Date: Mon, 12 Nov 2018 22:58:13 -0600 From: Kyle Evans <kevans@freebsd.org> To: marklmi@yahoo.com Cc: Sean Bruno <sbruno@freebsd.org>, ports-list freebsd <freebsd-ports@freebsd.org>, Warner Losh <imp@freebsd.org> Subject: Re: qemu-arm-static: bsd-user/arm/target_syscall.h: #define TARGET_HW_MACHINE_ARCH "armv6" // what of armv7? Message-ID: <CACNAnaGesvCJp=eip3319BQwXixwyaYFLmZMGpisnQjp8w8_XA@mail.gmail.com> In-Reply-To: <3B210EB1-E7A8-4C77-B418-D82C58AE7290@yahoo.com> References: <3B210EB1-E7A8-4C77-B418-D82C58AE7290@yahoo.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, Nov 12, 2018 at 10:41 PM Mark Millard <marklmi@yahoo.com> wrote: > > 11.x: > o 11.2-STABLE armv6 BANANAPI > o 11.2-STABLE armv6 BEAGLEBONE > o 11.2-STABLE armv6 CUBIEBOARD > o 11.2-STABLE armv6 CUBIEBOARD2 > o 11.2-STABLE armv6 CUBOX-HUMMINGBOARD > o 11.2-STABLE armv6 RPI-B > o 11.2-STABLE armv6 RPI2 > o 11.2-STABLE armv6 PANDABOARD > o 11.2-STABLE armv6 WANDBOARD > > 12.x+ (I got the list from a 13.0 snapshot announcement): > o 13.0-CURRENT armv6 RPI-B > o 13.0-CURRENT armv7 BANANAPI > o 13.0-CURRENT armv7 BEAGLEBONE > o 13.0-CURRENT armv7 CUBIEBOARD > o 13.0-CURRENT armv7 CUBIEBOARD2 > o 13.0-CURRENT armv7 CUBOX-HUMMINGBOARD > o 13.0-CURRENT armv7 RPI2 > o 13.0-CURRENT armv7 PANDABOARD > o 13.0-CURRENT armv7 WANDBOARD > o 13.0-CURRENT armv7 GENERICSD > > So as of 12.x+ most are armv7 --as are most new ones > expected to be. > > As stands, in my amd64 -> armv7 13.0 cross-build activity, > uname -p and the like under the chroot context are > returning armv6 instead of armv7 unless I override via > a UNAME_p definition. > > This appears to trace back to: bsd-user/arm/target_syscall.h > and its: > > #define TARGET_HW_MACHINE "arm" > #define TARGET_HW_MACHINE_ARCH "armv6" > > and lack context sensitivity, such as to the FreeBSD version > that it is in use under. > Indeed, I opened this a couple of hours ago: https://github.com/seanbruno/qemu-bsd-user/pull/70 -- It turns out this is basically wrong, though I'm not sure immediately how to rectify. I don't think we can reasonably decide at compile-time what this should look like since all 32-bit ARM are shoved into this one target, so perhaps the right answer is that armv6 and armv7 need to split off from arm.arm and we use a check like the one in the above PR. CC'ing imp for a wisdom drop.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CACNAnaGesvCJp=eip3319BQwXixwyaYFLmZMGpisnQjp8w8_XA>