Re: [PATCH] KVM PV Guest: Implement paravirtualized DMA

!MAILaRCHIVE_VOTE_RePLACE
Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]
To: Amit Shah <amit.shah@...>
Cc: <kvm-devel@...>, <chrisw@...>, <allen.m.kay@...>, <linux-kernel@...>, <gcosta@...>, <avi@...>, <virtualization@...>, <BENAMI@...>
Date: Tuesday, April 29, 2008 - 9:31 am

Amit Shah <amit.shah@qumranet.com> writes:
Bad variable names


I suspect real dma ops stacking will need some further thought than
your simple hacks

Note that on 32bit page_address() might be not mapped.



Are you sure that's correct? It looks quite bogus. order is a 2 logarithm,
normally npages = 1 << order

if you want  npages from order the correct  need 1 << order

Haven't read further, but to be honest the code doesn't seem to be anywhere
near merging quality.

-Andi

--
Previous message: [thread] [date] [author]
Next message: [thread] [date] [author]

Messages in current thread:
PV DMA for PCI passthrough devices for KVM, Amit Shah, (Tue Apr 29, 6:37 am)
Re: PV DMA for PCI passthrough devices for KVM, Andi Kleen, (Tue Apr 29, 9:15 am)
[PATCH] KVM PV Guest: Implement paravirtualized DMA, Amit Shah, (Tue Apr 29, 6:37 am)
Re: [PATCH] KVM PV Guest: Implement paravirtualized DMA, Andi Kleen, (Tue Apr 29, 9:31 am)