[libvirt] [PATCH 2/6] port allocator: remove range on manual port reserving

[Michal Privoznik]

On 12/20/2017 07:35 AM, Nikolay Shirokovskiy wrote:
> Range check in virPortAllocatorSetUsed is not useful anymore
> when we manage ports for entire unsigned short range values.
> ---
>  src/bhyve/bhyve_command.c   |  4 +---
>  src/bhyve/bhyve_process.c   |  4 +---
>  src/qemu/qemu_process.c     | 41 +++++++++++------------------------------
>  src/util/virportallocator.c |  9 +--------
>  src/util/virportallocator.h |  4 +---
>  5 files changed, 15 insertions(+), 47 deletions(-)
> 

> diff --git a/src/util/virportallocator.c b/src/util/virportallocator.c
> index cd64356..76346c7 100644
> --- a/src/util/virportallocator.c
> +++ b/src/util/virportallocator.c
> @@ -293,8 +293,7 @@ int virPortAllocatorRelease(virPortRangePtr range,
>      return ret;
>  }
>  
> -int virPortAllocatorSetUsed(virPortRangePtr range,
> -                            unsigned short port,
> +int virPortAllocatorSetUsed(unsigned short port,
>                              bool value)
>  {
>      int ret = -1;
> @@ -305,12 +304,6 @@ int virPortAllocatorSetUsed(virPortRangePtr range,
>  
>      virObjectLock(pa);
>  
> -    if (port < range->start ||
> -        port > range->end) {
> -        ret = 0;
> -        goto cleanup;
> -    }
> -


1: ^^^

>      if (value) {
>          if (virBitmapIsBitSet(pa->bitmap, port) ||
>              virBitmapSetBit(pa->bitmap, port) < 0) {
> diff --git a/src/util/virportallocator.h b/src/util/virportallocator.h
> index e9b9038..8511eca 100644
> --- a/src/util/virportallocator.h
> +++ b/src/util/virportallocator.h
> @@ -45,8 +45,6 @@ int virPortAllocatorAcquire(virPortRangePtr range,
>  int virPortAllocatorRelease(virPortRangePtr range,
>                              unsigned short port);
>  
> -int virPortAllocatorSetUsed(virPortRangePtr range,
> -                            unsigned short port,
> -                            bool value);
> +int virPortAllocatorSetUsed(unsigned short port, bool value);

I'm not a big fan of this. Since virPortRange (or whatever name you'll
give it) is going to have a pointer to virPortAllocator object I rather
leave this signature as is. The idea (hunk 1) makes sense though

Michal