netdev
[Top] [All Lists]

Re: [PATCH][6/12][RFC/v1] Add IPoIB (IP-over-InfiniBand) driver

To: Roland Dreier <roland@xxxxxxxxxxx>
Subject: Re: [PATCH][6/12][RFC/v1] Add IPoIB (IP-over-InfiniBand) driver
From: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Date: Mon, 22 Nov 2004 20:04:39 +0000
Cc: Christoph Hellwig <hch@xxxxxxxxxxxxx>, netdev@xxxxxxxxxxx
In-reply-to: <521xelemsm.fsf@xxxxxxxxxxx>
References: <200411181046.6Dz1IPtDKfpgSXN9@xxxxxxxxxxx> <200411181046.Jj0jsF5E2KmiGN8f@xxxxxxxxxxx> <20041118200017.GA26976@xxxxxxxxxxxxx> <52sm71ex0l.fsf@xxxxxxxxxxx> <20041122153456.GA4905@xxxxxxxxxxxxx> <521xelemsm.fsf@xxxxxxxxxxx>
Sender: netdev-bounce@xxxxxxxxxxx
User-agent: Mutt/1.4.1i
On Mon, Nov 22, 2004 at 10:55:21AM -0800, Roland Dreier wrote:
>     Christoph> That just means avoidable interface changes sooner or
>     Christoph> later.  The dma_ interfaces also have the added benefit
>     Christoph> of allowing better control of the type of consistant
>     Christoph> allocation interface.
> 
> Is there a dma_xxx equivalent to DECLARE_PCI_UNMAP_ADDR(),
> pci_unmap_addr_set() and friends, or does a driver always have to save
> off the DMA address from dma_map_xxx?

Good question.  In pratice I can't find a case were DECLARE_PCI_UNMAP_ADDR() 
wouldn't simply work for dma_*, but we should probably document that
somewhere.


<Prev in Thread] Current Thread [Next in Thread>