Hello! Thank you very much for your feedback.
> Nack, I don't like the patch as is. Starting with the unoptimized code
> movement combined with functional changes
Will it be OK if i split it into two parts:
1. Restructuring, introduction of object_property_add_single()
2. Introduction of cach
Am 03.07.2015 um 15:20 schrieb Pavel Fedin:
>> Ping!
>
> May be you could PULL it? Andreas seems to be down...
Nack, I don't like the patch as is. Starting with the unoptimized code
movement combined with functional changes, up to the use of [*] for the
new property iiuc.
Andreas
--
SUSE Linu
> Ping!
May be you could PULL it? Andreas seems to be down...
Kind regards,
Pavel Fedin
Expert Engineer
Samsung Electronics Research center Russia
s+p.fedin=samsung@nongnu.org] On Behalf Of Peter Crosthwaite
> Sent: Saturday, June 27, 2015 6:28 AM
> To: Pavel Fedin
> Cc: QEMU Developers; Andreas Färber
> Subject: Re: [Qemu-devel] [PATCH v3] QOM: object_property_add() performance
> improvement
>
> Ping!
>
>
Ping!
Regards,
Peter
On Tue, Jun 16, 2015 at 12:50 AM, Pavel Fedin wrote:
> The function originally behaves very badly when adding properties with "[*]"
> suffix. Normally these are used for numbering IRQ pins. In order to find the
> correct starting number the function started from zero and che
The function originally behaves very badly when adding properties with "[*]"
suffix. Normally these are used for numbering IRQ pins. In order to find the
correct starting number the function started from zero and checked for
duplicates. This takes incredibly long time with large number of CPUs beca
