Date: Sat, 29 Sep 2007 10:20:41 -0600 From: Scott Long <scottl@samsco.org> To: Hans Petter Selasky <hselasky@c2i.net> Cc: freebsd-arch@freebsd.org Subject: Re: Request for feedback on common data backstore in the kernel Message-ID: <46FE7B59.7080003@samsco.org> In-Reply-To: <200709291309.05747.hselasky@c2i.net> References: <200709262157.02712.hselasky@c2i.net> <200709281753.30367.hselasky@c2i.net> <20070928.141345.78789158.imp@bsdimp.com> <200709291309.05747.hselasky@c2i.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Hans Petter Selasky wrote:
> On Friday 28 September 2007, Warner Losh wrote:
>
>>Hans Petter Selasky wrote:
>>
>>>>Well, the point is that I'm not sure why you're so worried about
>>>>performance issues with USB and busdma. Do you have any test data that
>>>>shows that it's a problem?
>>>
>>>It is a problem on embedded devices. Typically not for an ordinary PC
>>>user.
>>
>>I'm curious. What's the problem?
>
>
> Technically, if you load a buffer pointer and length that is not cache aligned
> nor host aligned then you should bounce all the data no matter what. If you
> do not bounce under those conditions you might get trouble.
>
> struct my_softc {
> uint8_t hdr[2];
> uint8_t eth_pkt[1500];
> uint8_t other_important_stuff[546];
> } __aligned(PAGE_SIZE);
>
> Can "eth_pkt" be loaded into DMA ? No, it has an offset of 2.
>
> struct my_softc {
> uint8_t eth_pkt[1500];
> uint8_t other_important_stuff[548];
> } __aligned(PAGE_SIZE);
>
> Can "eth_pkt" be loaded into DMA ? No, "other_important_stuff" might get lost
> when the cache is invalidated.
>
> struct my_softc {
> uint8_t eth_pkt[1500];
> } __aligned(PAGE_SIZE);
>
> Can "eth_pkt" be loaded into DMA ? Yes, I see no problem.
>
> Is it guaranteed that the "m_data" field of a mbuf will always be correctly
> aligned for the CPU/DMA cache in addition to the host alignment ?
>
As Warner said, a lot of this is dependent on the device hardware, not
the CPU, and it's a problem that is already well understood and
addressed in those areas in FreeBSD.
> Does bus_dma support bouncing using multiple pages instead of using one
> contiguous segment ?
Yes, busdma bounces on a page-by-page basis. In fact, when doing
re-alignments, it'll attempt to bounce only the begining segment instead
of the entire buffer.
>
> Please correct me if I am wrong.
busdma is something that is both woefully underdocumented and much more
powerful and featureful than is what is available in most other OS's, so
please feel free to ask more questions like these.
Scott
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?46FE7B59.7080003>