Additional notes about details of switching, at least in my powerpc64 context:
Looks like root@FBSDG5M1:/usr/ports/x11-drivers/xorg-drivers # make config does not list a selection for x11-drivers/xf86-video-ati-ums : no mention of UMS/ums and the only one ATI/ati line is present to select/deselect. That selection is what creates the dependency that indirectly prevents more overall builds that do not explicitly list x11-drivers/xf86-video-ati-ums . So ATI needs to be disabled there and x11-drivers/xf86-video-ati-ums needs to be separately/explicitly portmaster'd. And pkg delete xf86-video-ati-6.14.6_4 or analogous is likely required first if one had such prior history already in place. Since I use a ports_origins file overall this adds (at least) x11-drivers/xf86-video-ati-ums to the file. My updated ports_origins ended up as shown by: root@FBSDG5M1:~/fbsd_rebuild_materials # more ports_origins graphics/ImageMagick textproc/asciidoc devel/automake x11-fonts/bdftopcf x11/bigreqsproto devel/bison sysutils/bsdadminscripts devel/cmake ports-mgmt/dialog4ports x11/dri3proto x11/evieext devel/gdb devel/git x11/glproto devel/gmake devel/gmake-lite archivers/gtar misc/help2man x11-themes/icon-naming-utils textproc/intltool devel/libatomic_ops devel/libcheck x11/libgnome devel/libtool devel/makedepend ports-mgmt/pkg ports-mgmt/portmaster x11/presentproto textproc/py-sphinx x11/resourceproto devel/scons security/sudo archivers/unzip multimedia/v4l_compat lang/vala x11/xcb-proto x11/xcmiscproto x11-drivers/xf86-video-ati-ums x11-drivers/xf86-video-scfb x11-fonts/xf86bigfontproto x11/xf86driproto x11-wm/xfce4 textproc/xmlto x11/xorg devel/xorg-macros archivers/zip If my occasional PowerMac G5 experiments with Jean-Sébastien Pédron's kms-drm-update-38 branch when Jean requests get afar enough along I will have to locally patch back in allowing powerpc64 builds of x11-drivers/xf86-video-ati . In early November we did get to the point of kldload for radeonkms getting no failure messages/rejections. But for the Radeon X1950 that I have for PowerMac G5's (PCIExpress) no video bios could be found and the video was then munged up. (The card works in Mac OS X 10.5.) The Radeon type X1950 may be odder than usual but it is the only Radeon for powerpc64 that I have access to and so is the only Radeon that I can indirectly give Jean access to for the PowerMac G5 (PCIExpress) context. > === Mark Millard markmi at dsl-only.net Context note: I build a powerpc64/GENERIC64 variant with both vt and sc enabled and use the same SSD for both NIVIDA based PowerMac G5's and one Radeon X1950 based one. Currently I do not use X11 at all for the Radeon X1950 but the SSD is designed to also cover NVIDA contexts where I do sometimes use X11/xfce4. Looks like the earlier change to force 7.2.0 created the issue in the Makefile and probably should have had a UPDATING notice: Revision 372171 - (view) (download) (annotate) - [select for diffs] Modified Tue Nov 4 20:32:43 2014 UTC (2 weeks, 5 days ago) by dumbbell File length: 455 byte(s) Diff to previous 371270 xf86-video-ati: Update to 7.5.0 This update brings several bug fixes. Furthermore, it improves Glamor support, though this feature isn't supported by FreeBSD yet. Differential Revision: https://reviews.freebsd.org/D1106 Tested by: Many on freebsd-x11@ Reviewed by: bapt@, kwm@ Approved by: bapt@ Revision 371270 - (view) (download) (annotate) - [select for diffs] Modified Mon Oct 20 14:27:59 2014 UTC (4 weeks, 6 days ago) by kwm File length: 442 byte(s) Diff to previous 367997 Remove the switch code between 6.14.6 and 7.2.0. Update to 7.2.0, since we have xf86-video-ati-ums for 6.14.6. For example compared to 367997: ONLY_FOR_ARCHS= i386 amd64 Note that I did not originally explicitly list xf86-video-ati as to be updated when I rebuilt: the automatic dependencies are what brought it in. So as stands far more than ati is blocked by default for powerpc64 rebuilds since the problem stops builds of other things explicitly requested when dependencies automatically also try to also build xf86-video-ati on powerpc64. I rebuilt with... portmaster -tDK --no-confirm `cat ports_origins` My ports_origins is as shown by (the explicit xf86-video-scfb is from when Nathan W. suggested trying it for something but it did not work): root@FBSDG5M1:~/fbsd_rebuild_materials # more ports_origins graphics/ImageMagick devel/automake x11-fonts/bdftopcf x11/bigreqsproto devel/bison sysutils/bsdadminscripts devel/cmake ports-mgmt/dialog4ports x11/dri3proto x11/evieext devel/gdb x11/glproto devel/gmake devel/gmake-lite misc/help2man x11-themes/icon-naming-utils textproc/intltool devel/libatomic_ops devel/libcheck x11/libgnome devel/libtool devel/makedepend ports-mgmt/pkg ports-mgmt/portmaster x11/presentproto textproc/py-sphinx x11/resourceproto security/sudo archivers/unzip multimedia/v4l_compat lang/vala x11/xcb-proto x11/xcmiscproto x11-drivers/xf86-video-scfb x11-fonts/xf86bigfontproto x11/xf86driproto x11-wm/xfce4 textproc/xmlto x11/xorg devel/xorg-macros archivers/zip === Mark Millard markmi at dsl-only.net On 23-11-2014 21:49, Mark Millard wrote: > I do not see anything about it in /usr/ports/UPDATING but when I attempted to > update my powerpc64/GENERIC64vtsc context's ports after updating to > 10.1-RELEASE I got: I think radeonKMS doesn't work/tested on anything other then i386 or amd64. I think you should switch to the x11-drivers/xf86-video-ati-ums port. > ===>>> Launching child to update xf86-video-ati-6.14.6_4 to > xf86-video-ati-7.5.0 > > ===>>> x11/xorg 1/2 >> xf86-video-ati-6.14.6_4 (1/3) > > ===>>> Currently installed version: xf86-video-ati-6.14.6_4 > ===>>> Port directory: /usr/ports/x11-drivers/xf86-video-ati > > ===>>> Starting check for all dependencies > ===>>> Gathering dependency list for x11-drivers/xf86-video-ati from ports > ===>>> Dependency check complete for x11-drivers/xf86-video-ati > > ===>>> x11/xorg 1/2 >> xf86-video-ati-6.14.6_4 (1/3) > > ===> Cleaning for xf86-video-ati-7.5.0 > ===> xf86-video-ati-7.5.0 is only for i386 amd64, while you are running > powerpc64. > *** Error code 1 > > Stop. > make: stopped in /usr/ports/x11-drivers/xf86-video-ati > > > The context is: > > root@FBSDG5M1:/usr/src # freebsd-version -ku; uname -a > 10.1-RELEASE > 10.1-RELEASE > FreeBSD FBSDG5M1 10.1-RELEASE FreeBSD 10.1-RELEASE #0 r274912M: Sun Nov 23 > 03:21:22 PST 2014 root@FBSDG5M1:/usr/obj/usr/src/sys/GENERIC64vtsc > powerpc > > (Note: when one boots a PowerMac with an explicit "boot KERNELNAME" > freebsd-version -k returns /boot/kernel information instead of > /boot/KERNELNAME information. But I did not boot that way: This was a > /boot/kernel use.) > > root@FBSDG5M1:/usr/ports # svnlite info > Path: . > Working Copy Root Path: /usr/ports > URL: https://svn0.us-west.freebsd.org/ports/head > Relative URL: ^/head > Repository Root: https://svn0.us-west.freebsd.org/ports > Repository UUID: 35697150-7ecd-e111-bb59-0022644237b5 > Revision: 373147 > Node Kind: directory > Schedule: normal > Last Changed Author: crees > Last Changed Rev: 373147 > Last Changed Date: 2014-11-23 05:22:05 -0800 (Sun, 23 Nov 2014) > > root@FBSDG5M1:/usr/ports # svnlite status > ? .snap > ? restoresymtable > > root@FBSDG5M1:/usr/ports # more /etc/make.conf > WRKDIRPREFIX=/usr/obj/portswork > > root@FBSDG5M1:/usr/src # svnlite status > ? .snap > ? restoresymtable > M sys/ddb/db_main.c > M sys/ddb/db_script.c > M sys/powerpc/ofw/ofw_machdep.c > M sys/powerpc/ofw/ofwcall64.S > > root@FBSDG5M1:/usr/src # more /etc/src.conf > WITH_DEBUG_FILES= > WITHOUT_CLANG= > WITH_DEBUG= > > (WITHOUT_CLANG is just because the last time I tried WITH_DEBUG_FILES with a > clang build involved it failed. It has been some time since I retried that.) > > root@FBSDG5M1:/usr/src # more sys/powerpc/conf/GENERIC64vtsc > include GENERIC64 > ident GENERIC64vtsc > > nooptions PS3 #Sony Playstation 3 > HACK!!! to allow sc > > options DDB # HACK!!! to dump early crash info > options GDB # HACK!!! ... > #options KTR > #options KTR_MASK=KTR_TRAP > #options KTR_CPUMASK=0xF > #options KTR_VERBOSE > > # HACK!!! to allow sc for 2560x1440 display on Radeon X1950 that vt mishandled > device sc > #device kbdmux # HACK: already listed by vt > options SC_OFWFB # OFW frame buffer > options SC_DFLT_FONT # compile font in > makeoptions SC_DFLT_FONT=cp437 > > > === > Mark Millard > markmi at dsl-only.net > > _______________________________________________ > freebsd-ports@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-ports > To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org" _______________________________________________ freebsd-ports@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-ports To unsubscribe, send any mail to "freebsd-ports-unsubscr...@freebsd.org"
