Re: ACPI broke going from 8 to 9
On Sat, Dec 31, 2011 at 3:31 PM, Jeremy Chadwick wrote: > On Sat, Dec 31, 2011 at 04:17:16PM -0700, Dan Allen wrote: >> On 31 Dec 2011, at 12:34 PM, Garrett Cooper wrote: >> >> > Not yet. Add 'nooptions NEW_PCIB' to your KERNCONF, recompile, and >> > try booting the new kernel. See if this works. >> >> It worked! No hang, power button works. Nice. I hope this experimental >> option stays in. >> >> Thank you everyone for your help. Happy New Years! > > This option isn't documented **anywhere** in the entire src tree. It's > purely #ifdef all over. > > The code in question was committed 7 months ago. It was MFC'd to > RELENG_8 6 months ago. Here's the HEAD commit message: > > http://www.freebsd.org/cgi/cvsweb.cgi/src/sys/dev/pci/pci.c#rev1.420 > > The RELENG_8 MFC is revision 1.386.2.15. > > The committer is jhb@, with mav@ being the individual who tested it, so > I imagine either of these folks will have some excellent insights as to > what's causing Dan's problem. I'm CC'ing them both directly on this > thread. > > In the meantime: Dan, when you say in your original mail, "I just > upgraded my Dell OptiPlex GX270 from RELENG_8 to RELENG_9", can you > please provide uname -a output from the system when it was running > RELENG_8? I'm looking specifically for the exact time when the kernel > was built, because there may have been fixes (that broke things for you) > between the above commit and present-day RELENG_8 (I have not examined > all commits). It's going to be the feature that's going to cause headaches post-9.0-RELEASE based on my observations of several mailing list posts and the fact that 9.0 isn't actually RELEASEd yet (people have run into issues with acpi, atkbdc, mfi, and usb so far, but that's probably not everything). If it could be made into a runtime tunable, that would be awesome, but that would require changes to driver structures and methods. With a little pointer aliasing and tunable guard sprinkling it wouldn't be hard to solve -- but it's still work. In the meantime, could someone please commit PR # 163748 to note what NEW_PCIB is and MFC it to RELENG_9 and could we consider disabling NEW_PCIB on i386 and pc98 until all the issues are ironed out? Thanks, -Garrett ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: ACPI broke going from 8 to 9
On 31 December 2011 16:08, Dan Allen wrote: > Almost every day I csup from RELENG_x and build. The traces of RELENG_8 are > gone, so no, unfortunately I cannot give you a uname -a from those days. Would you consider having a small partition to do the same for HEAD? :P > However, I have a build log file, and I see that I moved from RELENG_8 to > RELENG_9 on Friday, Dec 23, 2011. I csup'd at 12:24:26 MST and discovered > the failure at 15:41 MST. > > This "nooptions NEW_PCIB" fix does seem rather tenuous if it is not > documented. Wouldn't a better route be something like > > if (ACPI < 2.0) > oldCode(); > else > newCodeForNewACPI(); > > so that it will always work for everyone without having to build a special > kernel? After all, I went from a working system to a hung system which is > not the best upgrade path... ;-) Well it's hard to test stuff out without the hardware. :) And it's quite possible a lot of silly looking issues are actually working around real bugs in the hardware. I'm glad this issue was solved so quickly for you. Let's hope we can get you onto testing out HEAD (in a separate partition!) so we can ensure we don't break the basic stuff. :) Adrian ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: ACPI broke going from 8 to 9
On 31 Dec 2011, at 4:31 PM, Jeremy Chadwick wrote: > In the meantime: Dan, when you say in your original mail, "I just > upgraded my Dell OptiPlex GX270 from RELENG_8 to RELENG_9", can you > please provide uname -a output from the system when it was running > RELENG_8? I'm looking specifically for the exact time when the kernel > was built Almost every day I csup from RELENG_x and build. The traces of RELENG_8 are gone, so no, unfortunately I cannot give you a uname -a from those days. However, I have a build log file, and I see that I moved from RELENG_8 to RELENG_9 on Friday, Dec 23, 2011. I csup'd at 12:24:26 MST and discovered the failure at 15:41 MST. This "nooptions NEW_PCIB" fix does seem rather tenuous if it is not documented. Wouldn't a better route be something like if (ACPI < 2.0) oldCode(); else newCodeForNewACPI(); so that it will always work for everyone without having to build a special kernel? After all, I went from a working system to a hung system which is not the best upgrade path... ;-) Dan ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: ACPI broke going from 8 to 9
On Sat, Dec 31, 2011 at 04:17:16PM -0700, Dan Allen wrote: > On 31 Dec 2011, at 12:34 PM, Garrett Cooper wrote: > > > Not yet. Add 'nooptions NEW_PCIB' to your KERNCONF, recompile, and > > try booting the new kernel. See if this works. > > It worked! No hang, power button works. Nice. I hope this experimental > option stays in. > > Thank you everyone for your help. Happy New Years! This option isn't documented **anywhere** in the entire src tree. It's purely #ifdef all over. The code in question was committed 7 months ago. It was MFC'd to RELENG_8 6 months ago. Here's the HEAD commit message: http://www.freebsd.org/cgi/cvsweb.cgi/src/sys/dev/pci/pci.c#rev1.420 The RELENG_8 MFC is revision 1.386.2.15. The committer is jhb@, with mav@ being the individual who tested it, so I imagine either of these folks will have some excellent insights as to what's causing Dan's problem. I'm CC'ing them both directly on this thread. In the meantime: Dan, when you say in your original mail, "I just upgraded my Dell OptiPlex GX270 from RELENG_8 to RELENG_9", can you please provide uname -a output from the system when it was running RELENG_8? I'm looking specifically for the exact time when the kernel was built, because there may have been fixes (that broke things for you) between the above commit and present-day RELENG_8 (I have not examined all commits). -- | Jeremy Chadwickjdc at parodius.com | | Parodius Networking http://www.parodius.com/ | | UNIX Systems Administrator Mountain View, CA, US | | Making life hard for others since 1977. PGP 4BD6C0CB | ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: ACPI broke going from 8 to 9
On 31 Dec 2011, at 12:34 PM, Garrett Cooper wrote: > Not yet. Add 'nooptions NEW_PCIB' to your KERNCONF, recompile, and > try booting the new kernel. See if this works. It worked! No hang, power button works. Nice. I hope this experimental option stays in. Thank you everyone for your help. Happy New Years! Dan ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: ACPI broke going from 8 to 9
On 31 Dec 2011, at 1:01 PM, Adrian Chadd wrote: > So what I can only suggest is that you build and boot a variety of > -HEAD kernels. Start with HEAD from say, Jan 1 2011. Boot it, see if > it works. If it doesn't, go back 3 months at a time. If it does, go > forward three months until it breaks. Fair enough. I will see what I can accomplish. Thanks! Dan ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: ACPI broke going from 8 to 9
On 31 December 2011 11:41, Dan Allen wrote: > My Toshiba Satellite U205 used to work great with RELENG_7, but the boot code > of RELENG_8 will not recognize the 2nd core of my Core Duo (not Core 2 Duo) > processor. Nobody seems to care as few machines have Core Duo, or few people > use this era of Toshiba BIOS, or whatever. .. and if this is a problem, a boot -v from both releng_7 and releng_8 would be a good starting point. Surely detecting the CPU cores would be an easy thing to get right. Also, it may be a hyperthread CPU "thing" - I recall some changes in the releng_8 timeframe because of security vulnerabilities with classic hyperthreading and you have to do something to enable it again. Adrian ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: ACPI broke going from 8 to 9
Well, the problem is that the people working on this code don't have a variety of older hardware to test things on. Developers of free software rely on users to do testing of releases on the hardware they care about. It may not sound very good but it's the best that can be done with the given resources :) This is why developers ask for testing to be done on -HEAD before it becomes a RC. Otherwise the alternative is to build a huge regression testing lab and have volunteers staff it.. guess what the chances are of that happening right at the present? :) So what I can only suggest is that you build and boot a variety of -HEAD kernels. Start with HEAD from say, Jan 1 2011. Boot it, see if it works. If it doesn't, go back 3 months at a time. If it does, go forward three months until it breaks. Post the SVN revision numbers of the kernel versions that work and don't work. You don't have to do anything other than boot the kernel to see if it works, so you don't need to try and build an entire release. Thankfully. If you can spend a few hours doing that, you'll be helping out the pci/cardbus/acpi guys a _lot_. Chances are that they updated something that looked wrong, and broke some legacy thing. This happened with Atheros NICs and caused no end of heartache until someone actually did the above. John fixed it quick-smart. :) Good luck! Adrian ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: ACPI broke going from 8 to 9
On 31 Dec 2011, at 10:57 AM, Jeremy Chadwick wrote: > Do you have a necessary reason to upgrade to 9 given this situation? > Given the conditions I would stay you should stay with 8. This philosophy seems wrong, but it may be the way to go. My Toshiba Satellite U205 used to work great with RELENG_7, but the boot code of RELENG_8 will not recognize the 2nd core of my Core Duo (not Core 2 Duo) processor. Nobody seems to care as few machines have Core Duo, or few people use this era of Toshiba BIOS, or whatever. Now my Dell GX270 ACPI code is pre 2.0 (so Garrett tells me), so RELENG_9 is out. I guess I should run all of my older machines on RELENG_7 but -- and this is where the philosophy you suggest seems wrong -- I still want the latest apps, security fixes, etc. If the stable tree updates ls or tcsh or awk, I want these, but the core OS seems to have moved on from 2004 machines. In other words, there is no tree for me. Dan ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
[releng_9 tinderbox] failure on amd64/amd64
TB --- 2011-12-31 16:02:35 - tinderbox 2.8 running on freebsd-stable.sentex.ca TB --- 2011-12-31 16:02:35 - starting RELENG_9 tinderbox run for amd64/amd64 TB --- 2011-12-31 16:02:35 - cleaning the object tree TB --- 2011-12-31 16:03:34 - cvsupping the source tree TB --- 2011-12-31 16:03:34 - /usr/bin/csup -z -r 3 -g -L 1 -h cvsup.sentex.ca /tinderbox/RELENG_9/amd64/amd64/supfile TB --- 2011-12-31 16:04:25 - building world TB --- 2011-12-31 16:04:25 - CROSS_BUILD_TESTING=YES TB --- 2011-12-31 16:04:25 - MAKEOBJDIRPREFIX=/obj TB --- 2011-12-31 16:04:25 - PATH=/usr/bin:/usr/sbin:/bin:/sbin TB --- 2011-12-31 16:04:25 - SRCCONF=/dev/null TB --- 2011-12-31 16:04:25 - TARGET=amd64 TB --- 2011-12-31 16:04:25 - TARGET_ARCH=amd64 TB --- 2011-12-31 16:04:25 - TZ=UTC TB --- 2011-12-31 16:04:25 - __MAKE_CONF=/dev/null TB --- 2011-12-31 16:04:25 - cd /src TB --- 2011-12-31 16:04:25 - /usr/bin/make -B buildworld >>> World build started on Sat Dec 31 16:04:26 UTC 2011 >>> Rebuilding the temporary build tree >>> stage 1.1: legacy release compatibility shims >>> stage 1.2: bootstrap tools >>> stage 2.1: cleaning up the object tree >>> stage 2.2: rebuilding the object tree >>> stage 2.3: build tools >>> stage 3: cross tools >>> stage 4.1: building includes >>> stage 4.2: building libraries >>> stage 4.3: make dependencies >>> stage 4.4: building everything >>> stage 5.1: building 32 bit shim libraries >>> World build completed on Sat Dec 31 18:49:47 UTC 2011 TB --- 2011-12-31 18:49:47 - generating LINT kernel config TB --- 2011-12-31 18:49:47 - cd /src/sys/amd64/conf TB --- 2011-12-31 18:49:47 - /usr/bin/make -B LINT TB --- 2011-12-31 18:49:47 - cd /src/sys/amd64/conf TB --- 2011-12-31 18:49:47 - /usr/sbin/config -m LINT TB --- 2011-12-31 18:49:47 - building LINT kernel TB --- 2011-12-31 18:49:47 - CROSS_BUILD_TESTING=YES TB --- 2011-12-31 18:49:47 - MAKEOBJDIRPREFIX=/obj TB --- 2011-12-31 18:49:47 - PATH=/usr/bin:/usr/sbin:/bin:/sbin TB --- 2011-12-31 18:49:47 - SRCCONF=/dev/null TB --- 2011-12-31 18:49:47 - TARGET=amd64 TB --- 2011-12-31 18:49:47 - TARGET_ARCH=amd64 TB --- 2011-12-31 18:49:47 - TZ=UTC TB --- 2011-12-31 18:49:47 - __MAKE_CONF=/dev/null TB --- 2011-12-31 18:49:47 - cd /src TB --- 2011-12-31 18:49:47 - /usr/bin/make -B buildkernel KERNCONF=LINT >>> Kernel build for LINT started on Sat Dec 31 18:49:47 UTC 2011 >>> stage 1: configuring the kernel >>> stage 2.1: cleaning up the object tree >>> stage 2.2: rebuilding the object tree >>> stage 2.3: build tools >>> stage 3.1: making dependencies >>> stage 3.2: building everything [...] /src/sys/dev/usb/net/if_axe.c:1355: error: 'AXE_FLAG_CSUM_FRAME' undeclared (first use in this function) /src/sys/dev/usb/net/if_axe.c:1357: error: 'struct axe_softc' has no member named 'sc_lenmask' /src/sys/dev/usb/net/if_axe.c:1357: error: 'AXE_CSUM_HDR_LEN_MASK' undeclared (first use in this function) /src/sys/dev/usb/net/if_axe.c:1359: error: 'struct axe_softc' has no member named 'sc_lenmask' /src/sys/dev/usb/net/if_axe.c:1359: error: 'AXE_HDR_LEN_MASK' undeclared (first use in this function) /src/sys/dev/usb/net/if_axe.c: In function 'axe_ioctl': /src/sys/dev/usb/net/if_axe.c:1499: warning: implicit declaration of function 'uether_init' /src/sys/dev/usb/net/if_axe.c:1499: warning: nested extern declaration of 'uether_init' [-Wnested-externs] *** Error code 1 Stop in /obj/amd64.amd64/src/sys/LINT. *** Error code 1 Stop in /src. *** Error code 1 Stop in /src. TB --- 2011-12-31 19:01:11 - WARNING: /usr/bin/make returned exit code 1 TB --- 2011-12-31 19:01:11 - ERROR: failed to build LINT kernel TB --- 2011-12-31 19:01:11 - 7918.41 user 1064.17 system 10715.99 real http://tinderbox.freebsd.org/tinderbox-releng_9-RELENG_9-amd64-amd64.full ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
[releng_9 tinderbox] failure on i386/i386
TB --- 2011-12-31 16:39:24 - tinderbox 2.8 running on freebsd-stable.sentex.ca TB --- 2011-12-31 16:39:24 - starting RELENG_9 tinderbox run for i386/i386 TB --- 2011-12-31 16:39:24 - cleaning the object tree TB --- 2011-12-31 16:40:20 - cvsupping the source tree TB --- 2011-12-31 16:40:20 - /usr/bin/csup -z -r 3 -g -L 1 -h cvsup.sentex.ca /tinderbox/RELENG_9/i386/i386/supfile TB --- 2011-12-31 16:41:04 - building world TB --- 2011-12-31 16:41:04 - CROSS_BUILD_TESTING=YES TB --- 2011-12-31 16:41:04 - MAKEOBJDIRPREFIX=/obj TB --- 2011-12-31 16:41:04 - PATH=/usr/bin:/usr/sbin:/bin:/sbin TB --- 2011-12-31 16:41:04 - SRCCONF=/dev/null TB --- 2011-12-31 16:41:04 - TARGET=i386 TB --- 2011-12-31 16:41:04 - TARGET_ARCH=i386 TB --- 2011-12-31 16:41:04 - TZ=UTC TB --- 2011-12-31 16:41:04 - __MAKE_CONF=/dev/null TB --- 2011-12-31 16:41:04 - cd /src TB --- 2011-12-31 16:41:04 - /usr/bin/make -B buildworld >>> World build started on Sat Dec 31 16:41:06 UTC 2011 >>> Rebuilding the temporary build tree >>> stage 1.1: legacy release compatibility shims >>> stage 1.2: bootstrap tools >>> stage 2.1: cleaning up the object tree >>> stage 2.2: rebuilding the object tree >>> stage 2.3: build tools >>> stage 3: cross tools >>> stage 4.1: building includes >>> stage 4.2: building libraries >>> stage 4.3: make dependencies >>> stage 4.4: building everything >>> World build completed on Sat Dec 31 18:49:29 UTC 2011 TB --- 2011-12-31 18:49:29 - generating LINT kernel config TB --- 2011-12-31 18:49:29 - cd /src/sys/i386/conf TB --- 2011-12-31 18:49:29 - /usr/bin/make -B LINT TB --- 2011-12-31 18:49:29 - cd /src/sys/i386/conf TB --- 2011-12-31 18:49:29 - /usr/sbin/config -m LINT TB --- 2011-12-31 18:49:29 - building LINT kernel TB --- 2011-12-31 18:49:29 - CROSS_BUILD_TESTING=YES TB --- 2011-12-31 18:49:29 - MAKEOBJDIRPREFIX=/obj TB --- 2011-12-31 18:49:29 - PATH=/usr/bin:/usr/sbin:/bin:/sbin TB --- 2011-12-31 18:49:29 - SRCCONF=/dev/null TB --- 2011-12-31 18:49:29 - TARGET=i386 TB --- 2011-12-31 18:49:29 - TARGET_ARCH=i386 TB --- 2011-12-31 18:49:29 - TZ=UTC TB --- 2011-12-31 18:49:29 - __MAKE_CONF=/dev/null TB --- 2011-12-31 18:49:29 - cd /src TB --- 2011-12-31 18:49:29 - /usr/bin/make -B buildkernel KERNCONF=LINT >>> Kernel build for LINT started on Sat Dec 31 18:49:29 UTC 2011 >>> stage 1: configuring the kernel >>> stage 2.1: cleaning up the object tree >>> stage 2.2: rebuilding the object tree >>> stage 2.3: build tools >>> stage 3.1: making dependencies >>> stage 3.2: building everything [...] /src/sys/dev/usb/net/if_axe.c:1355: error: 'AXE_FLAG_CSUM_FRAME' undeclared (first use in this function) /src/sys/dev/usb/net/if_axe.c:1357: error: 'struct axe_softc' has no member named 'sc_lenmask' /src/sys/dev/usb/net/if_axe.c:1357: error: 'AXE_CSUM_HDR_LEN_MASK' undeclared (first use in this function) /src/sys/dev/usb/net/if_axe.c:1359: error: 'struct axe_softc' has no member named 'sc_lenmask' /src/sys/dev/usb/net/if_axe.c:1359: error: 'AXE_HDR_LEN_MASK' undeclared (first use in this function) /src/sys/dev/usb/net/if_axe.c: In function 'axe_ioctl': /src/sys/dev/usb/net/if_axe.c:1499: warning: implicit declaration of function 'uether_init' /src/sys/dev/usb/net/if_axe.c:1499: warning: nested extern declaration of 'uether_init' [-Wnested-externs] *** Error code 1 Stop in /obj/i386.i386/src/sys/LINT. *** Error code 1 Stop in /src. *** Error code 1 Stop in /src. TB --- 2011-12-31 19:00:45 - WARNING: /usr/bin/make returned exit code 1 TB --- 2011-12-31 19:00:45 - ERROR: failed to build LINT kernel TB --- 2011-12-31 19:00:45 - 6568.91 user 801.92 system 8481.04 real http://tinderbox.freebsd.org/tinderbox-releng_9-RELENG_9-i386-i386.full ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: ACPI broke going from 8 to 9
On Sat, Dec 31, 2011 at 10:26:09AM -0700, Dan Allen wrote: > (** Originally posted to freebsd-curr...@freebsd.org but I noticed that there > are 9.0 RC3 questions here on freebsd-stable; I am not sure which forum is > appropriate. **) > > I just upgraded my Dell OptiPlex GX270 from RELENG_8 to RELENG_9. The > machine no longer boots. However, if I put > > hint.acpi.0.disabled=1 > > in /boot/loader.conf then the machine runs fine. With RELENG_8 the machine > had no loader.conf, and the power button worked on my desktop machine. Now > with ACPI disabled my power button does not work. I have found that the > machine hangs at boot during a scan of the PCI bus, but if I disable that > (hw.acpi.disable=pci) then the machine cannot find a boot drive. > > So I have lost functionality that worked fine in BSD 8. Do you have a necessary reason to upgrade to 9 given this situation? Given the conditions I would stay you should stay with 8. > Thoughts? Suggestions? Provide full output from a verbose boot so developers can try to figure out where the problem is. You will either need serial console or firewire console for this task, as the amount of text will be too much to type in (many, many screens worth). Simply showing a photo of the last screen where it locks up may not be sufficient. -- | Jeremy Chadwickjdc at parodius.com | | Parodius Networking http://www.parodius.com/ | | UNIX Systems Administrator Mountain View, CA, US | | Making life hard for others since 1977. PGP 4BD6C0CB | ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
ACPI broke going from 8 to 9
(** Originally posted to freebsd-curr...@freebsd.org but I noticed that there are 9.0 RC3 questions here on freebsd-stable; I am not sure which forum is appropriate. **) I just upgraded my Dell OptiPlex GX270 from RELENG_8 to RELENG_9. The machine no longer boots. However, if I put hint.acpi.0.disabled=1 in /boot/loader.conf then the machine runs fine. With RELENG_8 the machine had no loader.conf, and the power button worked on my desktop machine. Now with ACPI disabled my power button does not work. I have found that the machine hangs at boot during a scan of the PCI bus, but if I disable that (hw.acpi.disable=pci) then the machine cannot find a boot drive. So I have lost functionality that worked fine in BSD 8. Thoughts? Suggestions? Thanks, Dan Allen ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"