From: "ext Aguirre, Sergio" <saaguirre@xxxxxx> Subject: RE: [Query][omap iommu] Consulting iommu if a physical region is "mappable" before actually mapping it Date: Mon, 3 May 2010 21:23:46 +0200 > > >> -----Original Message----- >> From: Kanigeri, Hari >> Sent: Monday, May 03, 2010 2:09 PM >> To: Aguirre, Sergio; Hiroshi DOYU >> Cc: linux-omap@xxxxxxxxxxxxxxx >> Subject: RE: [Query][omap iommu] Consulting iommu if a physical region is >> "mappable" before actually mapping it >> >> Sergio, >> >> > >> > Can the iommu driver be "consulted" if a certain area (contiguous or >> not) >> > can be mapped or not, before even trying to do it? >> > >> >> -- As long as there are physical pages backing the area it should be >> mappable right ? > > Ok, well.. maybe my question was more about that, if the ISP MMU has > some kind of limited translation table size, in which it has a > limited of pages to map to device addresses that the ISP can use. > > The need for this is that, in camera, there's a Video4Linux2 IOCTL > to know if a certain amount of buffers can be handled (and that > includes if they could be mapped or not), even before actually doing > the whole allocation/mapping. > > If there's no limitation in the translation table size, then I guess > the only concern is to ensure we have enough free RAM to use. (Which > I'll like to think that an userspace resource manager will be able > to figure out for the camera application, which is the one > allocating the buffers.) Right, currently a whole 4GB address space is available, so virtually no limitation for the translation table size. -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
