? > > Michal > I never gave it a second thought ;-) I haven't really dug into the bitmap code, but I see virBitmapExpand, virBitmapSetBitExpand, and virBitmapClearBitExpand that would seem to be useful for someone willing to make that kind of alteration.