Date: Mon, 28 Nov 2016 20:17:34 -0700 From: Ian Lepore <ian@freebsd.org> To: Mori Hiroki <yamori813@yahoo.co.jp>, "freebsd-arm@freebsd.org" <freebsd-arm@freebsd.org> Subject: Re: FDT_DTB_STATIC and LINUX_BOOT_ABI Message-ID: <1480389454.1889.153.camel@freebsd.org> In-Reply-To: <105386.51359.qm@web101710.mail.ssk.yahoo.co.jp> References: <105386.51359.qm@web101710.mail.ssk.yahoo.co.jp>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, 2016-11-28 at 23:33 +0900, Mori Hiroki wrote: > Hi. > > I have problem at u-boot arm module. > > I found double size memory. > > Because of u-boot parameter have ATAG_MEM and dts have memory entory. > It's doble count. > > I think check FDT_DTB_STATIC at linux_parse_boot_param() is good. > > #if !defined(FDT_DTB_STATIC) > > case ATAG_MEM: > arm_physmem_hardware_region(walker->u.tag_mem.start, > walker->u.tag_mem.size); > break; > #endif > > How about this? > > Regards > > Hiroki Mori What if someone wants to use a static dtb, but leaves the memory out of it so that u-boot can report the actual memory via atag? I think a better fix might be to have arm_physmem_hardware_region() handle the case where the same region is added twice. Can you do a verbose boot, or on a running system break into the kernel debugger using "<CR> ~ ^B" and do a "show physmem" (then you can "cont" to resume running the kernel). I'm looking for this output: Physical memory chunk(s): 0x00001000 - 0x1dffffff, 479 MB ( 122879 pages) Excluded memory regions: 0x00400000 - 0x00d44fff, 9 MB ( 2373 pages) NoAlloc 0x08000000 - 0x0fffffff, 128 MB ( 32768 pages) NoAlloc NoDump -- Ian
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1480389454.1889.153.camel>