Date: Fri, 22 Jan 1999 16:28:04 -0700 (MST) From: "Justin T. Gibbs" <gibbs@narnia.plutotech.com> To: mjacob@feral.com Cc: scsi@FreeBSD.ORG Subject: Re: Problems with QLogic SCSI Message-ID: <199901222328.QAA10113@narnia.plutotech.com> In-Reply-To: <Pine.LNX.4.04.9901221110430.1882-100000@feral-gw>
next in thread | previous in thread | raw e-mail | index | archive | help
In article <Pine.LNX.4.04.9901221110430.1882-100000@feral-gw> you wrote: > > > >> > And what *might* be happening (which would be *really* annoying) is if >> > bus_dmamem_alloc is allocating something that spans pages, although I >> > doubt that what you see would be the symptoms if this were the case. >> > >> > -matt >> >> It will only span pages if the allocation size is > PAGE_SIZE. > > It doesn't (currently at 4096)- and I need to figure out how to make the > bus_dma_tag_create call in isp_pci_mbxdma set things up so that you can at > least *try* and get > pagesized allocations (maybe with the filter > function- or maybe just try with a boundary limit and if you lose, you > lose). Currently the allocator will give you pages that our contiguous in bus space if the allocation spans pages. The interface does allow for you to specify the number of discontinuities in the allocation (segment count in the dma tag), but this is currently ignored. All pages are allocated from the 'safe area' specified by the tag so the filter func is never used for this kind of allocation. > -matt -- Justin To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199901222328.QAA10113>