Re: pkg_rolling-replace and mismatched version
Hi, Greg Troxel wrote: I am pretty convinced that something is wrong with your system. just for the record. I had an obsolete |PKG_DBDIR| in |/etc/pkg_install.conf| set to /var/db/pkg That was probably the root of evil, of a half-assed transition - by miracle it worked until I tried to upgrade. Let's start from scratch! Riccardo
Re: pkg_rolling-replace and mismatched version
Hi, Greg Troxel wrote: Riccardo Mottola writes: sudo pkg_add -v /usr/packages/All/perl-5.38.2.tgz pkg_add: no pkg found for '/usr/packages/All/perl-5.38.2.tgz', sorry. pkg_add: 1 package addition failed -rw-r--r-- 1 root wheel 18M May 17 18:32 perl-5.38.2.tgz run "tar tfvz" on it and see what's inside. The archive is corrupted... not a valid tar file. That's not good I wonder if these SS4 and SS20 have stability issues. I removed the binary file, it will be rebuilt. Check that you have only one PKG_DBDIR (/usr/pkg/pkgdb, and not /var/db/pkg). I am pretty convinced that something is wrong with your system. I have both. Strange. I migrated all by boxen back then. I did some scary now: sudo rm -rf /var/db/pkg* sudo rm -rf /usr/pkg and I hope this is enough to start over make-installing each package from scratch. I will also discard binary packages created on this system for safety, since it could be inconsistent. Riccardo
Re: framebuffer console on old ATI
Hi Michael Michael wrote: did so. Still get a black screen. Before going black, I am able to see a "failed". Also, it hangs, because if I wait some time I can't login remotely (left dhcp on). This is odd, but it's a data point at least. Before trying to recompile the kernel once again, I tried external monitor with RADEON_BIOS_INIT It doesn't hang. I get some garbage blue pixels in the first lines of the screen, but e.g. caps-lock works! Powering down did not, so the machine did not complete a boot, but is not totally dead. This is different than without screen - there it is "dead". Does it hang without RADEONFB_BIOS_INIT as well? I recompiled a radeonfb* kernel again without init. Internal screen: freeze, no caps-lock action External screen: blue garbage on top of the screen, caps-lock active. Same behaviour as with init. If it doesn't and you just get no useful output - does the laptop in question have a VGA or DVI output? If it does, could you plug something into it? Yes it has classic VGA and I reported above how it behaves. The reason I'm asking is, most/maybe all radeons have at least two outputs, how they're wired to connectors, LCDs etc. is up to the vendors. Normally there is a table in the video BIOS telling us exactly what is connected to what. That's what RADEONFB_BIOS_INIT tries to find. If we get it wrong we may very well end up with a black display. On macppc we don't have a connector table but pretty much all *Books wire up the internal display in exactly the same way, there are a few exceptions but we can distinguish those from the model string in OF. Thats a possible explanation. Also it probably misdetects the external screen size The other possible cause would be some endianness screwup, radeonfb has been used mostly on macppc and sparc64, and not much else. IIRC some brave soul made it work on alpha a while ago, but it got pretty much zero testing on x86. That could be an explanation, if someone can check if it still works on alpha, we would have endianness information. So the machine probably doesn't have a video BIOS, just like most Apple *Books. Strange.. .it is a fairly standard Pentium4M PC which did run Windows XP... Does Xorg run with a radeon driver on any OS at all? If so, could you send me the xorg.0.log from such a thing? That should give at least a hint on what we get wrong. Xorg runs with NetBSD and default kernel. Log attached. It is only very very slow in startup, like blank screen with white cursor on top and for several seconds, then garbage... then mouse pointer. also xrand doesn't display a second monitor if attached.. never noticed since I didn't try, but since I was testing, I tried. I remember xrand displaying LVDS, VGA and such on other systems, here not. "default" and no monitor. Internal monitor is 1024x768 classic [ 646.081] (II) VESA(0): initializing int10 [ 646.082] (II) VESA(0): Primary V_BIOS segment is: 0xc000 [ 646.096] (II) VESA(0): VESA BIOS detected [ 646.096] (II) VESA(0): VESA VBE Version 2.0 [ 646.096] (II) VESA(0): VESA VBE Total Mem: 16320 kB [ 646.096] (II) VESA(0): VESA VBE OEM: ATI MOBILITY RADEON 7500 [ 646.096] (II) VESA(0): VESA VBE OEM Software Rev: 1.0 [ 646.096] (II) VESA(0): VESA VBE OEM Vendor: ATI Technologies Inc. [ 646.096] (II) VESA(0): VESA VBE OEM Product: P7 [ 646.096] (II) VESA(0): VESA VBE OEM Product Rev: 01.00 [ 646.109] (II) VESA(0): virtual address = 0xaf80, VGAbase = 0xaf7f physical address = 0xe800, size = 16711680 [ 646.321] (II) VESA(0): Setting up VESA Mode 0x123 (1024x768) [ 648.893] (==) VESA(0): Default visual is TrueColor [ 649.210] (==) VESA(0): Backing store enabled [ 649.292] (==) VESA(0): DPMS enabled Riccardo [ 591.666] X.Org X Server 1.21.1.9 X Protocol Version 11, Revision 0 [ 591.667] Current Operating System: NetBSD orthanc.westernesse.arda 10.0 NetBSD 10.0 (GENERIC) #0: Thu Mar 28 08:33:33 UTC 2024 mkre...@mkrepro.netbsd.org:/usr/src/sys/arch/i386/compile/GENERIC i386 [ 591.667] [ 591.775] Current version of pixman: 0.38.4 [ 591.775] Before reporting problems, check http://wiki.x.org to make sure that you have the latest version. [ 591.775] Markers: (--) probed, (**) from config file, (==) default setting, (++) from command line, (!!) notice, (II) informational, (WW) warning, (EE) error, (NI) not implemented, (??) unknown. [ 592.218] (==) Log file: "/var/log/Xorg.0.log", Time: Tue May 7 00:59:32 2024 [ 593.000] (II) Loader magic: 0xe59c80 [ 593.001] (II) Module ABI versions: [ 593.001] X.Org ANSI C Emulation: 0.4 [ 593.001] X.Org Video Driver: 25.2 [ 593.001] X.Org XInput driver : 24.4 [ 593.001] X.Org Server Extension : 10.0 [ 593.436] (--) PCI:*(1@1:0:0) 1002:4c57:1014:0517 rev 0, Mem @ 0xe800/134217728, 0xd010/65536, I/O @ 0x3000/256 [ 593.638] (==) Using default built-in configuration (39 lines) [ 593.638] (==) --- Start of built-in
Re: strange issues with building ArcticFox - full rebuilds everytime
Hi, RVP wrote: My first guess is that configure is re-run unneeded, leading thus to a major recompile. You could make the 2nd & 3rd lines of configure this and confirm: echo >&2 dying... exit 1 I actually executed in an xterm, interrupted after a couple of minutes and checked the scrollback - configure is rerun. This does not happen on linux or FreeBSD. Just as an idea, I checked dates: -rwxr-xr-x 1 multix users 221 Jun 6 01:24 configure -rw-r--r-- 1 multix users 1154 Mar 5 01:29 configure.in -rw-r--r-- 1 multix users 873016 Jun 6 01:25 old-configure -rw-r--r-- 1 multix users 253669 Jun 6 01:19 old-configure.in the ".in" files are suspiciously new, as in the previous run. But I didn't update them. Comparing to linux, I see they are not touched since the last commit to them: -rwxr-xr-x 1 multix users 221 May 29 16:08 configure -rw-r--r-- 1 multix users 1154 Feb 19 11:24 configure.in -rw-r--r-- 1 multix users 874090 May 29 16:08 old-configure -rw-r--r-- 1 multix users 253669 May 29 15:30 old-configure.in the checked out repository is the same... Just to be sure, I yanked the build dir and retried. no avail. Riccardo
Re: pkg_rolling-replace and mismatched version
Hi! replying to myself. I wanted to run lintpkgsrc too I think maybe perl is missing? I have my locally build binary package. However, it doesn't install. sudo pkg_add -v /usr/packages/All/perl-5.38.2.tgz pkg_add: no pkg found for '/usr/packages/All/perl-5.38.2.tgz', sorry. pkg_add: 1 package addition failed -rw-r--r-- 1 root wheel 18M May 17 18:32 perl-5.38.2.tgz Mystery is that lintpkgsrc did check: => Full dependency perl>=5.36: found perl-5.36.0 But pkg_info told me it wasn't there.. how can this check pass? eowyn$ perl -sh: perl: not found I thus got perl from the prebuild binaries and installed it. Now lintpgsrc is running... it will take some time on this oldie. But I wonder what's going on. Riccardo
Re: pkg_rolling-replace and mismatched version
Hi Greg! thanks for replying... Greg Troxel wrote: It sounds like you are not having a pkg_rolling-replace problem, in that the underlying make replace does not work. Please see the pkg_rolling-replace man page where it asks that problems with a make replace (that was ordered reasonably) not be reported as pkg_rolling-replace issues. I didn't mean it is a pkg_rolling-replace issue, it is just started from that. If make replace itself fails, the problem is underlying. In this case it appears to do nothing. I thought the order of packages rebuilt could be wrong. I thought that perhaps the pkg tools need to be updated first? so I tried "make replace" in pkg_install => Automatic manual page handling => Generating post-install file lists => Checking file-check results for pkg_install-20240307 => Creating binary package /usr/pkg-workdir/pkgtools/pkg_install/work/.packages/pkg_install-20240307.tgz ===> Building binary package for pkg_install-20240307 => Creating binary package /usr/packages/All/pkg_install-20240307.tgz ===> Replacing for pkg_install-20240307 *** Error code 1 Stop. make[1]: stopped in /usr/pkgsrc/pkgtools/pkg_install *** Error code 1 Stop. make: stopped in /usr/pkgsrc/pkgtools/pkg_install this is for certain not a very helpful error! I would suggest checking your pkgdb (that you only have one, rebuild, rebuild-tree, check). I did build & rebuild, but since I manually removed some packages to ease the burden of upgrade, I reissued it. Didn't think to run check - good point. pkg_admin check finds issues: e.g. pkg_admin: x11-links-1.35: Symlink `/usr/pkg/share/x11-links/include/xorg/BT.h' exists and is in +CONTENTS but target does not exist! pkg_admin: x11-links-1.35: Symlink `/usr/pkg/share/x11-links/include/xorg/IBM.h' exists and is in +CONTENTS but target does not exist! but there are several failed checksums, e.g.: ../usr/pkg/lib/libffi.a fails MD5 checksum ./usr/pkg/bin/libtool fails MD5 checksum ./usr/pkg/lib/libpkgconf.a fails MD5 checksum ./usr/pkg/lib/libhistory.a fails MD5 checksum /usr/pkg/lib/libreadline.a fails MD5 checksum I rebuilt libtool and readline manually, they replaced fine! But still I cannot replace pkg_install Riccardo
Re: framebuffer console on old ATI
Hi John & Michael, since I borrowed the monitor of my workstation to do tests for radeonfb I also thougt to take the time and test what the radeondrm would do with an external monitor! set bool radeon_pci_ignore_r100_r200 = false; and recompile (commenting out radeonfb* just not to mix things up! At boot... I get a framebuffer high resolution console with a nice font! It is "correct" in size for the external 24" display and just displays a part clipped on the internal 13" display. Progress! Clearly, there is some display misdetection... it did not "switch" completely.. kind of bad mirroring, but with the resolution of the external one. It should have used internal resolution for mirror or drive only the external monitor. Monitor setup in BIOS is set to use VGA as default and that works for the BIOS and bootmanager. Try more? X11! Doesn't work. I get "ordered" garbage on the screen. Since the system is not stalled, I can telnet in and check dmesg. These appear at regular intevals [ 216.997284] radeon0: autoconfiguration error: error: ring 0 stalled for more than 82140msec [ 216.997284] radeon0: warn: GPU lockup (current fence id 0x0007 last fence id 0x002d on ring 0) [ 217.497291] radeon0: autoconfiguration error: error: ring 0 stalled for more than 82640msec [ 217.497291] radeon0: warn: GPU lockup (current fence id 0x0007 last fence id 0x002d on ring 0) [ 217.997299] radeon0: autoconfiguration error: error: ring 0 stalled for more than 83140msec [ 217.997299] radeon0: warn: GPU lockup (current fence id 0x0007 last fence id 0x002d on ring 0) out of test, I tried then rebooting again this radeon test kernl without external display, I get a perfect working internal console. It spits out this information [ 5.304228] [drm] initializing kernel modesetting (RV200 0x1002:0x4C57 0x1014:0x0517 0x00). [ 5.304228] [drm] register mmio base: 0xd010 [ 5.304228] [drm] register mmio size: 65536 [ 5.314229] radeon0: GTT: 64M 0xE000 - 0xE3FF [ 5.314229] radeon0: VRAM: 128M 0xE800 - 0xEFFF (16M used) [ 5.314229] [drm] Detected VRAM RAM=80M, BAR=128M [ 5.314229] [drm] RAM width 64bits DDR [ 5.314229] Zone kernel: Available graphics memory: 535062 KiB [ 5.314229] [drm] radeon: 16M of VRAM memory ready [ 5.314229] [drm] radeon: 64M of GTT memory ready. [ 5.314229] radeon0: WB disabled [ 5.324301] radeon0: fence driver on ring 0 use gpu addr 0xe000 and cpu addr 0x0xdb9a1000 [ 5.324301] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013). [ 5.324301] [drm] Driver supports precise vblank timestamp query. [ 5.324301] radeon0: interrupting at irq 11 (radeon0) [ 5.324301] [drm] radeon: irq initialized. [ 5.324301] [drm] Loading R100 Microcode [ 5.394230] [drm] radeon: ring at 0xE0001000 [ 5.394230] [drm] ring test succeeded in 1 usecs [ 5.404243] [drm] ib test succeeded in 0 usecs [ 5.404243] [drm] Panel ID String: 1024x768 [ 5.404243] [drm] Panel Size 1024x768 [ 5.404243] [drm] No TV DAC info found in BIOS [ 5.404243] [drm] Radeon Display Connectors [ 5.404243] [drm] Connector 0: [ 5.404243] [drm] VGA-1 [ 5.404243] [drm] DDC: 0x60 0x60 0x60 0x60 0x60 0x60 0x60 0x60 [ 5.404243] [drm] Encoders: [ 5.404243] [drm] CRT1: INTERNAL_DAC1 [ 5.404243] [drm] Connector 1: [ 5.404243] [drm] DVI-D-1 [ 5.404243] [drm] HPD1 [ 5.404243] [drm] DDC: 0x64 0x64 0x64 0x64 0x64 0x64 0x64 0x64 [ 5.404243] [drm] Encoders: [ 5.404243] [drm] DFP1: INTERNAL_TMDS1 [ 5.404243] [drm] Connector 2: [ 5.404243] [drm] LVDS-1 [ 5.404243] [drm] Encoders: [ 5.404243] [drm] LCD1: INTERNAL_LVDS [ 5.404243] [drm] Connector 3: [ 5.404243] [drm] SVIDEO-1 [ 5.404243] [drm] Encoders: [ 5.404243] [drm] TV1: INTERNAL_DAC2 [ 5.444234] radeondrmkmsfb0 at radeon0 [ 5.444234] [drm] Initialized radeon 2.50.0 20080528 for radeon0 on minor 0 [ 5.444234] radeondrmkmsfb0: framebuffer at 0xe804, size 1024x768, depth 16, stride 2048 [ 5.544232] wsdisplay0 at radeondrmkmsfb0 kbdmux 1: console (default, vt100 emulation), using wskbd0 what changed ot my first test? I don't know.. except I took the care to disable nouveau... but that shouldn't matter Interestingly here we see 4 connectors... but I only have VGA and SVIDEO as ports, so I guess DVI is internal? no HDMI/DVI port available and we don't know what is being used. X11 with internal video? it comes up fairly quickly (compared to the standard kernel, where I report that it does work, but takes a long time with a black screen, some garbage...) and I get a working mousr cursor on a black backround, nothing else telnet in and dmesg shows: [ 253.288036] radeon0: warn: GPU lockup (current fence id
pkg_rolling-replace and mismatched version
Hi, I am running pkg_rolling-replace -uv on my SparcStation 4 which has been upgraded from 9.3 to 10.0 I save binary packages, since I have two SS, I can share the compilation effort. I run, the package builds but then fails to install. It happened to me twice. E.g.: RR> Selecting gmake (devel/gmake) as next package to replace RR> Replacing gmake-4.4 ===> Cleaning for gmake-4.4.1 => Bootstrap dependency digest>=20211023: found digest-20220214 => Checksum BLAKE2s OK for make-4.4.1.tar.gz => Checksum SHA512 OK for make-4.4.1.tar.gz ===> Installing dependencies for gmake-4.4.1 <.> /bin/ln -sf /usr/pkg/man/man1/gmake.1 /usr/pkg-workdir/devel/gmake/work/.destdir/usr/pkg/gnu/man/man1/make.1 => Automatic manual page handling => Generating post-install file lists => Checking file-check results for gmake-4.4.1 => Creating binary package /usr/pkg-workdir/devel/gmake/work/.packages/gmake-4.4.1.tgz ===> Building binary package for gmake-4.4.1 => Creating binary package /usr/packages/All/gmake-4.4.1.tgz ===> Replacing for gmake-4.4.1 ===> Updating using binary package of gmake-4.4.1 /usr/bin/env /usr/pkg/sbin/pkg_add -K /usr/pkg/pkgdb -U -D /usr/pkg-workdir/devel/gmake/work/.packages/gmake-4.4.1.tgz *** package gmake is still a mismatched version. *** Please read the errors listed above, fix the problem, *** then re-run pkg_rolling-replace to continue. the package was freshly built and I checked, I had 4.4 and jump to 4.4.1 and it was detected correct.y If I go into /usr/pkgsrc/devel/gmake and type "make replace" It will quickly install the binary package without errors but it did nothing! Before: gmake-4.4 GNU version of 'make' utility After: gmake-4.4 GNU version of 'make' utility Build package: /usr/packages/All/gmake-4.4.1.tgz I'm confused... Riccardo
strange issues with building ArcticFox - full rebuilds everytime
Hi, I build ArcticFox from source - I am the main developer :) I noticed a strange behavior on NetBSD not present on other platforms. "mach build" does a full rebuild even if nothing changed. This doesn't happen on Linux. I noticed this 100% on NetBSD/sparc64 and NetBSD/ppc. On these system, being slow / very slow, I notice it... triggering up to 20hrs of build! I don't know if it happens always on amd64, I thought not, but I just tried and it happened. I run 9.4 and 10 on the systems, I didn't notice it when running 9.3 (but ArcticFox build system changed too). My first guess is that configure is re-run unneeded, leading thus to a major recompile. Any ideas? Riccardo
Re: framebuffer console on old ATI
Hi, John D. Baker wrote: Interesting that this works for you. I have several machines with the ATI ES1000 rev. 0x02 (RN50) video device: HP DL380G5, HP ML310G4, IBM x3650, Dell PowerEdge 2850 and they all exhibit the "almost-black-on-black" video problem when booting a kernel built with the patch applied and radeondrmkms, etc. enabled. when you have black on black, the computer is still functioning correctly... e.g. it responds to keyboard (caps-lock) and most important it responds to network and you can telnet/ssh into it. Mine does not.. this is what worries me, I think something else is misconfigured or misinitialized. Riccardo
Re: framebuffer console on old ATI
Hi Martin, Martin Neitzel wrote: A zillion thanks to John Baker for mailing this! I changed this on my old HP Proliant DL360 G6 with its ATI ES1000 (RV100) onboard graphics two weeks ago and --voila-- instant karma! I envy you! I have a newer chip but it still doesn't like to run :) Riccardo
Re: intel videocard performance degraded severely after 9.4 update
Sorry for the multiple-posts.. apparently SeaMonkey sent out drafts while I was composing the message! Very strange.
Re: framebuffer console on old ATI
Hi John sorry for the late reply... needed some NetBSD hack time to find to patch and recompile the kernel. John D. Baker wrote: > On Tue, 7 May 2024, Riccardo Mottola wrote: > > That looks like the same description of the radeon device in my T42 > ThinkPad: IN some sense the T43 is successor of the T30... I'd expect yours to be a little bit newer. But all these R/RV/RS revisions are confusing. > (radeondrm0 is the legacy UMS "radeondrm* at drm?" device) > > As an R*200 device, it is not matched by the DRM2 code in NetBSD. > >> Since I am endeavouring in compiling a kernel, I might try enabling it, >> if you tell me how. Just as a test, maybe magic happens. > > The following patch (against -current from about 4 years ago) will > enable matching R100/R200 devices: > > > +Index: sys/external/bsd/drm2/radeon/radeon_pci.c > +=== > +RCS file: /cvsroot/src/sys/external/bsd/drm2/radeon/radeon_pci.c,v > +retrieving revision 1.14 > +diff -u -p -r1.14 radeon_pci.c > +--- sys/external/bsd/drm2/radeon/radeon_pci.c24 Jan 2020 11:44:27 > - 1.14 > sys/external/bsd/drm2/radeon/radeon_pci.c28 May 2020 23:49:44 > - > +@@ -126,7 +126,7 @@ extern const struct pci_device_id *const > + extern const size_t radeon_n_device_ids; > + > + /* Set this to false if you want to match R100/R200 */ > +-bool radeon_pci_ignore_r100_r200 = true; > ++bool radeon_pci_ignore_r100_r200 = false; > + > + static bool > + radeon_pci_lookup(const struct pci_attach_args *pa, unsigned long *flags) > This patch was a little hard to read, looks like a patch of a patch? But the core is to swap true to false. in a match. I did... However, it does not work properly. During boot, when FB should attach, the whole scren displays garbage. It is not hi-res garbage, the screen still looks like being in VGA resolution and contains just blocks of green and white stripes (as of the console dmesg). So indeed the code does not seem to handle my RS250 without further tweaks! > > I really don't know. My understanding is that on OpenBSD and various > Penguin-OSes, these older ATI devices work fine. Indeed and we shouldn't be "of less" :) :) > I saw your message about trying "radeonfb* at pci?" in your kernel config. For that I am just trying Michael's suggestions, after having enabled again radeon_pci_ignore_r100_r200 Will report there. For most users it would be probably to have drm working... but let's start getting a framebuffer at all. Riccardo
Re: framebuffer console on old ATI
Hi, Michael wrote: > Did it crash or did if continue booting? If it continued to boot > radeonfb likely got the wiring of outputs to displays wrong. > With that, please try with options RADEONFB_BIOS_INIT, which adds code > to scan the BIOS for output info. This is not enabled by default > because most hardware we run radeonfb on doesn't have a PC BIOS. Should > probably be on by default on x86. did so. Still get a black screen. Before going black, I am able to see a "failed". Also, it hangs, because if I wait some time I can't login remotely (left dhcp on). thanks to easy videos with iPhones... I was able to extract a frame: radeonfb0: BIOS initialization failed hmm... looks like the option was not enough, I did not do it correctly or there is a further bug. Riccardo
Re: intel videocard performance degraded severely after 9.4 update
Hi, trying and testing... I can get even more scarier errors from i915. I tried various combination of enabling wifi before/after As you can see, the buffer underrun appears after some moments, after I completed login. So we have 1) CPU Pipe A FIFO underrun 2) CPU Pipe B FIFO underrun 3) failed to load DMA map for stolen object: -22 4) failed to load DMA map for stolen object: -22 again 5) after a while CPU Pipe A FIFO underrun again are there other old i915 owners? [ 5.327841] kern info: [drm] Memory usable by graphics device = 128M [ 5.337843] kern info: [drm] Supports vblank timestamp caching Rev 2 (21.10.2013). [ 5.337843] kern info: [drm] Driver supports precise vblank timestamp query. [ 5.337843] i915drmkms0: interrupting at irq 11 (i915drmkms0) [ 5.367842] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/i915/intel_fifo_underrun.c:363)intel_cpu_fifo_underrun_irq_handler] *ERROR* CPU pipe A FIFO underrun [ 5.367842] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/i915/intel_fifo_underrun.c:363)intel_cpu_fifo_underrun_irq_handler] *ERROR* CPU pipe B FIFO underrun [ 5.377843] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/i915/i915_gem_stolen.c:555)i915_pages_create_for_stolen] *ERROR* failed to load DMA map for stolen object: -22 [ 5.377843] kern info: [drm] initialized overlay support [ 5.427842] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/i915/i915_gem_stolen.c:555)i915_pages_create_for_stolen] *ERROR* failed to load DMA map for stolen object: -22 [ 5.437847] intelfb0 at i915drmkms0 [ 5.437847] intelfb0: framebuffer at 0xdc295000, size 1024x768, depth 32, stride 4096 [ 6.067853] wsdisplay0 at intelfb0 kbdmux 1: console (default, vt100 emulation), using wskbd0 [ 6.078116] wsmux1: connecting to wsdisplay0 [10.377920] wsdisplay0: screen 1 added (default, vt100 emulation) [10.387923] wsdisplay0: screen 2 added (default, vt100 emulation) [10.387923] wsdisplay0: screen 3 added (default, vt100 emulation) [10.387923] wsdisplay0: screen 4 added (default, vt100 emulation) [15.507997] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/i915/intel_fifo_underrun.c:363)intel_cpu_fifo_underrun_irq_handler] *ERROR* CPU pipe A FIFO underrun
Re: intel videocard performance degraded severely after 9.4 update
Hi, nia wrote: > What are the permissions on /dev/dri/*? > > Does your user have the ability to write to the devices? it should: R51$ ls -l /dev/dri* total 0 crw-rw 1 root wheel 180, 0 Aug 8 2018 card0 crw-rw 1 root wheel 180, 1 Aug 8 2018 card1 crw-rw 1 root wheel 180, 2 Aug 8 2018 card2 crw-rw 1 root wheel 180, 3 Aug 8 2018 card3 R51$ groups users wheel
Re: i386 - 10.0 gecko browsers
Hi Benny! Benny Siegert wrote: According to the latest bulk build data (https://releng.netbsd.org/bulktracker/build/724/www/), Firefox 123 failed to build, Firefox 115 also failed to build but Firefox 102 was OK. So in principle, there should be a firefox102 package. Oh! I will check... an up-to-date browser is needed, on systems where I have enough RAM... I see the is ArcticFox packaged. but unfortunately it is not enough for many sites... not even SeaMonkey is enough for github. The firefox115 and firefox error is the same: /pbulk/work/www/firefox115/work/build/dist/include/js/Utility.h:482:20: error: static assertion failed: over-aligned type is not supported by JS_DECLARE_NEW_METHODS 482 | alignof(T) <= alignof(max_align_t), \ | ~~~^~~ /pbulk/work/www/firefox115/work/firefox-115.10.0/js/src/vm/MallocProvider.h:234:3: note: in expansion of macro 'JS_DECLARE_NEW_METHODS' 234 | JS_DECLARE_NEW_METHODS(new_, pod_malloc, MOZ_ALWAYS_INLINE) Here I have no idea... I wonder if gentoo has some patches for 32bit. | ^~ For Seamonkey, there is also a failed assertion in the build log: In file included from /pbulk/work/www/seamonkey/work/build/ipc/chromium/Unified_cpp_ipc_chromium0.cpp:110: /pbulk/work/www/seamonkey/work/seamonkey-2.53.16/ipc/chromium/src/base/message_pump_libevent.cc: At global scope: /pbulk/work/www/seamonkey/work/seamonkey-2.53.16/ipc/chromium/src/base/message_pump_libevent.cc:31:40: error: static assertion failed: bad EVENT__SIZEOF_TIME_T 31 | static_assert(EVENT__SIZEOF_##TYPE == sizeof(type), \ /pbulk/work/www/seamonkey/work/seamonkey-2.53.16/ipc/chromium/src/base/message_pump_libevent.cc:43:1: note: in expansion of macro 'CHECK_EVENT_SIZEOF' 43 | CHECK_EVENT_SIZEOF(TIME_T,time_t); | ^~ I don't know how to fix these, but maybe firefox102 works for you? SM is at 2.53.18.2 now... wonder if that helps. However i think this is really a check of time_t size. That could be a 32bit issue. I did not try to build on linux 32bit. There is a mismanc of TIME_T sizes. I did build Firefox on Linux successfully. Riccardo
Re: where does pkg_admin come from?
Hi, Jeremy C. Reed wrote: I cannot update/replpace packages because make_replace complains with pkg_admin error, but maybe I can find the beginning of this vicious loop. Do you also have /usr/sbin/pkg_* ? Your/usr/pkg/ versions are from the pkg_install package. Maybe just remove that package, maybe with: /usr/sbin/pkg_delete pkg_install yes! the missing part is there -ff (double f!) because it is marked not for deletion. Now I started replacing, I'm back in business. I will run pkg_rolling-replace to see if the machine is stable. Major kernel upgrarde! I had issues with 9.3 on other systems, this is why I delayed the upgrade of this one. 8.2 was rock solid on the SS10. Riccardo
where does pkg_admin come from?
Hi! I upgraded from 8.2 to 9.4 To conserve space, I had to clean up obsolete libraries in /lib, by sparcstation is old and has a small disk) whenI try to upgrade packages, i get this error: /usr/pkg/sbin/pkg_admin: Shared object "libssl.so.12" not found we are now at /usr/lib/libssl.so.14.0 where does pkg_admin come from? a set that did not install or is outdated or some of the pkg ? I cannot update/replpace packages because make_replace complains with pkg_admin error, but maybe I can find the beginning of this vicious loop. Riccardo
Re: intel videocard performance degraded severely after 9.4 update
Hi, John McCue wrote: > > I have a R52e and the internal wifi does now work with the > router provided by my ISP (comcast). I believe a "Edimax > N150" dongle may work, but I have not tried it yet. Maybe you have a different card... IBM/Lenovo inserted "equivalent" cards (for windows) but it may not be iwi or the same revision. I believe on certain ThinkPads I inserted a compatible card coming from a Dell or HP system. Just doen't have IBM/Lenovo BIOS so it needs the a patch not to blame on 1802 error... > > I have a similar issue with X artifacts with NetBSD 10.0 amd64 > on my Thinkpad T420. It occurs when under a heavy load. > I created this PR for the issue: > > https://gnats.netbsd.org/cgi-bin/query-pr-single.pl?number=58217 > > I wonder if this is related ? Do you have intel video? >> In Xorg.0.log I see: >> [ 1561.785] (EE) intel(0): Failed to submit rendering commands (Invalid >> argument), disabling acceleration. >> >> that doesn't look promising! > > Yes it does, but I do not get that in my Xorg.0.log It is a bad error I think. I rebooted and started X, no issues. I did not configure network. I have the gut feeling that... network disturbs video! Being Intel WiFi and Intel Video... need to confirm this and maybe testing wired vs wireless. Riccardo
Re: NetBSD 10 and unreliable git
Hi, I think the situation is a little more complex and there may be some network card bug. Below my reasoning. Martin Husemann wrote: > I have never seen that (neither on 10.0 nor on -current). It is very > hard to tell what causes the errors for you. Two potential reasons > would be network issues (e.g. a bug in the driver for the network device > you are using) or the limit of simultaneous open files. - on core i3 latop, amd64, I had repeated several times on different WiFis... repeatable issue (checkout of gecko git). Attached wired and it worked. This points to be a network related issue, not open file or others. - on a vintage T43 with Pentium Mobile and troublesome iwi network card... the operation completed the first time plain over WiFi! And this system has half of the half performance of the above one... Riccardo
intel videocard performance degraded severely after 9.4 update
Hi, I updated my trusty R52 ThinkPad to 9.4 - i915. All went quite well. wifi is unreliable - but that is most probably the same isue I have on the T30 and I am already investigating that with Martin. X11 performance however is close to unusable, even a simple xterm has issues scrolling and there are artifacts in redrawing, sometimes garbage, something wrong colors in the title bar! With 9.3 it wasn't a a gaming platform, but quite usable for terminal and programming. I think that console switch sometimes had it loose acceleration, but it was never so bad and never had artifacts. wscons look and perform fine, the issue is X11. [ 5.106546] kern.module.path=/stand/i386/9.4/modules [ 5.106546] kern info: [drm] Memory usable by graphics device = 128M [ 5.106546] kern info: [drm] Supports vblank timestamp caching Rev 2 (21.10.2013). [ 5.106546] kern info: [drm] Driver supports precise vblank timestamp query. [ 5.106546] i915drmkms0: interrupting at irq 11 (i915drmkms0) [ 5.106546] WARNING: clock gained 4 days [ 5.126546] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/i915/intel_fifo_underrun.c:363)intel_cpu_fifo_underrun_irq_handler] *ERROR* CPU pipe A FIFO underrun [ 5.126546] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/i915/intel_fifo_underrun.c:363)intel_cpu_fifo_underrun_irq_handler] *ERROR* CPU pipe B FIFO underrun [ 5.136547] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/i915/i915_gem_stolen.c:555)i915_pages_create_for_stolen] *ERROR* failed to load DMA map for stolen object: -22 [ 5.136547] kern info: [drm] initialized overlay support [ 5.176547] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/i915/i915_gem_stolen.c:555)i915_pages_create_for_stolen] *ERROR* failed to load DMA map for stolen object: -22 [ 5.186566] intelfb0 at i915drmkms0 [ 5.186566] intelfb0: framebuffer at 0xdc295000, size 1024x768, depth 32, stride 4096 [ 5.816557] wsdisplay0 at intelfb0 kbdmux 1: console (default, vt100 emulation), using wskbd0 these errors look a bit scary, but I am pretty sure the fifo underrun was there before. However *ERROR* failed to load DMA map for stolen object: -22 is new. In Xorg.0.log I see: [ 1561.785] (EE) intel(0): Failed to submit rendering commands (Invalid argument), disabling acceleration. that doesn't look promising! Riccardo
Re: framebuffer console on old ATI
Hi, Michael wrote: >> How do I enable radeonfb? Can I do it through configurations or do I >> need to compile my own kernel to try? > radeonfb* at pci? > I copied that line from ALL to my custom kernel build which I did for iwi debug. I just get a big black screen... no text, no white cursor. Had to turn of via powerbutton. Riccardo
Re: framebuffer console on old ATI
Hi, John D. Baker wrote: > As you have seen, the "radeon*" and "radeondrmkmsfb*" drivers do not > attach on "r100" class devices. They have been explicitly excluded > because early in the DRMKMS integration in NetBSD, they exhibited the > "(almost) black-on-black" video problem described in: I remember seeing that issue.. long time ago, but don't remember on which OS and hardware combination. However, i checked a little better. ATI Technologies Radeon Mobility M7 LW (rev. 0x00) would be an RS250 chip, so a little bit newer. Does the same issue apply? or is it "blacklisted" for any other reason? Since I am endeavouring in compiling a kernel, I might try enabling it, if you tell me how. Just as a test, maybe magic happens. > > http://mail-index.netbsd.org/current-users/2015/02/20/msg026732 > > and later in the following PR: > > kern/49744: console is blank drm2/radeon > > The suggestion for "radeonfb*" is interesting. I see it's in the "ALL" > kernel, but not in "GENERIC" (not even as a commented-out device), so > you'd have to compile a custom kernel with it added to the config. That's what I wrote, I didn't see it, copying config from GENERIC. I also tried to compare with OpenBSD on another Laptop where framebwuffer runs wonderfully. It uses radeondrm0. I don't know how this in OpenBSD speak this compares to NetBSD. However, it detects a Mobility M9, which corresponds to RV250, which is newer... too much marketing speak there. RV200 appers to be just a shrink of RV100, a budget version of R100. But found no wiki about RS250. According to [1] it is just another Rage 6 variant, something for mobile. Riccardo [1] https://www.techpowerup.com/gpu-specs/ati-rs250.g384
Re: Intel Wireless fatal error
Hi, Martin Husemann wrote: > The message is bogus, it has nothing to do with autoconfiguration. Misleading.. I was wondering what autoconfiguration had to be done after everything was connected. > > "Fatal error" is a bit in the interrupt cause register of the intel > chipset. The driver can not do anything about it and resets the device > (simmilar to a "ifconfig iwi0 down). This is not a good error handling > strategy, as you have noticed - and it is also done very wrong in the > iwi_softintr handler. Not very good and it explains aslo why, if done quickly enough, ifconfig up restores network connectivity and even open connection. Clearly, if time passes, dhcp lease might expire, connections drop, etc. > The driver should clear the interrupt, schedule a reset and ignore all > other activity untill the reset has happened. Of course the reset should > include bringing up the device to full working state again. I wonder which interrupt here happens that requires a rest - these disconnects happen also with a quite idle system. I had no X11 running and just network up and telnet in sometimes. > > However, this is obviously not easy to test and debug, as you will have to > be in a setup where this happens often enough (like your's). Let's try I wonder if 10 increased the frequency of drops or today is an unlucky day! > If you have patience and would be able to help with this: compile your > kernel with "options IWI_DEBUG" and see if that enables enough login to > give us a hint about the cause. If that is not enough you would need to > raise iwi_debug to more than the default level of 4, but that would > spam your log with every received packet. I will! The system should be fast enough. I just found that syssrc.tgz doesn't provide builds.h and one needs whole src.tgz to get it.. a little stupid. And also share. Building now will report back once done / rebooted. > Please file a PR and lets collect data there. Done. kern/58232 Riccardo
Re: framebuffer console on old ATI
Hi Michael, Michael wrote: > > Yeah, I fixed a few problems there, xrender support was broken for r1xx> > Sun's xvr-100 card is a rebadged rv100... Wonderful... eager to test. >> How do I enable radeonfb? Can I do it through configurations or do I >> need to compile my own kernel to try? > > radeonfb* at pci? Do you mean I need to compile and enable it? Since Martin is asking a debug kernel... I think I will try both then. However, I see no radeonfb* in kernel config: In the DRI legacy drivers: #radeondrm* at drm? # ATI Radeon DRM driver and then: radeon* at pci? dev ? function ? radeondrmkmsfb* at radeonfbbus? > > It's not a DRM driver, just an accelerated framebuffer console thing. > If you have an iBook you probably used it. Good to know. Yes, I have an iBook... and there I have strange issues... will ask about that separately. It is a very young 10.0 install and I am giving priority to getting all my Intel machines updated. I am comparing how the T41 behaves, which has a more modern Radeon card. There I get a framebuffer out of the box! and without any font issue either. However, I think at the end drm is enabled there, not radeonfb. [ 1.044664] radeon0 at pci1 dev 0 function 0: ATI Technologies Radeon Mobility M300 (M22) 5460 (rev. 0x00) [ 1.044664] radeon0: autoconfiguration error: unable to reserve VGA registers for i386 radeondrmkms hack [ 5.424688] radeon0: VRAM: 128M 0xC000 - 0xC7FF (64M used) [ 5.424688] radeon0: GTT: 512M 0xA000 - 0xBFFF [ 5.424688] [drm] radeon: 64M of VRAM memory ready [ 5.424688] [drm] radeon: 512M of GTT memory ready. [ 5.424688] [drm] radeon: 1 quad pipes, 1 Z pipes initialized [ 5.424688] radeon0: WB enabled [ 5.424688] radeon0: fence driver on ring 0 use gpu addr 0xa000 and cpu addr 0x0xdc697000 [ 5.424688] radeon0: radeon: MSI limited to 32-bit [ 5.424688] radeon0: radeon: using MSI. [ 5.424688] radeon0: interrupting at msi0 vec 0 (radeon0) [ 5.424688] [drm] radeon: irq initialized. [ 5.473861] [drm] radeon: ring at 0xA0001000 [ 5.513890] radeondrmkmsfb0 at radeon0 [ 5.513890] [drm] Initialized radeon 2.50.0 20080528 for radeon0 on minor 0 [ 5.513890] radeondrmkmsfb0: framebuffer at 0xc00c, size 1400x1050, depth 32, stride 5632 [ 5.603954] wsdisplay0 at radeondrmkmsfb0 kbdmux 1: console (default, vt100 emulation), using wskbd0 This is coherent to the generic kernel config I see. Riccardo
Re: framebuffer console on old ATI
Hi Michael, Michael wrote: > Radeonfb should Just Work(tm) - it's used mainly on macppc and sparc64 > but I'm pretty sure it's been run on at least some little endian hw > before. according to radeon manpage, mine would have an R100, the earliest supported on X11. How do I enable radeonfb? Can I do it through configurations or do I need to compile my own kernel to try? drm did not configure, dmesg says. Riccardo
Re: Intel Wireless fatal error
Hi, Riccardo Mottola wrote: > dmesg | grep iwi0 > [ 1.005536] iwi0 at pci2 dev 2 function 0: Intel PRO/Wireless LAN > 2200BG Mini-PCI Adapter (rev. 0x05) > [ 1.005536] iwi0: interrupting at irq 11 > [ 1.005536] iwi0: 802.11 address 00:16:6f:0c:0f:59 > [ 3574.054462] iwi0: autoconfiguration error: fatal error > > what autoconfiguration failed after it was running for a while? if those > are directly seconds it would be after 1h. just a little after writing this mail, the network disconnected again. So it lasted like 30 minutes at most :( Again another fatal error. [ 1.005536] iwi0 at pci2 dev 2 function 0: Intel PRO/Wireless LAN 2200BG Mini-PCI Adapter (rev. 0x05) [ 1.005536] iwi0: interrupting at irq 11 [ 1.005536] iwi0: 802.11 address 00:16:6f:0c:0f:59 [ 3574.054462] iwi0: autoconfiguration error: fatal error [ 31534.294058] iwi0: autoconfiguration error: fatal error This connection is stable, the router is literally 2 meters away in the same room, actually a small repeating access point that delivers WEB instead of WPA. For easy connection purposes so wpa_config is not needed while hacking+ Ifconfig shows that the ssid and key were still fine. Just as I write this.. it disconnected again ! Even if all parameters according to ifconfig look fine, including IP, reissuing dhcpcd is not enough! But "ifconfig iwi0 up" after a couple of attempts. [ 3574.054462] iwi0: autoconfiguration error: fatal error [ 31534.294058] iwi0: autoconfiguration error: fatal error [ 32554.279739] iwi0: autoconfiguration error: fatal error [ 33274.150790] iwi0: autoconfiguration error: fatal error Riccardo
Intel Wireless fatal error
Hi, I upgraded my trusty Thinkpad to T30 to 10.0 and now most rough edges have been sorted out. Yesterday I connected it via WiFi, connected remotely to it, then went to bed. Today I see the connection dropped and a green message on the console. Ifconfig says network down. dmesg | grep iwi0 [ 1.005536] iwi0 at pci2 dev 2 function 0: Intel PRO/Wireless LAN 2200BG Mini-PCI Adapter (rev. 0x05) [ 1.005536] iwi0: interrupting at irq 11 [ 1.005536] iwi0: 802.11 address 00:16:6f:0c:0f:59 [ 3574.054462] iwi0: autoconfiguration error: fatal error what autoconfiguration failed after it was running for a while? if those are directly seconds it would be after 1h. This issue happened on 9.3 too, sometimes - but I never checked console/dmesg and related things. Riccardo
framebuffer console on old ATI
Hi, I notice that on 10.0 (but was the same on 9.x too) my old ThinkPad T30 runs in VGA text. Is there a specific limitaiton for the embedded videocard? a choice? or something I can try enabling? dmesg reports it as: [ 1.005536] acpivga0 at acpi0 (VID): ACPI Display Adapter [ 1.005536] acpiout0 at acpivga0 (LCD0, 0x0110): ACPI Display Output Device [ 1.005536] acpiout1 at acpivga0 (CRT0, 0x0100): ACPI Display Output Device [ 1.005536] acpiout2 at acpivga0 (TV0, 0x0200): ACPI Display Output Device [ 1.005536] acpiout3 at acpivga0 (DVI0, 0x0210): ACPI Display Output Device [ 1.005536] acpivga0: connected output devices: [ 1.005536] acpivga0: 0x0110 (acpiout0): LCD Panel, head 0 [ 1.005536] acpivga0: 0x0100 (acpiout1): Ext. Monitor, head 0 [ 1.005536] acpivga0: 0x0200 (acpiout2): TV, head 0 [ 1.005536] acpivga0: 0x0210 (acpiout3): Unknown Output Device, head 0 [ 1.005536] vga0 at pci1 dev 0 function 0: ATI Technologies Radeon Mobility M7 LW (rev. 0x00) [ 1.005536] wsdisplay0 at vga0 kbdmux 1: console (80x25, vt100 emulation), using wskbd0 [ 1.005536] drm at vga0 not configured I suppose the last line is the reason for FB console not working. X11, which appears to run fine, reports it as: [ 646.096] (II) VESA(0): VESA VBE Total Mem: 16320 kB [ 646.096] (II) VESA(0): VESA VBE OEM: ATI MOBILITY RADEON 7500 [ 646.096] (II) VESA(0): VESA VBE OEM Software Rev: 1.0 [ 646.096] (II) VESA(0): VESA VBE OEM Vendor: ATI Technologies Inc. [ 646.096] (II) VESA(0): VESA VBE OEM Product: P7 [ 646.096] (II) VESA(0): VESA VBE OEM Product Rev: 01.00 Linux has FB console on ever lower spec ATIs, but I have no direct comparison with this specific one, I admit. Riccardo
i386 - 10.0 gecko browsers
Hi, no firefox and seamonkey as binbary packages for pkgin? Is there any failure reason? At least seamonkey sure runs on 32bit and on Linux Firefox does too... Riccardo
Re: NetBSD i386 missing packages on 10.0 after upgrade - cairo gobject conflict
Hi, Havard Eidnes wrote: > Hmm, I think it's cairo-gobject which needs removing, if I recall > correctly that functinality and files were integrated into the > new version of the cairo package: thanks, you recalled correctly. I removed it and was able to reinstall windowmaker too. X11 starts up... so far 10.0 looks good on the trusty T30! WiFi works too... let me have it compile a little and test and see. Riccardo
Re: NetBSD 10 and unreliable git
Hi Martin and Aryabhata,
network is probably the issue. I was around this weekend and used
wireless in two different places, the connection was good. I know both
places and used them often with different computers (parent's house,
girlfriend home.. not some hotel or airport).
However, at my home today I launched on both repositories via wired
ethernet and it worked.
Martin Husemann wrote:
> I have never seen that (neither on 10.0 nor on -current). It is very
> hard to tell what causes the errors for you. Two potential reasons
> would be network issues (e.g. a bug in the driver for the network device
> you are using) or the limit of simultaneous open files.
I suppose network issues. Since cloning the same repository and updating
the other - both failed before also with reboots - now worked. I hope
git is deterministic.
Now that I write, I remember having seen this error on one of my old
SparcStations. There, just reuissing the command helped. It was however
on a much smaller repository.
>
> If you are using /bin/sh (or bash) please tell us the output of "ulimit -a"
> [or for {t}csh: "limit").
sure:
sphere$ ulimit -a
time (-t seconds)unlimited
file (-f blocks )unlimited
data (-d kbytes ) 262144
stack (-s kbytes ) 4096
coredump (-c blocks )unlimited
memory(-m kbytes ) 7877324
locked memory (-l kbytes ) 2625774
thread(-r threads) 3991
process (-p processes ) 1024
nofiles (-n descriptors) 1024
vmemory (-v kbytes )unlimited
sbsize(-b bytes )unlimited
> What network device are you using? Can you show the relevant part of
> /var/run/dmesg.boot? Also the output of ifconfig for the device (sometimes
> checksum offload breaks things).
[ 1.048313] ral0 at pci3 dev 0 function 0: Ralink Technologies
RT3090 802.11b/g/n (rev. 0x00)
[ 1.048313] ral0: interrupting at ioapic0 pin 19
[ 1.048313] ral0: 802.11 address 70:f3:95:e2:36:11
[ 1.048313] ral0: MAC/BBP RT3090 (rev 0x3213), RF RT3020 (MIMO 1T1R)
Ifconfig in a non-connected status says:
ral0: flags=0x8802 mtu 1500
ssid ""
powersave off
address: 70:f3:95:e2:36:11
media: IEEE802.11 autoselect
status: no network
Do I need to connect?
Riccardo
NetBSD i386 missing packages on 10.0 after upgrade - cairo gobject conflict
Hi, I upgraded two of my classic ThinkPads (i386) from 9.3 to 10.0. Upgrade went apparently smooth. Then I issue pkgin update & upgrade. However, shock... startx fails. windowmaker is missing! I explictely try to install it: sudo pkgin install windowmaker Password: calculating dependencies...done. 1 package to upgrade: cairo-1.18.0nb1 2 packages to install: pango-1.52.1 windowmaker-0.96.0nb4 0 to remove, 0 to refresh, 1 to upgrade, 2 to install 0B to download, 4265K of additional disk space will be used proceed ? [Y/n] [1/3] upgrading cairo-1.18.0nb1... [2/3] installing pango-1.52.1... [3/3] installing windowmaker-0.96.0nb4... pkg_install warnings: 0, errors: 6 pkg_install error log can be found in /var/db/pkgin/pkg_install-err.log I get a strange cairo conflict: pkg_add: Conflicting PLIST with cairo-gobject-1.16.0nb6: include/cairo/cairo-gobject.h pkg_add: 1 package addition failed ---May 06 14:42:46: [2/3] installing pango-1.52.1... pkg_add: Conflicting PLIST with cairo-gobject-1.16.0nb6: include/cairo/cairo-gobject.h pkg_add: Can't install dependency cairo>=1.18.0 pkg_add: Expected dependency cairo>=1.18.0 still missing pkg_add: 1 package addition failed Do I need to uninstall some package that "pking upgrade" didn't catch? or are the packages inconsistent? Riccardo
Re: NetBSD 10 and unreliable git
Hi, Lucifer wrote: > you guys are using too many resources... I'just doing a git clone or a git pull... so don't blame "me". At most git.. or its NetBSD compilation. I am able to perform the same actions on the same repository on slower or less performing systems with other OSs though. Riccardo
NetBSD 10 and unreliable git
Hi, I'm updating fairly big git repositories (e.g. mozilla and similar) and continue getting these errors. I am over a fast WiFi. e$ git pull remote: Enumerating objects: 4822, done. remote: Counting objects: 100% (4822/4822), done. remote: Compressing objects: 100% (1027/1027), done. error: RPC failed; curl 92 HTTP/2 stream 7 was not closed cleanly: CANCEL (err 8) error: 5970 bytes of body are still expected fetch-pack: unexpected disconnect while reading sideband packet fatal: early EOF fatal: fetch-pack: invalid index-pack output I can insist as much as I want, different byte count expected, but I am unable to finish the operation. Being github, it is a decently reliable server... and on other systems I don't have this issue. I don't remember having many issues with 9.3, but maybe it is a mix of network and kernel. Riccardo
Re: netbsd crashes when using fat filesys
Hi Martin, Martin Husemann wrote: - the original issue reported here is*something else* mangling/breaking a FAT file system and NetBSD not dealing with the result. This can either be a bug in Solaris or in NetBSD's interpration of the FAT file system format - we just don't know (yet). An image of a file system in broken state is absolutely required here to debug the issue. If the OP can't provide such an image, we need a*working recipe* how to reproduce the issue, plus someone with a Solaris 10 installation to create the broken image, plus someone to debug the result. Currently we have neither the recipe nor someone with Solaris 10 installed willing to help. I have here locally a Solaris 8 and Solaris 10 box - both sparc64, however no easy way mount stuff since they are servers. I think there are hard-to-access USBs in the back, if they can mount USB keys, I could try to reproduce. Never tried. It would be best if NetBSD could e.g. just refuse to mount, dive a system log and warn with fsck or try to be smart otherwise, not crash! Riccardo
Re: netbsd crashes when using fat filesys
Hi "xuser", xuser wrote: Yes it is that netbsd will crash on driver errors And I found the problem it was that solaris autofs driver would mount an LBA fat 32 as CHS fat32 And so disabling automount support in solaris fixed it. still it would be nice to understand what changes Solaris was doing which disturb NetBSD. I don't think NetBSD should crash... Riccardo
Re: Please forgive a blatant plug: I reviewed v10 for the Reg
Ciao Liam! Liam Proven wrote: I really wish there were more technology sharing between the BSDs. There is actually, but it is never easy. I have seen good transfer between NetBSD and OpenBSD in the years, including drivers and such. Dragonfly has the best installer, IMHO, but of course it has many fewer options to cover. I only use the "canonical" three. I must say as a user I like NetBSD and OpenBSD best. Of course the less platforms the easier it is. Things like partitioning, bootloader complicate things. I think NetBSD has a quite good installer in many aspects. Quick to setup, has a very convenient utility, network setup. Essentially the worst part is partitioning, but it is a tricky matter. On classic BIOS PC setup it works quite well though... quick and fast. Try to partition MacPPC and you get crazy. FreeBSD is the worst inasmuch as it does the least complete job. I agree... however it has some interesting points. I think Debian has a good, but complicated, heavy installer. NetBSD could learn something from it, but not too much. Debian has a decent partitioning tool Some OpenBSD folks are angry with me because I criticise its disk partitioner. When I tell them the config I work with and they recoil and go "OMG that is _impossible!_" OpenBSD are complicated people.. but they do good stuff. Also the prompt based installer is quite good! Upgrading is excellent! But certain things are a bit extreme. like no dhcp setup (must test latest though, maybe they changed it again). The point being: cross-platform installers that work on multiple very different distros with different packaging tools are 100% a thing. I'm not expert there, but they should have peraps more per I am sure it would be possible to write a program which, when run, tests the console or terminal to determine if it can use colour and cursor controls, and if it can, which presents a cursor-key-driven-menu based UI with CUA-style controls -- but if the terminal does not, then falls back gracefully to simple numeric or letter-choice menus. Terminal type does that for you... and NetBSD install works well even ona 9600 baud serial vt100, which is really legacy technology. Long-term users often tell me that they do not notice the issues because they simply upgrade from one version to the next and never see the installer. Well, in that case, offer that opportunity to visitors as well: it would be to the benefit of all of the BSD family if the projects supplied pre-installed and pre-configured VM images for direct download, so that the curious could simply download an OVA file, import it into the hypervisor of their choice, and try the OS out without installing it at all. Yes, upgrading sometimes does not well test the bare install. However both are important applications. I tend to too to upgrade... In the case of NetBSD however you still test a big part of the install - except partitioning. You do all steps! I just did an upgrade on SPARC64 and it worked wonderfully. Cheers, Riccardo
Re: $squid is not set properly
Hi, Greg Troxel wrote: Almost certainly, you have squid installed via pkgsrc, or you have a leftover /etc/rc.d/squid because you used to. oh, got you! Squid was there. Actually, I really had the package installed and left unused. Possibly some testing. The logs indicated it did some runs in the past. I just removed it (no dependencies found) and instruction to remove squid user and such appeared. I cleaned everything up as per messages. Next reboot... all fine. Thanks, Riccardo
$squid is not set properly
Hello, on boot, I see this message: /etc/rc: WARNING: $squid is not set properly - see rc.conf(5). but I have no squid related lines in my /etc/rc.conf and I don't use squid, so I don't think I need it (or does the system need it?). Riccardo
Re: NetBSD 9.3 to 10.0 upgrade failure - check for DOS fs
Hi, Martin Husemann wrote: > On Tue, Apr 23, 2024 at 01:20:12PM +0000, Riccardo Mottola wrote: >> /dev/sd0e /media/usb msdos rw 0 0 > What is that partition? > Is that "usb" device available (as sd0) during your upgrade experiment? external hard disk. NO it was not mounted, but it is also "noauto" > > If you boot the installer from a USB disk, that might become sd0 and > things will go awfully wrong. This still is a bug - sysinst should recognize > the device clash, but I'm pretty sure there currently is no such code. True, but I also (first attempt actually) booted off CD, which should remain cd0. > > Maybe you could try commenting that partition out in /etc/fstab, do the > upgrade > and later manually enable it again? However, this was the trick. I commented that line out... and installer CD worked like a charm. So, it got really confused but that fstab I'm in! 10 boots. pkgin upgrade done. Framebuffer console works, X11 works, ethernet works. Touchpad was unusable before and still is :) Now I will stress it a bit, compile, check stuff like sleep, audio, wifi Riccardo
Re: NetBSD 9.3 to 10.0 upgrade failure - check for DOS fs
Hi, (removing Robert who may read the reply on the list, if) Martin Husemann wrote: > On Wed, Apr 10, 2024 at 11:52:04PM +0200, Riccardo Mottola wrote: >> 0: NetBSD (sysid 169) >> bootmenu: NetBSD >> start 2048, size 625140400, Active >> 1: >> 2: >> 3: >> Bootselector disabled. >> First active partition: 0 > Can you show us the /etc/fstab file from the NetBSD partition? of course, here it is. Inspected from booting 9.3 [1] - I hope that is what you asked for. sphere$ cat /etc/fstab # NetBSD /etc/fstab # See /usr/share/examples/fstab/ for more examples. /dev/wd0a / ffs rw 1 1 /dev/wd0b noneswapsw,dp0 0 kernfs /kern kernfs rw ptyfs /dev/ptsptyfs rw procfs /proc procfs rw tmpfs /var/shmtmpfs rw,-m1777,-sram%25 /dev/sd0e /media/usb msdos rw 0 0 /dev/cd0a /media/cdromcd9660 ro,noauto it confirms that "wd" is completely dedicated to NetBSD. If we had nice BSD stickers (especially with the Daemon) I'd remove the Windows7 for it :) Riccardo [1] wow, I missed 9.4 release, should I upgrade to that at least? getting to 10 should be still easy. However this is my main NetBSD laptop,s o it should have latest stable. I have then one with testing and one I will keep on 9.x :)
Re: Please forgive a blatant plug: I reviewed v10 for the Reg
Hi Liam. Liam Proven wrote: I thought this might interest folks here... Nice share and thanks for taking the time to write it. NetBSD 10 proves old tech can still kick apps and take names three decades later Proper old-school Unix, not like those lazy, decadent Linux types https://www.theregister.com/2024/04/17/30yo_netbsd_releases_v10/ Comments and feedback welcomed! I mostly agree with the article, the citations, it gives the mood back. Then it shares some personal opinions about experience. Some considerations: * About "size" and "performance". NetBSD runs well on a lot of software, in my experience the core OS is very comparable to OpenBSD, but a litte bit easier since it doesn't have the relink stuff, and base Linux. However recent linux with systemd are definitely worse. However still there is "imported userland" and then "pkgsrc". Except for raw speed in system calls, kernel management, if you install package X it will about be similar in size on different operating systems. So for sure since XFCE has gotten fatty, you get the same fat on NetBSD. Same goes for GCC which gets fatter and slower with each release :) Of course, new features are added. You start to see the size of packages installed more than the OS itself * installation experience. I like the menu-driven installer, it has the classic semi-GUI with windows and arrows, but still very usable on a serial console (handy if you install on a server!). The rough edges in my opinion are in the partitioning. I am myself bitten right now by an upgrade on a "standard x86-64 laptop". Resizing partitions, selecting, labels... that part. Of course it is complicated by all the many differences in BIOS, UEFI... OpenFirmware on mac, etc etc However, when things go smooth, e.g. on a classic BIOS-based PC where you take over the whole disk, it is very fine and convenient. Riccardo
Re: NetBSD 9.3 to 10.0 upgrade failure - check for DOS fs
Hi, Martin Husemann wrote: Can you boot the original install CD you used, exit the installer and check the output of: sysctl machdep.bootmethod This should either tell you BIOS or UEFI, and this is how the installer decides what kind of setup the booted system later will need. Original install CD... as in 10.0 standard CD or "one used to install... years ago" ? That was lost in time... probably 9.0 or even 8.x With the 10.0 install ISO amd64 the command you say yields machdep.bootmethod = BIOS Which is correct I believe, since UEFI is disabled in the BIOS. Riccardo
Re: NetBSD 10 and framebuffer consoles setup vs 9.3 (font, multiple...)
Hi, RVP wrote: and the system will now use the larger font if you have a high enough display. Otherwise, you would end up with an unreadable chars. on, say, 4K displays. For 4K displays maybe, but on any laptop I have at hand - without prehaps a "Retina" or equivalent display, the small font is readable enough and actually the more information makes installing, merging files reading dmesg much much easier. Up to my HD display on 22" standard font is fine. Maybe it is just which threshold you consider "high enough". While installing, you can escape into the shell, then do: wsconsctl -dw font=Boldface Will try at the next run - it works in normal boot. Perhaps an installer option could be also useful? like keyboard setup, display setup. Is this mis-behaviour seen always or only if you sudo? https://mail-index.netbsd.org/pkgsrc-users/2024/03/23/msg039239.html I see this only when using a recent sudo, and only on constty. now that you told me, I checked on two consoles with and without sudo. It is indeed the combination of sudo & first console. does it have consequences to change constty to ttyE0? loosing "console" status? Riccardo
Re: NetBSD 10 and framebuffer consoles setup vs 9.3 (font, multiple...)
Hi, comparing to my old 9.3 system and doing some experimenting... I found some difference and have questions. Riccardo Mottola wrote: I used the sysupgrade method for the first time, I hope I didn't mess up with etcupdate step with required a lot of manual work. Also, since the console was in 80x24 mode, checking diff of etcfiles was really cumbersome. I can reboot and login, however my consoles don't behave as expected comapred to 9.3. How can I fix it? Did I miss some configuration? 1) Font size. The framebuffer is enabled (I notice during boot the switch from VGA) but the displayed font gives me 80x24 anyway... how can I do high-res as before (not 8x8 size on VGA, but classic expected framebuffer I got before... this new setup is very unusual) Do I need to mess in wscons.conf? boot.cfg? Here I found that enabling fonts helps. To restore the old behavior, I explicitely need to set: setvar ttyE0 font Terminus16B-ISO8859-1 apparently otherwise something which looks like 32B seems to be used... 2) multiple consoles. ctrl/alt/F1 doesn't open a new login... just the basic console Here I enabled: ttyE0 "/usr/libexec/getty Pc" wsvt25 off secure ttyE1 "/usr/libexec/getty Pc" wsvt25 on secure ttyE2 "/usr/libexec/getty Pc" wsvt25 on secure ttyE3 "/usr/libexec/getty Pc" wsvt25 on secure And now I have all my consoles. 3) basic console behaves strange. login and shell does work, but if I do vi, I cannot navigate, keyboard is mapped wrongly, cursor keys do not work... totally crazy. Also certain promts spit out withyout newline, only carriage-return. Like it is not behaving as vt100? This is a very nasty thing. However, now that I have 4 consoles, I notice that 1,2,3 work fine, it is console 0 that doesn't work well! But it is also... the first one and the only one I had at the beginning. On 10.0 I notice: console "/usr/libexec/getty Pc" vt100 off secure constty "/usr/libexec/getty Pc" vt100 on secure While on 9.3: console "/usr/libexec/getty Pc" wsvt25 off secure constty "/usr/libexec/getty Pc" wsvt25 on secure notice vt100 vs wsvt25? However if I compare dmesg, I see almost the same 10.0 [ 5.876542] wsdisplay0 at radeondrmkmsfb0 kbdmux 1: console (default, vt100 emulation), using wskbd0 [ 6.006667] wsmux1: connecting to wsdisplay0 [ 8.609158] wsdisplay0: screen 1 added (default, vt100 emulation) [ 8.609158] wsdisplay0: screen 2 added (default, vt100 emulation) [ 8.609158] wsdisplay0: screen 3 added (default, vt100 emulation) [ 8.609158] wsdisplay0: screen 4 added (default, vt100 emulation) 9.3 [ 6.442132] wsdisplay0 at intelfb0 kbdmux 1: console (default, vt100 emulation), using wskbd0 [ 6.467386] wsmux1: connecting to wsdisplay0 [ 14.397163] wsdisplay0: screen 1 added (default, vt100 emulation) [ 14.397163] wsdisplay0: screen 2 added (default, vt100 emulation) [ 14.397163] wsdisplay0: screen 3 added (default, vt100 emulation) [ 14.397163] wsdisplay0: screen 4 added (default, vt100 emulation) except intel vs. radeon, they both say vt100, so I am puzzled, it looks as the 9.3 is inconsistent ! Riccardo
Re: NetBSD 9.3 to 10.0 upgrade failure - check for DOS fs
Robert Elz wrote: | no "e" of course... and no MS-DOS in sight. It was already a fully | BSD-ized system. What does fdisk show? (ie: the MBR label). fdisk on wd0 run from the utility shell of the install cd says: 0: NetBSD (sysid 169) bootmenu: NetBSD start 2048, size 625140400, Active 1: 2: 3: Bootselector disabled. First active partition: 0 If I just boot into 9.3 an run it (so I can easily copy it instead of typing it from screen): Disk: /dev/rwd0 NetBSD disklabel disk geometry: cylinders: 620181, heads: 16, sectors/track: 63 (1008 sectors/cylinder) total sectors: 625142448, bytes/sector: 512 BIOS disk geometry: cylinders: 1022, heads: 255, sectors/track: 63 (16065 sectors/cylinder) total sectors: 625142448 Partitions aligned to 16065 sector boundaries, offset 63 Partition table: 0: NetBSD (sysid 169) bootmenu: NetBSD start 2048, size 625140400 (305244 MB, Cyls 0/32/33-38913/80/63), Active 1: 2: 3: Bootselector disabled. First active partition: 0 Drive serial number: 3120074710 (0xb9f88fd6) which seems consistent. Riccardo
Re: NetBSD 9.3 to 10.0 upgrade failure - check for DOS fs
Hi, Martin Husemann wrote: Try the BIOS-only install image instead - I bet your system boots the CD via UEFI, but the original installation was BIOS only. when booting from CD, I see: NetBSD/x86 BIOS Boot Revision 5.11 Is this meaningful? Riccardo
Re: NetBSD 9.3 to 10.0 upgrade failure - check for DOS fs
Hi Martin, Martin Husemann wrote: On Tue, Apr 09, 2024 at 10:28:46PM +0200, Riccardo Mottola wrote: So I have a perfect working 9.3 installation on on an HP ProBook laptop. I boot the CD which works fine Try the BIOS-only install image instead - I bet your system boots the CD via UEFI, but the original installation was BIOS only. Your explanation sounds plausible. However, "UEFI Boot Mode" is disabled in BIOS, I don't know if it is still used by the CD or "detected" somehow anyway. How can I tell further? I did get the "bios image". I see it is USB only? no ISO? Of course my USB Key decided to strike, but I used a spare external hard drive. Much faster to write too :) It boots. but tries to do the same silly thing, fsck a non-existing MS-DOS partition. Yes, this is a bug, the installer should find out and deal with it. Well, I suppose it can find out at least what partitions are there? Other attempts and workarounds? I already have a half-assed system using sysupgrade, so I want to avoid that. My goal was to use this system with the classic upgrade and then maybe compare files like wscons, etc. (Although I already have noticed that the framebuffer uses the silly low-resolution font). Do we have install kernels like I used on OpenBSD before the upgrade tool? download and boot from hard-disk and hope it figures out things better... Riccardo
NetBSD 9.3 to 10.0 upgrade failure - check for DOS fs
Hi! I want to upgrade my 9.3 amd64 laptop. Since my other upgrade using sysupgrade was a failure, i tried doing it via CD install. I still hope either to get help fixing my other laptop or being able to "compare" with a successful install So I have a perfect working 9.3 installation on on an HP ProBook laptop. I boot the CD which works fine I get a low-res framebuffer (it is detected as 1366x768, but VGA size fonts are displayed, please please put Hi-Res back! all other BSDs, including obscure Darwin or Linux do that!) I select b-> Upgrade Confirm -> yes Which disk? wd0 (it is the only one!) I thus get "Command failed" /sbin/fsck_msdos -p -q /dev/rsd0e device not configured why the heck is it trying to check for an msdos filesystem? I get into the utility menu, run fdisk on "wd0" and then see one partition table 0, NetBSD disklabel shows: a: 4.2BSD b: swap c unused d: unused no "e" of course... and no MS-DOS in sight. It was already a fully BSD-ized system. Riccardo
NetBSD 10 and framebuffer consoles setup vs 9.3 (font, multiple...)
Hi! I just upgraded from 9.3 to 10 on amd64 with AMD Radeon I used the sysupgrade method for the first time, I hope I didn't mess up with etcupdate step with required a lot of manual work. Also, since the console was in 80x24 mode, checking diff of etcfiles was really cumbersome. I can reboot and login, however my consoles don't behave as expected comapred to 9.3. How can I fix it? Did I miss some configuration? 1) Font size. The framebuffer is enabled (I notice during boot the switch from VGA) but the displayed font gives me 80x24 anyway... how can I do high-res as before (not 8x8 size on VGA, but classic expected framebuffer I got before... this new setup is very unusual) Do I need to mess in wscons.conf? boot.cfg? 2) multiple consoles. ctrl/alt/F1 doesn't open a new login... just the basic console 3) basic console behaves strange. login and shell does work, but if I do vi, I cannot navigate, keyboard is mapped wrongly, cursor keys do not work... totally crazy. Also certain promts spit out withyout newline, only carriage-return. Like it is not behaving as vt100? Riccardo
Hangs when trying to read a NTFS disk with 3g (No buffer space?)
Hi! I installed fuse-ntfs-3g-2022.10.3 to read on an external HD. I did this on an i386 laptop and an amd64 one, to double check.. and because I need it on both! I try to read this disk [ 7080.368720] sd0 at scsibus0 target 0 lun 0: disk fixed [ 7080.378727] sd0: fabricating a geometry [ 7080.378727] sd0: 1397 GB, 1430799 cyl, 64 head, 32 sec, 512 bytes/sect x 2930277168 sectors [ 7080.388733] sd0: fabricating a geometry I mount the disk this way and get the following messages: isengard$ sudo mount.ntfs-3g /dev/sd0e /mnt mount.ntfs-3g: perfuse_open: setsockopt SO_SNDBUF to 2162688 failed: No buffer space available mount.ntfs-3g: perfuse_open: setsockopt SO_RCVBUF to 2162688 failed: No buffer space available mount.ntfs-3g: perfuse_open: setsockopt SO_SNDBUF to 2162688 failed: No buffer space available mount.ntfs-3g: perfuse_open: setsockopt SO_RCVBUF to 2162688 failed: No buffer space available mount.ntfs-3g: perfuse_open: setsockopt SO_SNDBUF to 2162688 failed: No buffer space available mount.ntfs-3g: perfuse_open: setsockopt SO_RCVBUF to 2162688 failed: No buffer space available A quick check with "ls /mnt" shows the expected directories, but if I try to read into a subdirectory, I get a hang. In one attempt, the system did hang, I could not kill applications, only a (soft) power-down helped. Right now I could quit with crl-c, I see that I can't list several first-level directories, but some *do* work. I tried to copy some files... I get a hang. After a second attempt I get a hang I cannot ctrl-c. I see no output nor dmesg with a clue. Ideas? Experiences with 3f NTFS? Riccardo
laptop crashes - ATI radeon GPU issues?
Hi All! I installed NetBSD on a laptop with amd64 (genuine Turion64 x2) and ATI video card. I notice some errors in dmesg and wanted to report them without hurry, since everything appeared to work. However, I noticed a lot of crashes with this laptop. What happens at boot it may hang (quite often) or reset the moment the framebuffer gets initialized. Sometimes several times in a row. Sometimes, when I do get to the login, I can start X11 and see some garbage during init and then a crash+reset before actual X11 appears. If I keep trying/rebooting it works eventually and if it does, it is stable both in console and X11, also after hours of usage and compilation in background. I would thus exclude hardware issues. It is like go or not go, but if it goes, it works. To be sure, I also tried changing RAM and cleaning the cooling fan, without changes. I Think it is some issue the way the framebuffer gets initialized. Some days it just works at the first attempt. At the last crash, I got a core recovered. This is here: [ 133.1036804] wsdisplay0: screen 1 added (default, vt100 emulation) [ 133.1136884] wsdisplay0: screen 2 added (default, vt100 emulation) [ 133.1136884] wsdisplay0: screen 3 added (default, vt100 emulation) [ 133.1136884] wsdisplay0: screen 4 added (default, vt100 emulation) [ 416.5794537] ohci0: WARNING: addr 0x4596c560 not found [ 416.9773093] fatal protection fault in supervisor mode [ 416.9773093] trap type 4 code 0 rip 0x8047f1a0 cs 0x8 rflags 0x13282 c r2 0x72335dc07dd9 ilevel 0 rsp 0xb10054a36a30 [ 416.9773093] curlwp 0xad9b0732a0a0 pid 645.1 lowest kstack 0xb10054a34 2c0 [ 416.9773093] panic: trap [ 416.9773093] cpu0: Begin traceback... [ 416.9773093] vpanic() at netbsd:vpanic+0x160 [ 416.9773093] snprintf() at netbsd:snprintf [ 416.9773093] startlwp() at netbsd:startlwp [ 416.9873170] alltraps() at netbsd:alltraps+0xbb [ 416.9873170] usbd_create_xfer() at netbsd:usbd_create_xfer+0x18d [ 416.9873170] usbd_open_pipe_intr() at netbsd:usbd_open_pipe_intr+0x64 [ 416.9973225] uhidev_open() at netbsd:uhidev_open+0x17e [ 416.9973225] ums_enable() at netbsd:ums_enable+0x4e [ 416.9973225] wsmuxopen() at netbsd:wsmuxopen+0xf9 [ 417.0073279] cdev_open() at netbsd:cdev_open+0xeb [ 417.0073279] spec_open() at netbsd:spec_open+0x175 [ 417.0073279] VOP_OPEN() at netbsd:VOP_OPEN+0x4c [ 417.0173334] vn_open() at netbsd:vn_open+0x2b3 [ 417.0173334] do_open() at netbsd:do_open+0x103 [ 417.0273383] do_sys_openat() at netbsd:do_sys_openat+0x8b [ 417.0273383] sys_open() at netbsd:sys_open+0x24 [ 417.0273383] syscall() at netbsd:syscall+0x157 [ 417.0273383] --- syscall (number 5) --- [ 417.0373432] 723365e42d2a: [ 417.0373432] cpu0: End traceback... dmesg shows at boot: [ 4.896025] kern info: [drm] initializing kernel modesetting (RS480 0x1002:0x5975 0x103C:0x30B0). [ 4.896025] kern info: [drm] register mmio base: 0xd430 [ 4.896025] kern info: [drm] register mmio size: 65536 [ 4.896025] kern info: [drm] Generation 2 PCI interface, using max accessible memory [ 4.896025] radeon0: info: VRAM: 128M 0xB800 - 0xBFFF (128M used) [ 4.896025] radeon0: info: GTT: 512M 0x8000 - 0x9FFF [ 4.896025] kern info: [drm] Detected VRAM RAM=80M, BAR=128M [ 4.896025] kern info: [drm] RAM width 128bits DDR [ 4.896025] Zone kernel: Available graphics memory: 761766 kiB [ 4.896025] kern info: [drm] radeon: 128M of VRAM memory ready [ 4.896025] kern info: [drm] radeon: 512M of GTT memory ready. [ 4.896025] kern info: [drm] GART: num cpu pages 131072, num gpu pages 131072 [ 4.896025] kern info: [drm] radeon: 2 quad pipes, 1 z pipes initialized. [ 4.896025] kern info: [drm] PCIE GART of 512M enabled (table at 0x45972000). [ 4.896025] radeon0: info: WB enabled [ 4.896025] radeon0: info: fence driver on ring 0 use gpu addr 0x8000 and cpu addr 0x0xc1b717a07000 [ 4.896025] kern info: [drm] Supports vblank timestamp caching Rev 2 (21.10.2013). [ 4.896025] kern info: [drm] Driver supports precise vblank timestamp query. [ 4.896025] radeon0: info: radeon: MSI limited to 32-bit [ 4.896025] radeon0: interrupting at ioapic0 pin 17 (radeon0) [ 4.896025] kern info: [drm] radeon: irq initialized. [ 4.896025] kern info: [drm] Loading R300 Microcode [ 4.926041] ACPI Error: Aborting method \_GPE._L19 due to previous error (AE_AML_LOOP_TIMEOUT) (20190405/psparse-581) [ 4.926041] ACPI Error: AE_AML_LOOP_TIMEOUT, while evaluating GPE method [_L19] (20190405/evgpe-603) [ 4.926041] acpiacad0: AC adapter online. [ 4.956057] kern info: [drm] radeon: ring at 0x80001000 [ 5.086125] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/radeon/radeon_r100.c:3701)r100_ring_test] *ERROR* radeon: ring test failed (scratch(0x15E4)=0xCAFEDEAD) [ 5.086125] kern error:
pkgconf freetype and flags
Hi!
who sets what pkgconf returns for the packages? Is it upstream or does
it come from NetBSD?
I think there is an issue with freetype, missing the other part.
Here:
osgiliath: {64} pkg-config --libs freetype2
-L/usr/pkg/lib -lfreetype
osgiliath: {65} pkg-config --libs-only-other freetype2
if I compare it with nettle, which works fine:
osgiliath: {72} pkg-config --libs nettle
-Wl,-R/usr/pkg/lib -L/usr/pkg/lib -lnettle
osgiliath: {73} pkg-config --libs-only-other nettle
-Wl,-R/usr/pkg/lib
we see that for freetype -Wl,-R/usr/pkg/lib is missing and this causes
me various issues during configures and builds.
Riccardo
Re: would anybody use binary packages for NetBSD/i386 10?
Hi Michael, Michael van Elst wrote: What else? Self-compiling on a system you already consider outdated?:) Binary packages are more important on systems that we consider old, doesn't have to be a VAX. you nail it.. actually older or slower systems need binary things. Imagine compiling everything on an Raspberry! The same goes for older laptops, embedded systems, etc, etc I see that on SPARC systems where we don't have many packages and building stuff myself needs quite some time. Riccardo
Re: would anybody use binary packages for NetBSD/i386 10?
Hi Greg Greg Troxel wrote: In contemplating bulk builds and resources, I wonder if there are still people who: are running NetBSD/i386 (as opposed to amd64) are using the binary packges from quarterly branches on ftp.netbsd.org are running NetBSD 10 already, or who intend to move to it soon or after release I still have several i386 systems and I intend to gradually switch some or all of them to NetBSD 10 and binary packages are welcome. My guess is that at this point, i386 use is limited to a) old embedded-type systems (soekris) b) systems that are running i386 because they were first installed many years ago and haven't been converted to amd64 for no good reason or for some odd special case odd reason c) build systems to support category a/b systems, for testing or building private binary package sets d) retrocomputing I am mostly in category a) or d). that is, pure 32bit systems of various age. What can run 64bit has been upgraded. I add as a category also VM systems, since for most of them 32bit is enough, 64bit wll use more resources, but it is probably nowadays negligible for powerful hosts. Most of my systems are laptops, some just "old but good" and some... more vintage (but heck, the older, the nicer the keyboard...) If a system runs fine and is usable with 9.3, is there any good reason why it shouldn't run well with 10? Slower? More resource usage? Except where support of certain devices broke, it NetBSD has a long history of upgrading well for me. and that the amount of use with ftp.n.o binary packages is extremely small. With time, userland became very fat, so building a usable desktop environment with gui on a laptop is not desiderable, especially given RAM limitations of older laptops and ageing hard disks. So while I did build everything from source 5 years ago, I now welcome binary packages especially for things like GCC, browsers heck even installing git and subversion is fat work. It is no longer gcc 2.95 building cvs! Riccardo
Re: would anybody use binary packages for NetBSD/i386 10?
Hi David, David Brownlee wrote: I also think there are a potentially interesting (if small) set of people who would like a desktop with minimal web browser on an older 32bit x86 system, and NetBSD + ArcticFox pretty much delivers on that nice to see ArcticFox mentioned! Indeed, I use it on an OpenBSD workstation with a Pentium 4 HT and on a Thinkpad R52 with Gentoo Linux and ArcticFox delives still a very tood browsing experience for wikipedia, netbsd.org and other quick lookups. Not a daily browsing system, but enough to consult the web, get development stuff, consult CVS diffs, ecc. Very handy. But to get AF building a whole need of dependencies is needed which is good to have ready. [*] I also have ArcticFox on an almost identical Thinkpad with NetBSD, but there is usable, but less pleasant to use. I guess it is the inferior video card, but it should be analyzed. Riccardo [*] Except github and gitlab which now are extremely web browser demanding, unfortunately.
Re: Nouveau
Hi, Todd Gruhn wrote: How / when to upgrade nouveau? it is inside NetBSD core, so it gets updated for you with release updates. Riccardo
Re: Seting time
Hi, Todd Gruhn wrote: Last night I set clock 1hr ahead. how did you set it ahead? Today, X-clock still says time is 4:20pm (it should be 5:20pm). did you reboot/restart in the meanwhile? Why isnt this being set/changed? Must I use 'date' ??? what does "date" say in the console? The time you need or the old one? is the TimeZone correct? Riccardo
Re: firefox resource hog
Hi, On 1/7/23 18:49, Clay Daniels wrote: I find that firefox 105 or 107 are almost unusable on a laptop running NetBSD 10.0 BETA. My older 2014 machine has similar 4gb ram & I have found arcticfox works best there. but at the end ArcticFox is really close to Firefox, something like a mix of FF 40-52 things... backports, in features and fixes, but at the "core" it should behave as FF 40-45. This means its process setup, it does not use E10s/Electrolysis yet and has the old-old interface and is "full C/C++" without rust. Being the main author behind it, I do know. It comes from PaleMoon 27, but I "firefoxized" it a lot since the fork - trying not to sacrifice much. Besides startup times, on a normal Intel computer with 2 cores with 2-4GB of RAM I must say general usage of Firefox compares well with ArcticFox. Something like the NetBSD website, wikipedia, DuckDuckGo. What changes a lot is opening more tabs! but with less than 3GB or ram one shouldn't really have more than a couple of tabs nowadays. I too notice things are slower on NetBSD with Firefox and ArcticFox seems to do better, so the hint that "threads" and "processes" might be an issue is a hint. Also, acceleration does make a difference and on similar systems, Linux still seems to be a tad better and/or supporting more cards. Unfortunately, on my two best NetBSD laptops where I would browse, I can't test well. -> on one laptop, x86 + nvidia, nouveau doesn't work and the "nv" driver is so slow and full of scrolling artifacts that beyond compilation and startup, I can't "test and compare" ArcticFox -> my other, x86-64 + i915, is running current NetBSD and there ArcticFox stopped building. So if you like ArcticFox on NetBSD, help :) As another comparison, there is FreeBSD. Unfortunately, ArcticFox doesn't support most recent versions of FreeBSD either, but on my workstation with x86-64 and i915 I can still compare and there ArcitcFox and FireFox are "closer" like linux, although FF uses more ram with many tabs. Summary: - I think NetBSD trails in performance, due to some thread related thing + graphic acceleration, but it is just a gut feeling. - if you like ArcticFox, it needs help on BSDs :) Riccardo
Re: NetBSD and old nvidia laptop card / crash and unusable X11
Hi Julian, On 15/12/2022 18:28, Julian Coleman wrote: I wonder if it's crashing with the screen off? Are you able to build a kernel with: options DDB_ONPANIC=1 options DDB_COMMANDONENTER="trace;reboot" ? We might then be able to catch the crash details on the next boot. it took a little ride, I had to set up everything, but was able to compile current sources with the options you gave me. The kernel, at boot, crashes, I get a trace, a reboot (I suppose that's the second lines of options you gave me). But where should be the crash details? in /var/crash there is nothing new. I made a "screenshot" of the error. I'll post it on netbsd-current, since I think it is more correct than users for the current sources. Riccardo
fpudna: FPU busy with EagerFPU enabled
Hi, my i386 laptop running 9,3 was compiling. I attempted to switch consoles (text only nouveau is disabled) with ctrl-alt-F1 but it crashed with kernel fault. At reboot, I see savecore: rebood after panic: panic: fpudna: FPU busy with EagerFPU enabled crash is in netbsd.8.core.gz Never seen this error, does it ring a bell? Riccardo -- Sent with GNUMail from GNUstep on a Raspberry PI 4.
Re: NetBSD and old nvidia laptop card / crash and unusable X11
Hi, On 2022-12-23 14:30:35 + Greg Troxel wrote: It is entirely possible to build a full current or 10 release on 9. You can then run either kernel with 9 userland. To build, just "./build.sh release". The time you have to wait will be less than the time to figure out how to avoid that. But you can build tools, and then kernel, without building all of userland. I hoped to skimp on tools :) But well... I did try not to build full release. I did build tools ./build.sh -U tools then kernel ./build.sh kernel=myconf but kernel build fails. I thought to have gootten a bad CVS release, updated and now tool fails :( I will report on netbsd current, since for pratical purposes it is an issue of building current on i386... debugging nouveau will come later! Riccardo -- Sent with GNUMail from GNUstep on a Raspberry PI 4.
Re: NetBSD and old nvidia laptop card / crash and unusable X11
Hi, Riccardo Mottola wrote: I can do that very easily on the 9.3 kernel sources which I have installed and the patch by Robert... because I have those sources already setup up and ready ro rebuild. I don't have the current kernel stuff here though. answering to my self after private with Robert: current needs to be build. I got the sources, upgraded to current. I have of course full 9.3 userland. Is a build possible at all without a switch? I configured the kernel, added the options you suggested and it dies with: In file included from ../../../../netipsec/ipsec.c:103:0: ../../../../netipsec/ipsec.c: In function 'ipsec_checkpcbcache': ../../../../netipsec/key_debug.h:64:8: error: expected ')' before '__VA_OPT__' __VA_OPT__(,) __VA_ARGS__); \ ^ ../../../../netipsec/ipsec.c:253:2: note: in expansion of macro 'KEYDEBUG_PRINTF' KEYDEBUG_PRINTF(KEYDEBUG_IPSEC_STAMP, ^~~ ../../../../netipsec/key_debug.h:63:19: error: format '%d' expects a matching 'int' argument [-Werror=format=] log(LOG_DEBUG, "%s: " fmt, __func__ \ ^ ../../../../netipsec/ipsec.c:253:2: note: in expansion of macro 'KEYDEBUG_PRINTF' KEYDEBUG_PRINTF(KEYDEBUG_IPSEC_STAMP, <..cut..> do I need to build full tools first? O did I get a bad release? Soon I need to move this discussion to "netbsd-current"... darn. Riccardo
Re: NetBSD and old nvidia laptop card / crash and unusable X11
Hi Julian, Julian Coleman wrote: I wonder if it's crashing with the screen off? Are you able to build a kernel with: options DDB_ONPANIC=1 options DDB_COMMANDONENTER="trace;reboot" ? We might then be able to catch the crash details on the next boot. Regards, I can do that very easily on the 9.3 kernel sources which I have installed and the patch by Robert... because I have those sources already setup up and ready ro rebuild. I don't have the current kernel stuff here though. Robert: is your patch in current already? Riccardo
Re: NetBSD and old nvidia laptop card / crash and unusable X11
Hi Benny, Benny Siegert wrote: Hi Riccardo, On Thu, 8 Dec 2022, Riccardo Mottola wrote: I have a fine and nice laptop which unfortunately has an older nvidia card. I updated to 9.3, but the same issue persist as older versions. is it any better if you try booting a NetBSD-current snapshot? yes and no... it behaves differently. I just used the current kernel here http://nycdn.netbsd.org/pub/NetBSD-daily/HEAD/202212142120Z/i386/binary/kernel/ The screen becomes black when fb sould be turned on. This is similar (but not same) behaviour after applying a patch of Robert Swindell about correct PCI BAR detection. The screen is black, the boot does not complete (cannot remotely log-in), caps-lock does not work, power-off or ctrl-alt-canc neither. Riccardo
NetBSD and old nvidia laptop card / crash and unusable X11
Hi All, I have a fine and nice laptop which unfortunately has an older nvidia card. I updated to 9.3, but the same issue persist as older versions. lspci identifies the card as: 01:00.0 VGA compatible controller: NVIDIA Corporation G72M [Quadro NVS 110M/GeForce Go 7300] (rev a1) If I boot, it will crash the kernel during boot, also in the installer kernel, so it is quite inconvenient. If I disable nouveau (boot with -c) then the system comes up. The only X11 driver usable at this point is the NV one. Small extract of Xorg.log at the bottom, it matches lspci. Which we all know "sucks". It comes up and xterm works. glgxgears works! It ranges 100-1300 frames (very inconsistent) But most application are unusable due to refresh and redraw issues. Firefox comes up, but is extremely slow in redrawing, you literally see the menus, tabs drawing their controls. Youtube video is pathetic. Using gvim gtk is interesting, it come sup, but as soon as you scroll, most of the text disappears, like it is being redrawn by "drawing" without antialias in white over the old text, so it ghosts. GIMP leaves large black areas when clicking on controls, but mostly works. GNUstep apps then are almost unusable as gvim-gtk, scrolling areas leaves with text halos of the previous of the previous one.. IN the short-term, has somebody an idea on how to tweak the driver prefs, perhaps some acceleration is misbehaving? In the longer term, I hope the issues can be fixed the old X11 driver is, if I remember correctly, close to unmaintainable being just open-sourced proprietary code very hard to read. Xorg.log is spammed with entries like these: [ 2003.211] (EE) NvPrepareSolid 8 bpp [ 2003.212] (EE) NvPrepareSolid 8 bpp [ 2003.212] (EE) NvPrepareSolid 8 bpp [ 2003.237] (EE) NvPrepareSolid 8 bpp [ 2003.237] (EE) NvPrepareSolid 8 bpp [ 2003.237] (EE) NvPrepareSolid 8 bpp [ 2003.237] (EE) NvPrepareSolid 8 bpp [ 2003.237] (EE) NvPrepareSolid 8 bpp [ 2003.237] (EE) NvPrepareSolid 8 bpp [ 2003.237] (EE) NvPrepareCopy 8 bpp [ 2003.237] (EE) NvPrepareCopy 8 bpp [ 2003.237] (EE) NvPrepareSolid 8 bpp [ 2003.237] (EE) NvPrepareSolid 8 bpp [ 2003.238] (EE) NvPrepareSolid 8 bpp [ 2003.238] (EE) NvPrepareSolid 8 bpp [ 2003.238] (EE) NvPrepareSolid 8 bpp [ 2003.238] (EE) NvPrepareSolid 8 bpp Maybe The best options would be to fix nouveau. The Go 7300 according to https://www.techpowerup.com/gpu-specs/geforce-go-7300.c1579 should have a G72 Curie processor,. It should be here: https://nouveau.freedesktop.org/CodeNames.html#NV40 a NV46, alias for G72. https://nouveau.freedesktop.org/FeatureMatrix.html It should be quite well supported. First would be not having crash with the console framebuffer :) How to debug further? Drivers are in the kernel right? A tedious option would be to somehow test linux, maybe with a LiveCD or with another hard disk. Riccardo [ 1540.510] (--) NV: Found NVIDIA GeForce Go 7300 at 01@01:00:0 [ 1540.510] (WW) VGA arbiter: cannot open kernel arbiter, no multi-card support [ 1540.510] (II) Loading sub module "int10" [ 1540.510] (II) LoadModule: "int10" [ 1540.511] (II) Loading /usr/X11R7/lib/modules/libint10.so [ 1540.532] (II) Module int10: vendor="X.Org Foundation" [ 1540.532] compiled for 1.20.5, module version = 1.0.0 [ 1540.532] ABI class: X.Org Video Driver, version 24.0 [ 1540.532] (II) NV(0): Initializing int10 [ 1540.533] (II) NV(0): Primary V_BIOS segment is: 0xc000 [ 1540.533] (--) NV(0): Chipset: "GeForce Go 7300" [ 1540.533] (==) NV(0): Depth 24, (--) framebuffer bpp 32 [ 1540.533] (==) NV(0): RGB weight 888 [ 1540.533] (==) NV(0): Default visual is TrueColor [ 1540.533] (II) Loading sub module "vgahw" [ 1540.533] (II) LoadModule: "vgahw" [ 1540.535] (II) Loading /usr/X11R7/lib/modules/libvgahw.so [ 1540.536] (II) Module vgahw: vendor="X.Org Foundation" [ 1540.536] compiled for 1.20.5, module version = 0.1.0 [ 1540.536] ABI class: X.Org Video Driver, version 24.0 [ 1540.536] (==) NV(0): Using HW cursor [ 1540.536] (--) NV(0): Linear framebuffer at 0xC000 [ 1540.536] (--) NV(0): MMIO registers at 0xDD00 [ 1540.542] (II) Loading sub module "i2c" <..> [ 1540.684] (II) Module "ramdac" already built-in [ 1540.988] (II) EXA(0): Offscreen pixmap area of 128443904 bytes [ 1540.988] (II) EXA(0): Driver registered support for the following operations: [ 1540.988] (II) Solid [ 1540.988] (II) Copy [ 1540.988] (II) UploadToScreen [ 1540.988] (II) DownloadFromScreen [ 1540.988] (==) NV(0): Backing store enabled [ 1540.988] (==) NV(0): Silken mouse enabled [ 1541.019] (==) NV(0): DPMS enabled
Re: switched from 9.99 to 9.3, init dies
Hi Martin, On 11/18/22 08:30, Martin Husemann wrote: Or the symlinks used for the versioning - I would suggest to upgrade to ahttps://nycdn.netbsd.org/pub/NetBSD-daily/HEAD/latest/ snapshot and start tracking netbsd-10 in ~two weeks after a week I replied to the wrong email sorry... I did not use the BIOS legacy image, but used the NetBSD-daily image as suggested here. Two different machines... both HP laptops... So I saved the upgrade one with the daily-snaphost, then upgraded pkgsrc, etc etc. Riccardo
Re: Upgrade 9.2->9.3 amd64 issues - device not configured
Hi Martin, On 11/10/22 07:45, Martin Husemann wrote: On Thu, Nov 10, 2022 at 04:30:18AM +0700, Robert Elz wrote: The error you saw is harmless (to you), there is nothing to fix. Not quite, the installer needs to detect this and deal with it - I will fix it (but not before next week, sorry). One immediate workaround would be to use the legacy bios USB boot image to do the update: https://nycdn.netbsd.org/pub/NetBSD-daily/netbsd-9/latest/images/NetBSD-9.3_STABLE-amd64-bios-install.img.gz this was a good idea... I installed it (well, a couple of days ago, took some time to test before writing this email( and it worked right out, the laptop is saved. Boot blocks updated too. I upgraded all pkgsrc from source, that took some compiling and worked being stable. X11, ethernet, wifi seem to work. Well WiFi has the same issues it had when working with Christos on it - needs some kick to come up, gives some errors. Suspend-to-RAM is also flakey with X11. Buit I will report on issues separately. Riccardo
switched from 9.99 to 9.3, init dies
Hi all, I "upgraded" (quotes intented) from a non-better-specified version of netbsd current of about April 2022 to NetBSD 9.3 I thought that I no longer had the time to compile and upgrade on my laptop so I switched. After the upgrade (which I believe is partially a downgrade, I don't know when 9.3 was forked) boot dies with "panic: init died" apparently in snprintf. If I boot "netbsd.old"... it dies too! Fine, so I suppose it is not the kernel, but something in the userland that is messed up. Ideas? Maybe I can "upgrade" again with a certain releng version - get back to some non-release stuff, but salvage the laptop and then maybe have a better way to "switch back to release"? Thanks, Riccardo
Re: Upgrade 9.2->9.3 amd64 issues - device not configured
Hi Robert. Robert Elz wrote: Unless there is something wrong, nothing.On a 9.x system you don't need any boot code updates with upgrades, so not having one done is harmless. The error you saw is harmless (to you), there is nothing to fix. perhaps harmless, but the installer appears to halt 1) upgrade 2) choose wd0 3) command fails -> continue 4) try mounting it anyway? if no, installer halts, which is what I always answered. if yes, i get another error on sd0e, then continue again and then indeed I was able to continue, install... and upgrade and get a usable, updated 9.3. Boot manager is correctly BIOS 9.3... all looks fine You must admit it is not exactly user friendly. Can you or Martin file a bug giving a description of what is happening? I tihnk I understood but not enough to write a bug report. Put me in CC, so I can follow it. about how 9.3 performs here, I'll report later. Thanks! Riccardo
Re: Upgrade 9.2->9.3 amd64 issues - device not configured
Hi Martin, Martin Husemann wrote: now what MS-DOS partition does it want to mess around and not finding? The EFI boot partition (to upgrade the bootloader) - this is a bug, apparently your hard disk uses BIOS booting, while the installer booted via UEFI from CD, and this confuses sysinst. seems I always put my foot in bugs... If I go into the StartUp menu on boot I go into BIOS, however there is a setting "UEFI Mode" which is unset. A warning appears that UEFI is only for development, not warranted or supported by HP. So I don't suppose I ever enabled it. Could it be that a setting changed when installing compared to upgrading? As a test, I enabled and booting drom HDD still says "BIOS Boot" The CDROM says then UEFI boot. Without the flag, it says "BIOS Boot". So could it be that while booting BIOS there are UEFI traces? What do you suggest to do now? Riccardo
Re: sysupgrade usage, expected file format
Hi Martin, Martin Husemann wrote: PS: what now? boot with CD-ROM and try to upgrade "over" again? Yes, that should be the easy way out (or USB install image instead of CD). one question, is sysupgrade uspposed to update the bootloader? I noticed the two upgrade systems are not fully equivalent in some details sysupgrade will show the sysmerge of all files and let you decide, while booted from CDROM I did not get the questions, just the auto post-install checks. Riccardo
9.3 sysupgrade on Sparc successful
Hi! as general news, proved that sysupgrade works... I used it to update my venerable SparcStation 4 (SPARC 32bit v8) as well as my Sun Netra (UltraSparc 64bit) and in both cases worked fine, very convenient since both are running headless, remote log-in, no need for serial console or such! A little bit of impatience when the SS4 took a bit to reboot and be accessible, but.. new kernel. Minimal testing shows functioning... stress test will follow and in case of issues, will be discussed separately for sparc! Another SPARC system will follow tomorrow. Good news sometimes, not just problems! Riccardo
Upgrade 9.2->9.3 amd64 issues - device not configured
Hi, part of the upgrading campaign, after partial success on x86... let's go to amd64. Candidate is an HP ProBook I boot from CD-ROM, select upgrade and the only HD which is "wd0", installation is straightforward. I get: Status: command failed Command: /sbin/fsck_msdos -o -q /dev/rsd0e Can't open 'dev/rsd0e' (device not configured) now what MS-DOS partition does it want to mess around and not finding? disklabel tells me that wd0 has two partitions: a is 4.2BSD and b is swap. Riccardo
Re: sysupgrade usage, expected file format
Hi! thanks, Martin, Bartosz! Martin Husemann wrote: No, there are no (known) binary compatibility issues. Are you sure you upgraded with sets from the same architecture (e.g. i386 vs. amd64)? unfortunately, that is exactly what happened. Bartosz spotted it. Since the first issue I had, I connected remotely to my laptop to write the email, then resuming, copy-pasted the wrong architecture :( PS: what now? boot with CD-ROM and try to upgrade "over" again? Yes, that should be the easy way out (or USB install image instead of CD). Luckily, on this R51 the cdrom usage was smooth (it is the same CDROM I am using with the other x86 laptops I reported in the other mail) and upgrading over... fixed everything! I add that I had another odd system left, which has no CD-ROM... and there sysugrade did run perfectly smooth on the first attempt. Other systems to upgrade remain :) And then of course usage will tell what 9.2 does. Riccardo
Re: sysupgrade usage, expected file format
Hi, Bartosz Maciejewski wrote: Hi, Please look at the bottom of sysupgrade.conf file:) #Archive file extension of the sets. For now it can be either 'tgz' or # 'tar.xz'. #ARCHIVE_EXTENSION=tar.xz found it :) used it and broke my system! argh! It indeed makes upgrade start. However, at some point I get flooded with: postinstall: Cannot execute ELF binary /bin/mkdir postinstall: Cannot execute ELF binary /bin/mkdir postinstall: Cannot execute ELF binary /bin/mkdir from another console I see that not even "ls" is executable anmore. Wow. I think somehow 9.3 is not binary compatible with 9.2 the userland gets extracted, the sysupgrade breaks. Am I unlucky on this i386 system or is it a general issue? I never upgraded with sysupgrade, but it would be so convenient for systems without CD-ROM to boot, servers, etc. Riccardo PS: what now? boot with CD-ROM and try to upgrade "over" again?
sysupgrade usage, expected file format
Hello, while upgrading various systems to 9.3, I wanted to test also the sysupgrade option. But how is it supposed to work? Can I use the sysupgrade installed in 9.2 and use it? On x86, I would have this version: sysupgrade-1.5nb10 However, if I try to use it this way: # sysupgrade auto http://cdn.netbsd.org/pub/NetBSD/NetBSD-9.3/amd64 it fails: sysupgrade: I: Starting auto-update with stages: fetch modules kernel sets etcupdate postinstall clean sysupgrade: I: Downloading http://cdn.netbsd.org/pub/NetBSD/NetBSD-9.3/amd64/binary/sets/base.tgz into /var/cache/sysupgrade Trying [2a04:4e42:39::262]:80 ... ftp: Can't connect to `2a04:4e42:39::262:80': No route to host Trying 146.75.53.6:80 ... Requesting http://cdn.netbsd.org/pub/NetBSD/NetBSD-9.3/amd64/binary/sets/base.tgz ftp: Error retrieving file `404 Not Found' sysupgrade: E: Failed to fetch http://cdn.netbsd.org/pub/NetBSD/NetBSD-9.3/amd64/binary/sets/base.tgz Took me a while, but then I found it: http://cdn.netbsd.org/pub/NetBSD/NetBSD-9.3/amd64/binary/sets/ all files are .xz now, not tgz! Riccardo
Upgrade 9.2->9.3 x86 issues - Part 1
Hi, I started upgrading my various computers to 9.3, starting from the most "standard" ones. Unfortunately my first two attempts already went havoc. Upgrade method chosen was to use the CD ISO. Candidate is a ThinkPad T30, running classic intel x86 32bit, running 9.2 with a very standard installation. I insert the CD, select upgrade and start the process. No big issues until postinstall, which I did not write down but appeared on minor packages, like fontconfig. I reboot and find the system partially upgraded, boot blocks are new, but the kernel is still old: netbsd 9.2. I find in the root the new kernel called netbsd where is a random suffix. Can I re-run postinstall from the "new system" like sysupgrade on OpenBSD and get a log of the error so that I can share them? The non-installed kernel is due to the installer not finishing postinstall and not replacing te kernel for safety, perhaps? Second candidate, still x86, is an Acer Aspire. This is also very standard, except I had a second kernel on the HDD but with a custom name, so that it could boot without crashing. This notebook has an nvidia card and it crashes. Unfortunately it still happens - I'll report separate on this. So I have to boot the cdrom with "-c", disable nouveau and continue installation. Here I have a lot of strange issues, first attempt I got up to downloading the kernel set, then got a hardware error. I reboot with a second attempt, I get an error in the ethernet autoconfiguration I/O error. If I try a reboot from menu, I get a segfault Either the kernel is unstable on this system, or the HW broke. I will stress-test it a little running 9.2 and report back. Third candidate is another ThinkPad, T43, which is extremely stable on 9.2. I boot the CD without issues, but while downloading kern-GENERIC.tgz I get an issue again with cd0 which says cd removal requested, sense error and fails. I try to reboot from menu, get a segfault. This looks similar the Acer Aspire! I made a second attempt. This time I got as far as postinstall and got these failed postinstall fixes: gid, x11, uid gid and uid both fail becasue _nsd, _unbound, nvmm missing. I wonder why this was not automated? Without a log, I don't know what failed in x11. In this case, however, reboot completed, system boots and pkgin upgraded all packages! X11 & Firefox work... so most looks fine. Suggestions welcome. I have other machines to install and will report separately. Cheers, Riccardo PS: summary: Thinkad T30 - i386 - CD: postinstall fails, kernel not installed Thinkpad T43 - i386 - CD: first attempt HW issues, second attempt postinstall failures, but works Acer Aspire - i386 - CD: fails boot for nvidia nouveau if enabled, fails for HW issues during install
Re: Arcticfox
Hello nia, nia wrote: >> It has a couple specific issues on NetBSD regarding WebGL, but I do >> relularly test it to build on NetBSD i386, amd64. >> It also builds on Sparc64, but will crash on startup (as on Linux, so >> not specific to NetBSD. > Do you know why it installs so many header files? actually no, never checked. Swhich header files? does PaleMoon and/or Firefox 52 do that? In case maybe I broke something or maybe something can be upgraded Riccardo
Re: Arcticfox
Hi, nia wrote: > greetz, > > I just packaged up arcticfox in pkgsrc and it seems to run well on > NetBSD (enough to watch youtube). It's a fork of Firefox with tons > of backported fixes that aims to target platforms like powerpc and > mips. Oh, that's nice to hear, being one of the main authors behind it... cool that there is some interest- It has a couple specific issues on NetBSD regarding WebGL, but I do relularly test it to build on NetBSD i386, amd64. It also builds on Sparc64, but will crash on startup (as on Linux, so not specific to NetBSD. > If anyone wants to try building it on some "unusual" NetBSD platforms, > I'd love to hear from you. I'd love to hear if it compiles and works on arm and PPC 32 or 64, since on Linux it compiles and runs on these. Anybody up to MIPS32 LE ? It used to run on Linux, but does no longer for me. Hints & help welcome. Riccardo
Synaptics button - movement issues
Hello, I have a laptop with a Synaptics touchpad which has "integrated" buttons, that is the lower part of the touchpad has buttons, but it also makes part of the scroll surface. I have issues that when clicking on the buttons, the touch movement remains active and makes the cursor jump around. This makes it very had to click on something and impossible to click-drag on a menu or context-menu! Can this be "fixed" with some sysctl setting? hw.synaptics.up_down_emulation = 3 hw.synaptics.up_down_motion_delta = 1 hw.synaptics.gesture_move = 200 hw.synaptics.gesture_length = 20 hw.synaptics.edge_left = 1632 hw.synaptics.edge_right = 5312 hw.synaptics.edge_top = 4288 hw.synaptics.edge_bottom = 1568 hw.synaptics.edge_motion_delta = 32 hw.synaptics.finger_high = 35 hw.synaptics.finger_low = 20 hw.synaptics.two_fingers_emulation = 0 hw.synaptics.scale_x = 16 hw.synaptics.scale_y = 16 hw.synaptics.scale_z = 32 hw.synaptics.max_speed_x = 32 hw.synaptics.max_speed_y = 32 hw.synaptics.max_speed_z = 2 hw.synaptics.movement_threshold = 4 hw.synaptics.movement_enable = 1 hw.synaptics.button_boundary = 2288 hw.synaptics.button2_edge = 2858 hw.synaptics.button3_edge = 4085 hw.synaptics.finger_scroll-min = 13 hw.synaptics.finger_scroll-max = 14 hw.synaptics.finger_scroll-hysteresis = 30 hw.synaptics.aux_mid_button_scroll = 1 I have not seen somee specific setting, but maybe the button boundaris/edges control this behavioiur and are detected wrong? Riccardo
pkg db fixing split brain situation
Hi all, I noticed by chance that I have a "split brain" situation when I couldnpt find pkg_install installing pkgin! I am following this guide: http://mail-index.netbsd.org/pkgsrc-users/2020/12/30/msg032987.html In that part I am up step 7 - manual preparations. Then I launch the check script, but I get: # sh _NetBSD-pkgdb check FATAL: mismatch /usr/pkg/pkgdb directory pair what is this error? and in case how can I fix it... Riccardo
Issues with wired and wifi cards
Hi, I bite the bullet and upgraded my "vintage" Toshiba Portégé from NetBSD 8.x to 9.2. Most things seem fine. My issue are network cards. The laptop has none internally, only a very small hard-to-access USB port, so the choice is PCMCIA/CardBus of which there are two slots. I don't remember which cards I used, I have a bunch of "pool card". I have issues with WiFi cards The bwi0 card required firmware. I got it according to the manpage, untarred it and put it here: portege# ls /libdata/firmware/bwi/v3/ a0g0bsinitvals2.fw a0g0initvals5.fw b0g0bsinitvals2.fw b0g0initvals5.fw ucode11.fw ucode5.fw a0g0bsinitvals5.fw a0g1bsinitvals5.fw b0g0bsinitvals5.fw pcm4.fw ucode2.fw a0g0initvals2.fw a0g1initvals5.fw b0g0initvals2.fw pcm5.fw ucode4.fw Does it look correct? [ 103.173212] bwi0 at cardbus0 function 0: Broadcom Wireless [ 103.173212] bwi0: BBP id 0x4318, BBP rev 0x2, BBP pkg 2 [ 103.173212] bwi0: MAC: rev 9 [ 103.183220] bwi0: PHY type 2, rev 7, ver 3 [ 103.183220] bwi0: RF manu 0x17f, type 0x2050, rev 8 [ 103.193214] bwi0: autoconfiguration error: invalid antenna gain in sprom [ 103.193214] bwi0: 11b rates: 1Mbps 2Mbps 5.5Mbps 11Mbps [ 103.193214] bwi0: 11g rates: 1Mbps 2Mbps 5.5Mbps 11Mbps 6Mbps 9Mbps 12Mbps 18Mbps 24Mbps 36Mbps 48Mbps 54Mbps When I try to pull up the card I get this: [ 111.763344] bwi0: autoconfiguration error: firmware image v3/ucode5.fw size mismatch, fw f139, real 5700 Then I have this: [ 257.405587] (manufacturer 0x0, product 0x0) Texas Instruments ACX111 802.11b/g (miscellaneous network) at cardbus1 function 0 not configured I guess this is just not supported under NetBSD - bad luck :) Then this one: [ 268.535753] ral0 at cardbus1 function 0: Ralink Technologies RT2561 802.11b/g (rev. 0x00) [ 268.545762] ral0: 802.11 address 00:22:b0:64:44:06 [ 268.545762] ral0: MAC/BBP RT2561C, RF RT2527 [ 268.545762] ral0: 11b rates: 1Mbps 2Mbps 5.5Mbps 11Mbps [ 268.545762] ral0: 11g rates: 1Mbps 2Mbps 5.5Mbps 11Mbps 6Mbps 9Mbps 12Mbps 18Mbps 24Mbps 36Mbps 48Mbps 54Mbps [ 461.768722] ral0: detached It looks fine, I configure the ssid, but it never connects. I see no evident errors And even worse with Wired cards... I only got one working (which I surely did not use in the past, it is "fixed" in another laptop) [ 141.523801] ne2 at pcmcia1 function 0: [ 141.523801] ne2: Ethernet address 00:13:46:34:0d:62 [ 141.523801] nsphyter0 at ne2 phy 1: DP83815/DP83846A 10/100 media interface, rev. 0 [ 141.533805] nsphyter0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto [ 158.554063] ne2: device timeout [ 164.554155] ne2: device timeout [ 168.554216] ne2: device timeout [ 174.554309] ne2: device timeout [ 176.554339] ne2: device timeout [ 179.554385] ne2: device timeout [ 181.554416] ne2: device timeout [ 197.624663] nsphyter0: detached [ 197.624663] ne2: detached This one makes me really wonder! If I put it in the upper slot, I see the light blinking for 100Mbit (actually - it is connected to a 1Gbit switch, but the card is 100) and then when I run dhcp it either gets no address, or gets an address but then doesn't work. The many device timeouts come from these attempts. If putin the lower slot, I get no green light at all. Issue with the lower slot? No. If I insert an old and trusty Toshiba card which is double-size and fits only in the lower slot: [ 482.421145] tlp0 at cardbus0 function 0: Xircom X3201-3 Ethernet, pass 0.3 [ 482.421145] tlp0: Ethernet address 00:10:a4:9f:1c:21 [ 482.421145] tqphy0 at tlp0 phy 0: 78Q2120 10/100 media interface, rev. 11 [ 482.421145] tqphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto And yay it connects without issues! Riccardo. PS: I will test these cards on another NetBSD laptop and then also on an OpenBSD laptop and see...
nouveau driver & installation hang with G72/NV46
Hi all, as written some times, I have issues with nVidia (pesky cards!) - NetBSD crashes as soon as initializes the framebuffer. I asked for help several times... up to 9.1... I just upgraded to 9.2 and had a bad experience: the bood kernel of the CD fails and hangs! I think this is really a bad experience. Since I debugged this a lot, I quickly knew "boot netbsd -c" and "disable nouveau" to get the classic text menus. A very nice thing would be of course to have this issue fixed for "my" G72 nVidia card [*]; but more broadly, this could happen for other cards. I see three solutions 1) disable drm framebuffers in install kernels totally (OpenBSD does this I think) 2) have a "whitelist" approach, what works, enabled, else... blocked. At every confirmed card and/or fixed bug, add it to the list 3) have a "blacklist" approach... at every bug, block one, if bug solved, remove it What do you think? We should not scare users away with a crashing boot. I found this related bug (another card than mine): http://gnats.netbsd.org/cgi-bin/query-pr-single.pl?number=53492 If someone wants to hack, I can as I did in the past share logs, try thing, disable stuff in kernel... debugging is the fun part - but I propose also a broader solution for newcomers. Riccardo [*] The chip should be supported - in nouveau speak it is a NV46 https://linux.die.net/man/4/nouveau https://nouveau.freedesktop.org/FeatureMatrix.html
Re: OpenGL issues - glxgears failure or segfault
Hello, no clue about this? I noticed that the numbers with the hex-value can go on forever. with a fresh reboot and no X11, I can read this information: [ 1.053049] agp0 at pchb0: G4X-family chipset [ 1.053049] agp0: detected 32252k stolen memory [ 1.053049] agp0: aperture at 0xc000, size 0x1000 [ 1.053049] i915drmkms0 at pci0 dev 2 function 0: vendor 8086 product 0046 (rev. 0x02) [ 1.053049] vendor 8086 product 3b64 (miscellaneous communications, revision 0x06) at pci0 dev 22 function 0 not configured <...> [ 5.269417] kern.module.path=/stand/amd64/9.2/modules [ 5.269417] kern info: [drm] Memory usable by graphics device = 512M [ 5.269417] kern info: [drm] Supports vblank timestamp caching Rev 2 (21.10.2013). [ 5.269417] kern info: [drm] Driver supports precise vblank timestamp query. [ 5.269417] i915drmkms0: interrupting at ioapic0 pin 16 (i915drmkms0) [ 5.319449] intelfb0 at i915drmkms0 [ 5.319449] intelfb0: framebuffer at 0xb8806507, size 1366x768, depth 32, stride 5504 Thanks, Riccardo Riccardo Mottola wrote: Hi, I traced down my WebGL problems to core OpenGL issues. If I run glxgears I get bad results. The display does not show rotating gears but either two images vibrating between (strobo-effect like) or even just one frozen image. Sometimes I get a crash, like this: $ glxgears Running synchronized to the vertical refresh. The framerate should be approximately the same as the monitor refresh rate. i965: Failed to submit batchbuffer: Input/output error This is on a laptop with integrated Iron Lake graphics running NetBSD/amd64. I just did check dmesg to type out the details here, but see it flooded with this errors: [ 128.377135] 0001327c : [ 128.377135] 00013280 : 7a006002 [ 128.377135] 00013284 : 00024104 [ 128.377135] 00013288 : [ 128.377135] 0001328c : [ 128.377135] 00013290 : 7a006002 [ 128.377135] 00013294 : 00024184 [ 128.377135] 00013298 : [ 128.377135] 0001329c : [ 128.377135] 000132a0 : 7a006002 [ 128.377135] 000132a4 : 00024204 [ 128.377135] 000132a8 : [ 128.377135] 000132ac : [ 128.377135] 000132b0 : 7a006002 [ 128.377135] 000132b4 : 00024284 [ 128.377135] 000132b8 : [ 128.377135] 000132bc : [ 128.377135] 000132c0 : 7a006002 [ 128.387142] 000132c4 : 00024304 [ 128.387142] 000132c8 : [ 128.387142] 000132cc : [ 128.387142] 000132d0 : 7a005502 [ 128.387142] 000132d4 : 00024004 [ 128.387142] 000132d8 : f204 [ 128.387142] 000132dc : [ 128.387142] 000132e0 : 0222 [ 128.387142] 000132e4 : [ 128.387142] 000132e8 : 18800180 [ 128.387142] 000132ec : 1f9d9000 [ 128.387142] 000132f0 : 0200 [ 128.387142] 000132f4 : [ 128.387142] 000132f8 : 7a005402 [ 128.387142] 000132fc : 00024004 [ 128.387142] 00013300 : f205 [ 128.387142] 00013304 : [ 128.387142] 00013308 : 7a006002 [ 128.387142] 0001330c : 00024084 [ 128.387142] 00013310 : [ 128.387142] 00013314 : [ 128.387142] 00013318 : 7a006002 [ 128.387142] 0001331c : 000 [ 8185.721137] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/i915/intel_sprite.c:132)intel_pipe_update_start] *ERROR* Potential atomic update failure on pipe A: -35 [ 24956.434454] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/i915/intel_sprite.c:132)intel_pipe_update_start] *ERROR* Potential atomic update failure on pipe A: -35 [ 25038.667770] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/i915/intel_sprite.c:132)intel_pipe_update_start] *ERROR* Potential atomic update failure on pipe A: -35 [ 25866.454467] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/i915/intel_sprite.c:132)intel_pipe_update_start] *ERROR* Potential atomic update failure on pipe A: -35 now, If I rerun glxgears, I get a segfault: $ glxgears Running synchronized to the vertical refresh. The framerate should be approximately the same as the monitor refresh rate. [1] Segmentation fault (core dumped) glxgears Thread 1 "" received signal SIGSEGV, Segmentation fault. 0x717487fe5850 in brw_workaround_depthstencil_alignment () from /usr/X11R7/lib/modules/dri/i965_dri.so (gdb) bt #0 0x717487fe5850 in brw_workaround_depthstencil_alignment () from /usr/X11R7/lib/modules/dri/i965_dri.so #1 0x717487fa34d5 in ?? () from /usr/X11R7/lib/modules/dri/i965_dri.so #2 0x00017f001f8a in draw () #3 0x00017f00353d in main () doesn't look nice!! Riccardo
OpenGL issues - glxgears failure or segfault
Hi, I traced down my WebGL problems to core OpenGL issues. If I run glxgears I get bad results. The display does not show rotating gears but either two images vibrating between (strobo-effect like) or even just one frozen image. Sometimes I get a crash, like this: $ glxgears Running synchronized to the vertical refresh. The framerate should be approximately the same as the monitor refresh rate. i965: Failed to submit batchbuffer: Input/output error This is on a laptop with integrated Iron Lake graphics running NetBSD/amd64. I just did check dmesg to type out the details here, but see it flooded with this errors: [ 128.377135] 0001327c : [ 128.377135] 00013280 : 7a006002 [ 128.377135] 00013284 : 00024104 [ 128.377135] 00013288 : [ 128.377135] 0001328c : [ 128.377135] 00013290 : 7a006002 [ 128.377135] 00013294 : 00024184 [ 128.377135] 00013298 : [ 128.377135] 0001329c : [ 128.377135] 000132a0 : 7a006002 [ 128.377135] 000132a4 : 00024204 [ 128.377135] 000132a8 : [ 128.377135] 000132ac : [ 128.377135] 000132b0 : 7a006002 [ 128.377135] 000132b4 : 00024284 [ 128.377135] 000132b8 : [ 128.377135] 000132bc : [ 128.377135] 000132c0 : 7a006002 [ 128.387142] 000132c4 : 00024304 [ 128.387142] 000132c8 : [ 128.387142] 000132cc : [ 128.387142] 000132d0 : 7a005502 [ 128.387142] 000132d4 : 00024004 [ 128.387142] 000132d8 : f204 [ 128.387142] 000132dc : [ 128.387142] 000132e0 : 0222 [ 128.387142] 000132e4 : [ 128.387142] 000132e8 : 18800180 [ 128.387142] 000132ec : 1f9d9000 [ 128.387142] 000132f0 : 0200 [ 128.387142] 000132f4 : [ 128.387142] 000132f8 : 7a005402 [ 128.387142] 000132fc : 00024004 [ 128.387142] 00013300 : f205 [ 128.387142] 00013304 : [ 128.387142] 00013308 : 7a006002 [ 128.387142] 0001330c : 00024084 [ 128.387142] 00013310 : [ 128.387142] 00013314 : [ 128.387142] 00013318 : 7a006002 [ 128.387142] 0001331c : 000 [ 8185.721137] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/i915/intel_sprite.c:132)intel_pipe_update_start] *ERROR* Potential atomic update failure on pipe A: -35 [ 24956.434454] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/i915/intel_sprite.c:132)intel_pipe_update_start] *ERROR* Potential atomic update failure on pipe A: -35 [ 25038.667770] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/i915/intel_sprite.c:132)intel_pipe_update_start] *ERROR* Potential atomic update failure on pipe A: -35 [ 25866.454467] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/i915/intel_sprite.c:132)intel_pipe_update_start] *ERROR* Potential atomic update failure on pipe A: -35 now, If I rerun glxgears, I get a segfault: $ glxgears Running synchronized to the vertical refresh. The framerate should be approximately the same as the monitor refresh rate. [1] Segmentation fault (core dumped) glxgears Thread 1 "" received signal SIGSEGV, Segmentation fault. 0x717487fe5850 in brw_workaround_depthstencil_alignment () from /usr/X11R7/lib/modules/dri/i965_dri.so (gdb) bt #0 0x717487fe5850 in brw_workaround_depthstencil_alignment () from /usr/X11R7/lib/modules/dri/i965_dri.so #1 0x717487fa34d5 in ?? () from /usr/X11R7/lib/modules/dri/i965_dri.so #2 0x00017f001f8a in draw () #3 0x00017f00353d in main () doesn't look nice!! Riccardo
Re: OpenGL - browser and WebGL support - failed libGL.so
Hi, Martin Neitzel wrote: The convention is that the major number reflects the shared lib's API, while minor numbers are used for bug fixes and internal improvements. *IF* the GL folks have taken care to keep their API downwards-compatible, you can safely ln -s libGL.so.3 libGL.so.1 did that, but I still get an initialization error, nothing changed. So I suppose there is either an issue which FF88 works around or instead some other issue older codebases like SeaMonkey and ArcticFox have. I added in ArcticFox a fix that allows to use of libGL.so without number. I also tried the link trick must investigate more. Riccardo
Re: OpenGL - browser and WebGL support - failed libGL.so
Hello. Chavdar Ivanov wrote: FWIW WebGL works for me with the latest Firefox (both using my laptop with Intel 530 graphics and also using the software renderer under VirtualBox). From time to time some firefox version would behave somewhat strangely - i.e. the demos from webglsamples.org would work, buthttps://cad.onshape.com/check would complain of too slow updates and refuse to check; right now firefox 91 works ok for both. I have FF 88 Nightly and it doesn't work. Would you mind trying out seamonkey too? Does glxgears run fine for you? displays fine too? However, I also noticed today that I have issues with glxgears. So I fear I have two types of issues: SeaMonkey and ArcticFox do not correctly detect, initialize or use my GL libraries. Generally however, GL support is shakey for my videocard (I always had various issues with intel cards, but being older/slower computers I never bothered much with GL or even less WebGL. I'll ask a separate question about this browser- independent issue. Thanks, RIccardo
Re: OpenGL - browser and WebGL support - failed libGL.so
Hi Patrick,
On 2021-08-23 15:04:59 +0200 Patrick Welche wrote:
> On Mon, Aug 23, 2021 at 02:46:39PM +0200, Riccardo Mottola wrote:
>> GLXtest process failed (exited with status 1): Unable to load
>> libGL.so.1
>
> This reminds me of e.g., libepoxy hardcoding "libGL.so.1", when
>
> $ ls /usr/X11R7/lib/libGL.*
> /usr/X11R7/lib/libGL.a /usr/X11R7/lib/libGL.so.3
> /usr/X11R7/lib/libGL.so /usr/X11R7/lib/libGL.so.3.0
>
> (Why hardcode the major number?)
I don't know, but you gave me a hint and grepping found this:
// see e.g. bug 608526: it is intrinsically interesting to know whether
we have dynamically linked to libGL.so.1
// because at least the NVIDIA implementation requires an executable
stack, which causes mprotect calls,
// which trigger glibc bug
http://sourceware.org/bugzilla/show_bug.cgi?id=12225
#ifdef __OpenBSD__
libGLfilename = "libGL.so";
#else
libGLfilename = "libGL.so.1";
#endif
mOGLLibrary = PR_LoadLibrary(libGLfilename);
if (!mOGLLibrary) {
NS_WARNING("Couldn't load OpenGL shared library.");
return false;
}
at least, this is ArcticFox. I will try making NetBSD as OpenBSD and see what
happens.
I wonder if SeaMonkey, also based on older Gecko code but a little bit newer,
has a similar issue.
I will also check current FireFox code to see what they did/changed there and
see.
Thanks,
Riccardo
--
Sent with GNUMail from iMac PowerPC 64bit running GNUstep on Debian/PPC
OpenGL - browser and WebGL support - failed libGL.so
Hi, I am struggling to have WebGL support in browsers on NetBSD. First, I hope the drivers support my Intel graphics and that permissions in groups are set up correctly, etc. glxgears starts up, that's the first minimal tests I know. It doesn't display nice rotating gears though but at about 60fps they vibrate between two images... so not a good start. Then I go to run ArcticFox, the derivate browser I work myself on, it refuses to run WebGL (it does work however on several Linux systems). about:troubleshooting shows: GLXtest process failed (exited with status 1): Unable to load libGL.so.1 Then I test seamonkey as it comes from pkgin: The diagnostics is more advanced and says: GPU #1 Active Yes Description Intel Open Source Technology Center -- Mesa DRI Intel(R) Ironlake Mobile Vendor ID Intel Open Source Technology Center Device ID Mesa DRI Intel(R) Ironlake Mobile Driver Version 2.1 Mesa 18.3.4 However then: WebGL 1 Driver Renderer: WebGL creation failed: * Error during native OpenGL init. * Error during native OpenGL init. * Error during native OpenGL init. * Error during native OpenGL init. * Error during native OpenGL init. * Exhausted GL driver caps. * Exhausted GL driver options. Firefox 88 instead behaves differently. Apparently it loads and enables WebGL but when I run a sample (e.g. https://webglsamples.org/dynamic-cubemap/dynamic-cubemap.html) starts to load objects but doesn't display textures and the tab crashes (can be repeated ad infinitum), How does it do for you? Any suggestions? Riccardo
Xorg and intel graphics - instant crash: *ERROR* Potential atomic update failure on pipe A: -35
Hi, I was able to reproduce several times an immediate crash of Xorg. I can open xterm... and if I want I can open many. But if I open xterm and then do "Launch" from windowmaker's menu on xterm (which usually just launches another xterm) I get an Xorg crash. I relaunched Xorg and got the crash again. In dmesg I see, however I am not sure this is the immediate cause of the crash - I have seen this error in dmesg without Xorg crashing. [ 1379.692811] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/i915/intel_sprite.c:132)intel_pipe_update_start] *ERROR* Potential atomic update failure on pipe A: -35 [ 2496.820019] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/i915/intel_sprite.c:132)intel_pipe_update_start] *ERROR* Potential atomic update failure on pipe A: -35 I smell a bug here... Intel graphics has issues, I noticed them on other systems too, but this I never had. [ 6.247928] i915drmkms0: interrupting at ioapic0 pin 16 (i915drmkms0) [ 6.297957] intelfb0 at i915drmkms0 [ 6.297957] intelfb0: framebuffer at 0x8d806507, size 1366x768, depth 32, stride 5504 Riccardo
Re: 9.2: pkgin packages do not register pkg_info
Hi. replying to myself Riccardo Mottola wrote: > I fear this "dirs" come from this step: > > In /var/db/pkg.refcount > tar cf - . | (cd /usr/pkg/pkgdb && tar xfv -) > > since pkg.refcount contained "dirs" > > ls /usr/pkg/pkgdb/dirs/ > etc usr var > > is it out of place? the proper command should have been: In /var/db/pkg.refcount tar cf - . | (cd /usr/pkg/pkgdb.refcount && tar xfv -) otherwise one mixes up with th other files. Riccard
Re: 9.2: pkgin packages do not register pkg_info
Hi Greg, Greg Troxel wrote: > I don't either, but my advice is to *always* set PKG_DBDIR explicitly in > btoh mk.conf and pkg_install.conf. Check if you have the split brain > situation, or something else: > > https://pkgsrc.org/pkgdb-change/ oh, I fear I have split. But, permit me, this is bad: I have a super-fresh install; I used pkg_add to instappl pkgin and nothing more! So what went wrong? localhost$ /usr/sbin/pkg_add -V 20201218 localhost$ /usr/pkg/sbin/pkg_add -V 20210410 The versions are exactly the two which are cited as "good"! I will follow the script... but it is bad to land this way without actually "doing" an upgrade... should we migrate essentially wth the script a clean 9.2 install? in /usr/pkg/pkgdb I essentially only have pkgin -> the one which I added with pkg_add in /var/db/pkg I have all the packages installed with the said pkgin! Riccardo
Re: 9.2: pkgin packages do not register pkg_info
Hi Greg Greg Troxel wrote: > I don't either, but my advice is to *always* set PKG_DBDIR explicitly in > btoh mk.conf and pkg_install.conf. Check if you have the split brain > situation, or something else: > > https://pkgsrc.org/pkgdb-change/ I tried to perform the split migration according to the cited mail http://mail-index.netbsd.org/pkgsrc-users/2020/12/30/msg032987.html I got some warnings and I don't think it went totally well: localhost# pkg_admin check .pkg_admin: can't open /usr/pkg/pkgdb/dirs/+CONTENTS: No such file or directory localhost# pkg_admin rebuild-tree pkg_admin: Cannot read +CONTENTS of package dirs I fear this "dirs" come from this step: In /var/db/pkg.refcount tar cf - . | (cd /usr/pkg/pkgdb && tar xfv -) since pkg.refcount contained "dirs" ls /usr/pkg/pkgdb/dirs/ etc usr var is it out of place? Riccardo
9.2: pkgin packages do not register pkg_info
Hello, I just installed a fresh 9.2 on an HP ProBook amd64. It went quite smooth! (FreeBSD didn't even boot with SATA in AHCI mode...) Yay. Wireless worked out of the box, X11 too (well Intel graphics is a good bet there.) I installed using pkg_add pkgin. Then I installed a myriad of packages using "pkgin install", so to be quick and not compile stuff myself. I wonder however that "pkg_info" then does not list these packages, why? $ pkg_info pkg_install-20210410 Package management and administration tools for pkgsrc pkgin-20.12.1nb1 Apt / yum like tool for managing pkgsrc binary packages Yet I do know that I installed things like windowmaker, autoconf, etc etc and /usr/pkg/bin is full Where are they? Riccardo
Re: Recommended web browser?
Hi, nottobay wrote: Are there any properly supported browsers? I'm fine with compiling one I just want one that works with cites like YouTube. SeaMonkey & FireFox. You don't give us information about the specs of your system However, if you dare compiling you may try ArcticFox - I try my best to keep it running on NetBSD and it does! Especially if you have an older spec machines you may like it, I have it running on my T30 ThinkPad Pentium-4 with NetBSD (or appreciate speed on a Xeon...). Runs nice also on a RaspberryPI 3 or 4, but I did test only Linux for now. Riccardo
Display goes havoc (Radeon IRQ no handler installed)
Hi, It happened to me several times: I use a laptop for hours, also connecting remotely to it, and it works then suddenly the display goes completely out of sync! Not only in X11, but also consoles are affected. One see garbage (barely recognizing windows, text) as with old monitors loosing the sync signal. I'm running 9.1 on a genuine amd64. Usually I just had to reboot (power button respons, the computer is not locked, as proven when changing consoles). This time I was remotely logged in. So before rebooting I checked dmesg and saw this: [ 5824.943995] warning: /usr/src/sys/external/bsd/drm2/dist/drm/radeon/radeon_r100.c:714: Can't enable IRQ/MSI because no handler is installed [ 5824.954001] warning: /usr/src/sys/external/bsd/drm2/dist/drm/radeon/radeon_r100.c:714: Can't enable IRQ/MSI because no handler is installed [ 10303.493972] warning: /usr/src/sys/external/bsd/drm2/dist/drm/radeon/radeon_r100.c:714: Can't enable IRQ/MSI because no handler is installed [ 10303.503979] warning: /usr/src/sys/external/bsd/drm2/dist/drm/radeon/radeon_r100.c:714: Can't enable IRQ/MSI because no handler is installed [ 38899.246495] warning: /usr/src/sys/external/bsd/drm2/dist/drm/radeon/radeon_r100.c:714: Can't enable IRQ/MSI because no handler is installed [ 38899.256501] warning: /usr/src/sys/external/bsd/drm2/dist/drm/radeon/radeon_r100.c:714: Can't enable IRQ/MSI because no handler is installed [ 82921.420180] warning: /usr/src/sys/external/bsd/drm2/dist/drm/radeon/radeon_r100.c:714: Can't enable IRQ/MSI because no handler is installed [ 82921.430186] warning: /usr/src/sys/external/bsd/drm2/dist/drm/radeon/radeon_r100.c:714: Can't enable IRQ/MSI because no handler is installed [ 83667.286532] warning: /usr/src/sys/external/bsd/drm2/dist/drm/radeon/radeon_r100.c:714: Can't enable IRQ/MSI because no handler is installed [ 83667.296538] warning: /usr/src/sys/external/bsd/drm2/dist/drm/radeon/radeon_r100.c:714: Can't enable IRQ/MSI because no handler is installed [ 85715.930107] warning: /usr/src/sys/external/bsd/drm2/dist/drm/radeon/radeon_r100.c:714: Can't enable IRQ/MSI because no handler is installed [ 85715.940114] warning: /usr/src/sys/external/bsd/drm2/dist/drm/radeon/radeon_r100.c:714: Can't enable IRQ/MSI because no handler is installed It is not enocuraging, isn't it? The error is actually seen on boot already, but just once. Here my dmesg excerpt: [ 5.353875] kern.module.path=/stand/amd64/9.1/modules [ 5.353875] kern info: [drm] initializing kernel modesetting (RS480 0x1002:0x5955 0x103C:0x308B). [ 5.353875] kern info: [drm] register mmio base: 0xd030 [ 5.353875] kern info: [drm] register mmio size: 65536 [ 5.353875] kern info: [drm] Generation 2 PCI interface, using max accessible memory [ 5.353875] radeon0: info: VRAM: 64M 0x7C00 - 0x7FFF (64M used) [ 5.353875] radeon0: info: GTT: 512M 0x8000 - 0x9FFF [ 5.353875] kern info: [drm] Detected VRAM RAM=40M, BAR=128M [ 5.353875] kern info: [drm] RAM width 128bits DDR [ 5.353875] Zone kernel: Available graphics memory: 421182 kiB [ 5.353875] kern info: [drm] radeon: 64M of VRAM memory ready [ 5.353875] kern info: [drm] radeon: 512M of GTT memory ready. [ 5.353875] kern info: [drm] GART: num cpu pages 131072, num gpu pages 131072 [ 5.363887] kern info: [drm] radeon: 1 quad pipes, 1 z pipes initialized. [ 5.363887] kern info: [drm] PCIE GART of 512M enabled (table at 0x43D78000). [ 5.363887] radeon0: info: WB enabled [ 5.363887] radeon0: info: fence driver on ring 0 use gpu addr 0x8000 and cpu addr 0x0xbd07b5188000 [ 5.363887] kern info: [drm] Supports vblank timestamp caching Rev 2 (21.10.2013). [ 5.363887] kern info: [drm] Driver supports precise vblank timestamp query. [ 5.363887] radeon0: info: radeon: MSI limited to 32-bit [ 5.363887] radeon0: interrupting at ioapic0 pin 17 (radeon0) [ 5.363887] kern info: [drm] radeon: irq initialized. [ 5.363887] kern info: [drm] Loading R300 Microcode [ 5.423919] ubt0 at uhub0 port 2 [ 5.423919] ubt0: Broadcom Corp (0x3f0) HP Integrated Module (0x171d), rev 2.00/1.00, addr 2 [ 5.423919] kern info: [drm] radeon: ring at 0x80001000 [ 5.564005] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/radeon/radeon_r100.c:3701)r100_ring_test] *ERROR* radeon: ring test failed (scratch(0x15E4)=0xCAFEDEAD) [ 5.564005] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/radeon/radeon_r100.c:1232)r100_cp_init] *ERROR* radeon: cp isn't working (-22). [ 5.564005] radeon0: autoconfiguration error: error: failed initializing CP (-22). [ 5.564005] radeon0: autoconfiguration error: error: Disabling GPU acceleration [ 5.694085] kern error: [drm:(/usr/src/sys/external/bsd/drm2/dist/drm/radeon/radeon_r100.c:1252)r100_cp_fini] *ERROR* Wait for CP idle timeout, shutting down CP. [ 5.824166] kern warning: Failed to
NetBSD base gcc & libatomic
Hi, on x86-32, I noticed our compiler emits calls to atomic functions, but does not ship libatomic, which is available only as an "addon" in pkgsrc. I think the base compiler should be consistent and either ship libraries it uses or do not use. Gcc from pkgsrc then is free to pull libatomic in instead. What do you think? Riccardo
