broken source upgrade 9.3-STABLE to 10.3-STABLE
Hi! I've got legacy FreeBSD server running 8.4-STABLE and I'm trying to upgrade it. Source upgrade for 8.4 to 9.3-STABLE r310015 went flawlessly. After reboot, I checked out stable/10 r310043 sources and ran buildworld again. It fails: ===> lib/clang/libllvmanalysis (all) c++ -O2 -pipe -I/usr/src/lib/clang/libllvmanalysis/../../../contrib/llvm/include -I/usr/src/lib/clang/libllvmanalysis/../../../contrib/llvm/tools/clang/include -I/usr/src/lib/clang/libllvmanalysis/../../../contrib/llvm/lib/Analysis -I. -I/usr/src/lib/clang/libllvmanalysis/../../../contrib/llvm/../../lib/clang/include -DLLVM_ON_UNIX -DLLVM_ON_FREEBSD -D__STDC_LIMIT_MACROS -D__STDC_CONSTANT_MACROS -DNDEBUG -fno-strict-aliasing -DLLVM_DEFAULT_TARGET_TRIPLE=\"x86_64-unknown-freebsd10.3\" -DLLVM_HOST_TRIPLE=\"x86_64-unknown-freebsd10.3\" -DDEFAULT_SYSROOT=\"/usr/obj/usr/src/tmp\" -I/usr/obj/usr/src/tmp/legacy/usr/include -fno-exceptions -fno-rtti -c /usr/src/lib/clang/libllvmanalysis/../../../contrib/llvm/lib/Analysis/LazyValueInfo.cpp -o LazyValueInfo.o /usr/src/lib/clang/libllvmanalysis/../../../contrib/llvm/lib/Analysis/LazyValueInfo.cpp: In member function 'llvm::Constant* llvm::LazyValueInfo::getConstant(llvm::Value*, llvm::BasicBlock*)': /usr/src/lib/clang/libllvmanalysis/../../../contrib/llvm/lib/Analysis/LazyValueInfo.cpp:1054: error: 'nullptr' was not declared in this scope *** Error code 1 Does it mean that FreeBSD clang version 3.4.1 (tags/RELEASE_34/dot1-final 208032) 20140512 is unable to build more recent clang version? ___ freebsd-stable@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
broken source upgrade 9.3-STABLE to 10.3-STABLE
Hi! I've got legacy FreeBSD server running 8.4-STABLE and I'm trying to upgrade it. Source upgrade for 8.4 to 9.3-STABLE r310015 went flawlessly. After reboot, I checked out stable/10 r310043 sources and ran buildworld again. It fails: ===> lib/clang/libllvmanalysis (all) c++ -O2 -pipe -I/usr/src/lib/clang/libllvmanalysis/../../../contrib/llvm/include -I/usr/src/lib/clang/libllvmanalysis/../../../contrib/llvm/tools/clang/include -I/usr/src/lib/clang/libllvmanalysis/../../../contrib/llvm/lib/Analysis -I. -I/usr/src/lib/clang/libllvmanalysis/../../../contrib/llvm/../../lib/clang/include -DLLVM_ON_UNIX -DLLVM_ON_FREEBSD -D__STDC_LIMIT_MACROS -D__STDC_CONSTANT_MACROS -DNDEBUG -fno-strict-aliasing -DLLVM_DEFAULT_TARGET_TRIPLE=\"x86_64-unknown-freebsd10.3\" -DLLVM_HOST_TRIPLE=\"x86_64-unknown-freebsd10.3\" -DDEFAULT_SYSROOT=\"/usr/obj/usr/src/tmp\" -I/usr/obj/usr/src/tmp/legacy/usr/include -fno-exceptions -fno-rtti -c /usr/src/lib/clang/libllvmanalysis/../../../contrib/llvm/lib/Analysis/LazyValueInfo.cpp -o LazyValueInfo.o /usr/src/lib/clang/libllvmanalysis/../../../contrib/llvm/lib/Analysis/LazyValueInfo.cpp: In member function 'llvm::Constant* llvm::LazyValueInfo::getConstant(llvm::Value*, llvm::BasicBlock*)': /usr/src/lib/clang/libllvmanalysis/../../../contrib/llvm/lib/Analysis/LazyValueInfo.cpp:1054: error: 'nullptr' was not declared in this scope *** Error code 1 Does it mean that FreeBSD clang version 3.4.1 (tags/RELEASE_34/dot1-final 208032) 20140512 is unable to build more recent clang version? ___ freebsd-stable@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: Huge hpet0 interrupt storm, but only if HGST HTS721010A9E630 HDD is present at the boot _and_ formatted with recognized fs (??)
No change after a # gpart destroy -F , like something permanently changed after first writes to disk. -- View this message in context: http://freebsd.1045724.x6.nabble.com/Huge-hpet0-interrupt-storm-but-only-if-HGST-HTS721010A9E630-HDD-is-present-at-the-boot-and-formatted-tp6151041p6151798.html Sent from the freebsd-stable mailing list archive at Nabble.com. ___ freebsd-stable@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
The EASIEST CHRISTMAS present for everyone..
Your email client cannot read this email. To view it online, please go here: http://e.bananabed.com.au/mail/display.php?M=167275=8648afca1a3d942b3722860809266e82=9=2=2 To stop receiving these emails:http://e.bananabed.com.au/mail/unsubscribe.php?M=167275=8648afca1a3d942b3722860809266e82=2=9 ___ freebsd-stable@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: CVE-2016-7434 NTP
We plan to issue an EN to update the base system ntp to 4.2.8p9. The high impact issue is Windows only by the way. Cheers, On Mon, Dec 12, 2016 at 6:18 PM, Michelle Sullivanwrote: > Dimitry Andric wrote: >> >> On 08 Dec 2016, at 06:08, Michelle Sullivan wrote: >>> >>> Are we going to get a patch for CVE-2016-7434 on FreeBSD 9.3? >> >> On Nov 22, in r309009, Xin Li merged ntp 4.2.8p9, which fixes this >> issue, to stable/9: >> >> https://svnweb.freebsd.org/changeset/base/309009 >> >> Unfortunately the commit message did not mention the CVE identifier. I >> can't find any corresponding security advisory either. >> >> -Dimitry >> > > > No updates needed to update system to 9.3-RELEASE-p52. > No updates are available to install. > Run '/usr/sbin/freebsd-update fetch' first. > [root@gauntlet /]# ntpd --version > ntpd 4.2.8p8-a (1) > > So no then... > > 9.3 is still so-say supported so I'm not talking about -STABLE. > > Michelle ___ freebsd-stable@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: make buildkernel does not respect KERNCONF or JOBS in /etc/make.conf
On Tue, 2016-12-13 at 10:13 -0700, Ian Lepore wrote: > On Tue, 2016-12-13 at 10:00 +, tech-lists wrote: > > > > On 12/12/2016 23:40, Herbert J. Skuhra wrote: > > > > > > > > > PORTS_MODULES does not work if KERNCONF contains multiple > > > kernels: > > > > > > The problem is obviously in /usr/src/sys/conf/kern.post.mk (line > > > 66): > > > > > > WRKDIRPREFIX?= ${MAKEOBJDIRPREFIX}${SRC_BASE}/sys/${KERNCONF} > > hmm! I didn't know that. > > > > I can't confirm exactly when the old way stopped working and when > > I > > started defining modules in src.conf. > > > > If I wanted to install a known, good kernel as /boot/workingkernel > > with > > all of its modules, so that I can avoid kernel.old being a bad > > kernel > > and kernel being non-bootable, how would I go about doing it? > > > > many thanks, > > > I think the problem might have started with some changes to the > kernel > build infrastructure that result in reading make.conf and/or src.conf > when they didn't used to, so now KERNCONF with multiple entries is > defined differently in kern.post.mk than it used to be. > > I wonder if this patch might fix it (I'm not in a position to test it > myself right now -- this is purely a shot in the dark)... > > iIndex: sys/conf/kern.post.mk > === > --- sys/conf/kern.post.mk (revision 302505) > +++ sys/conf/kern.post.mk (working copy) > @@ -63,7 +63,7 @@ OSRELDATE!= awk > '/^\#define[[:space:]]*__FreeBSD_v > ${MAKEOBJDIRPREFIX}${SRC_BASE}/include/osreldate > .h > .endif > # Keep the related ports builds in the obj directory so that they > are only rebuilt once per kernel build > -WRKDIRPREFIX?= ${MAKEOBJDIRPREFIX}${SRC_BASE}/sys/${KERNCONF} > +WRKDIRPREFIX?= ${MAKEOBJDIRPREFIX}${SRC_BASE}/sys/${.OBJDIR} > PORTSMODULESENV=\ > PATH=${PATH}:${LOCALBASE}/bin:${LOCALBASE}/sbin \ > SRC_BASE=${SRC_BASE} \ > > -- Ian Actually, now that I look at it again, I wonder if it should be just: +WRKDIRPREFIX?= ${.OBJDIR} -- Ian ___ freebsd-stable@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: make buildkernel does not respect KERNCONF or JOBS in /etc/make.conf
On Tue, 2016-12-13 at 10:00 +, tech-lists wrote: > On 12/12/2016 23:40, Herbert J. Skuhra wrote: > > > > PORTS_MODULES does not work if KERNCONF contains multiple kernels: > > > > The problem is obviously in /usr/src/sys/conf/kern.post.mk (line > > 66): > > > > WRKDIRPREFIX?= ${MAKEOBJDIRPREFIX}${SRC_BASE}/sys/${KERNCONF} > hmm! I didn't know that. > > I can't confirm exactly when the old way stopped working and when I > started defining modules in src.conf. > > If I wanted to install a known, good kernel as /boot/workingkernel > with > all of its modules, so that I can avoid kernel.old being a bad > kernel > and kernel being non-bootable, how would I go about doing it? > > many thanks, > I think the problem might have started with some changes to the kernel build infrastructure that result in reading make.conf and/or src.conf when they didn't used to, so now KERNCONF with multiple entries is defined differently in kern.post.mk than it used to be. I wonder if this patch might fix it (I'm not in a position to test it myself right now -- this is purely a shot in the dark)... iIndex: sys/conf/kern.post.mk === --- sys/conf/kern.post.mk (revision 302505) +++ sys/conf/kern.post.mk (working copy) @@ -63,7 +63,7 @@ OSRELDATE!= awk '/^\#define[[:space:]]*__FreeBSD_v ${MAKEOBJDIRPREFIX}${SRC_BASE}/include/osreldate.h .endif # Keep the related ports builds in the obj directory so that they are only rebuilt once per kernel build -WRKDIRPREFIX?= ${MAKEOBJDIRPREFIX}${SRC_BASE}/sys/${KERNCONF} +WRKDIRPREFIX?= ${MAKEOBJDIRPREFIX}${SRC_BASE}/sys/${.OBJDIR} PORTSMODULESENV=\ PATH=${PATH}:${LOCALBASE}/bin:${LOCALBASE}/sbin \ SRC_BASE=${SRC_BASE} \ -- Ian ___ freebsd-stable@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: make buildkernel does not respect KERNCONF or JOBS in /etc/make.conf
On Mon, 2016-12-12 at 20:22 -0800, Kevin Oberman wrote: > On Mon, Dec 12, 2016 at 4:20 PM, Herbert J. Skuhrarg> > wrote: > > > > > Kevin Oberman skrev: > > > > > > > > > Clearly the documentation is a bit behind the times. For some > > > time people > > > have used KERNCONF to build multiple kernels, but that was a > > > lucky things > > > that was not officially supported. It just happened to work. > > > Then, with > > > 11.0, it no longer did in many cases sue to changes in the kernel > > > build > > > system. When people complained, there did not seem to be a way to > > > fix > > this > > > > > > without blocking future goals in speeding up and enhancinghte > > > standard > > > kernel build. > > > > > > The result was BUILDKERNELS. It was specifically designed to > > > allow the > > > building of multiple kernels and the appropriate modules. This > > > would > > always > > > > > > take longer and use more disk space, but it would work reliably. > > > Now, bit > > > building a single, local kernel, KERNCONF is the best way, though > > > I > > suspect > > > > > > that it make only a small difference until new capabilities are > > > added > > later > > > > > > in the life of 11. > > > > > > So, while it seems the man pages need to catch up, building > > > multiple > > > kernels should be done with KERNCONF in either make.conf or > > > src.conf and > > > multiple kernels with BUILDKERNELS. > > > > > > This is from my recollection of the discussion thread. I'll admit > > > to > > being > > > > > > too lazy to go find and read all of it. I suspect it was on > > > current@, > > but > > > > > > I'm not even sure of that. > > ??? > > > > From /usr/src/Makefile.inc1: > > > > 1137 .if ${TARGET_ARCH} == "powerpc64" > > 1138 KERNCONF?= GENERIC64 > > 1139 .else > > 1140 KERNCONF?= GENERIC > > 1141 .endif > > [...] > > 1149 BUILDKERNELS= > > 1150 INSTALLKERNEL= > > 1151 .if defined(NO_INSTALLKERNEL) > > 1152 # All of the BUILDKERNELS loops start at index 1. > > 1153 BUILDKERNELS+= dummy > > 1154 .endif > > 1155 .for _kernel in ${KERNCONF} > > 1156 .if exists(${KERNCONFDIR}/${_kernel}) > > 1157 BUILDKERNELS+= ${_kernel} > > 1158 .if empty(INSTALLKERNEL) && !defined(NO_INSTALLKERNEL) > > 1159 INSTALLKERNEL= ${_kernel} > > 1160 .endif > > 1161 .endif > > 1162 .endfor > > > > So setting BUILDKERNELS has no effect. > > > > -- > > Herbert > > > I think you miss the significance. BUILDKERNELS only is used to build > the > kernels. It is not used by installkernel as installing two kernels > does not > make sense. It is ONLY used to build multiple kernels. KERNCONF is > still > used to specify the kernel, with attendant modules, to be installed. > > I should also mention that, if you want to install a new kernel > without > overwriting the old kernel and modules (/boot/kernel.old/), "make > reinstallkernel". It replaces the existing kernel instead of renaming > the > kernel directory to kernel.old. I find this very handy, but it is > poorly > documented. So, if you want to install GENERIC and not lose your last > working kernel, "make reinstallkernel KERNCONF=GENERIC". That will > blow > away the bad kernel and modules and install GENERIC. Note that it > does not > touch the /usr/obj directory that PUMPKIN is built in, so PUMPKIN and > similarly be reinstalled. > > The man page for src.conf is automatically generated and only lists > those > values which are specific to src.conf (WITH_ and WITHOUT_) and > describes > its use. It is used as input for system builds and installs but > should not > be accessed for any other purpose. make.conf is always read by make > with no > regard to what is being made. (N.B. I believe that some people have > ignored > this in some ports stuff.) Anything that is put into make.conf may be > placed in src,conf if you only want to have it used when > building/installing the kernel and world. > > I'm probably forgetting something, but I hope this explains it a bit. The BUILDKERNELS variable is part of the build system's internal machinery for building multiple kernels when the user has set KERNCONF= to more than one name. It's not a thing that a user can set. -- Ian ___ freebsd-stable@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: Is System V IPC namespace still shared across jails?
2016-12-13 16:29, Alan Somers wrote: I've already added support for sysvmsg, sysvsem, and sysvshm to iocage. They all default to "new", which means you won't have to do anything special in your jail config to make postgres work. You can find the patch below. The only reason it hasn't been merged is because it can't (yet) be made to work correctly on the develop branch of iocage. But it works fine on the master branch. https://github.com/iocage/iocage/pull/370 -Alan Superb, appreciated! Mark On Tue, Dec 13, 2016 at 8:08 AM, Mark Martinecwrote: 2016-12-12 20:38, Christian Schwarz wrote: With the new jail parameters, new namespaces for SysV IPC are possible on FreeBSD 11. For those ezjail users, add something like this to the jail's config after creating it using 'ezjail-admin create': export jail_postgres_parameters="sysvmsg=new sysvsem=new sysvshm=new" Cheers, Christian Thank you, this is it! I missed it in the JAIL(8) man page, and is not mentioned in release notes. Now if only the iocage would recognized the sysvmsg, sysvsem, and sysvshm options: # iocage set sysvmsg='new' xxx ERROR: Unsupported property: sysvmsg! I guess I should file a bug report. Mark ___ freebsd-stable@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: CVE-2016-7434 NTP
On 13 Dec 2016, at 03:18, Michelle Sullivanwrote: > > Dimitry Andric wrote: >> On 08 Dec 2016, at 06:08, Michelle Sullivan wrote: >>> Are we going to get a patch for CVE-2016-7434 on FreeBSD 9.3? >> On Nov 22, in r309009, Xin Li merged ntp 4.2.8p9, which fixes this >> issue, to stable/9: >> >> https://svnweb.freebsd.org/changeset/base/309009 >> >> Unfortunately the commit message did not mention the CVE identifier. I >> can't find any corresponding security advisory either. ... > No updates needed to update system to 9.3-RELEASE-p52. > No updates are available to install. > Run '/usr/sbin/freebsd-update fetch' first. > [root@gauntlet /]# ntpd --version > ntpd 4.2.8p8-a (1) > > So no then... > > 9.3 is still so-say supported so I'm not talking about -STABLE. Well, as I mentioned, there was no Security Advisory (which is a little strange), so I didn't expect there to be any binary updates. As far as I know, binary updates are only built for Security Advisories and Errata Notices. -Dimitry signature.asc Description: Message signed with OpenPGP using GPGMail
Re: Is System V IPC namespace still shared across jails?
I've already added support for sysvmsg, sysvsem, and sysvshm to iocage. They all default to "new", which means you won't have to do anything special in your jail config to make postgres work. You can find the patch below. The only reason it hasn't been merged is because it can't (yet) be made to work correctly on the develop branch of iocage. But it works fine on the master branch. https://github.com/iocage/iocage/pull/370 -Alan On Tue, Dec 13, 2016 at 8:08 AM, Mark Martinecwrote: > 2016-12-12 20:38, Christian Schwarz wrote: >> >> With the new jail parameters, new namespaces for SysV IPC are possible >> on FreeBSD 11. >> >> For those ezjail users, add something like this to the jail's config >> after creating it using 'ezjail-admin create': >> >> export jail_postgres_parameters="sysvmsg=new sysvsem=new sysvshm=new" >> >> Cheers, >> Christian > > > > Thank you, this is it! > I missed it in the JAIL(8) man page, and is not mentioned in release notes. > > > Now if only the iocage would recognized the sysvmsg, sysvsem, and sysvshm > options: > > # iocage set sysvmsg='new' xxx > ERROR: Unsupported property: sysvmsg! > > I guess I should file a bug report. > > > Mark > > > >> man 8 jail >>> >>> ... >>> allow.sysvipc >>> A process within the jail has access to System V IPC >>> primitives. This is deprecated in favor of the per- >>> module parameters (see below). When this parameter is >>> set, it is equivalent to setting sysvmsg, sysvsem, and >>> sysvshm all to ``inherit''. >>> ... >>> >>>sysvmsg >>> Allow access to SYSV IPC message primitives. If set to >>> ``inherit'', all IPC objects on the system are visible to this >>> jail, whether they were created by the jail itself, the base >>> system, or other jails. If set to ``new'', the jail will have >>> its own key namespace, and can only see the objects that it has >>> created; the system (or parent jail) has access to the jail's >>> objects, but not to its keys. If set to ``disable'', the jail >>> cannot perform any sysvmsg-related system calls. >>> >>> sysvsem, sysvshm >>> Allow access to SYSV IPC semaphore and shared memory primitives, >>> in the same manner as sysvmsg. > > Regarding installation of PostgreSQL in a FreeBSD jail, the web hold plenty of warnings/advice that each postgres instance should have a unique UID, otherwise they stumble across each other's feet: | allow.sysvipc | A process within the jail has access to System V IPC primitives. In the | current jail implementation, System V primitives share a single namespace | across the host and jail environments, meaning that processes within a jail | would be able to communicate with (and potentially interfere with) processes | outside of the jail, and in other jails. Is this still the case in FreeBSD 11.0 ??? I remember hearing rumors that the System V namespace no longer is (will?) be shared across jails. (Couldn't find it being mentioned in release notes.) Mark > > ___ > freebsd-stable@freebsd.org mailing list > https://lists.freebsd.org/mailman/listinfo/freebsd-stable > To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org" ___ freebsd-stable@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: Is System V IPC namespace still shared across jails?
2016-12-12 20:38, Christian Schwarz wrote: With the new jail parameters, new namespaces for SysV IPC are possible on FreeBSD 11. For those ezjail users, add something like this to the jail's config after creating it using 'ezjail-admin create': export jail_postgres_parameters="sysvmsg=new sysvsem=new sysvshm=new" Cheers, Christian Thank you, this is it! I missed it in the JAIL(8) man page, and is not mentioned in release notes. Now if only the iocage would recognized the sysvmsg, sysvsem, and sysvshm options: # iocage set sysvmsg='new' xxx ERROR: Unsupported property: sysvmsg! I guess I should file a bug report. Mark man 8 jail ... allow.sysvipc A process within the jail has access to System V IPC primitives. This is deprecated in favor of the per- module parameters (see below). When this parameter is set, it is equivalent to setting sysvmsg, sysvsem, and sysvshm all to ``inherit''. ... sysvmsg Allow access to SYSV IPC message primitives. If set to ``inherit'', all IPC objects on the system are visible to this jail, whether they were created by the jail itself, the base system, or other jails. If set to ``new'', the jail will have its own key namespace, and can only see the objects that it has created; the system (or parent jail) has access to the jail's objects, but not to its keys. If set to ``disable'', the jail cannot perform any sysvmsg-related system calls. sysvsem, sysvshm Allow access to SYSV IPC semaphore and shared memory primitives, in the same manner as sysvmsg. Regarding installation of PostgreSQL in a FreeBSD jail, the web hold plenty of warnings/advice that each postgres instance should have a unique UID, otherwise they stumble across each other's feet: | allow.sysvipc | A process within the jail has access to System V IPC primitives. In the | current jail implementation, System V primitives share a single namespace | across the host and jail environments, meaning that processes within a jail | would be able to communicate with (and potentially interfere with) processes | outside of the jail, and in other jails. Is this still the case in FreeBSD 11.0 ??? I remember hearing rumors that the System V namespace no longer is (will?) be shared across jails. (Couldn't find it being mentioned in release notes.) Mark ___ freebsd-stable@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: make buildkernel does not respect KERNCONF or JOBS in /etc/make.conf
On 12/12/2016 23:40, Herbert J. Skuhra wrote: PORTS_MODULES does not work if KERNCONF contains multiple kernels: The problem is obviously in /usr/src/sys/conf/kern.post.mk (line 66): WRKDIRPREFIX?= ${MAKEOBJDIRPREFIX}${SRC_BASE}/sys/${KERNCONF} hmm! I didn't know that. I can't confirm exactly when the old way stopped working and when I started defining modules in src.conf. If I wanted to install a known, good kernel as /boot/workingkernel with all of its modules, so that I can avoid kernel.old being a bad kernel and kernel being non-bootable, how would I go about doing it? many thanks, -- J. ___ freebsd-stable@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: make buildkernel does not respect KERNCONF or JOBS in /etc/make.conf
On 13/12/2016 04:22, Kevin Oberman wrote: I should also mention that, if you want to install a new kernel without overwriting the old kernel and modules (/boot/kernel.old/), "make reinstallkernel". It replaces the existing kernel instead of renaming the kernel directory to kernel.old. I find this very handy, but it is poorly documented. So, if you want to install GENERIC and not lose your last working kernel, "make reinstallkernel KERNCONF=GENERIC". That will blow away the bad kernel and modules and install GENERIC. Note that it does not touch the /usr/obj directory that PUMPKIN is built in, so PUMPKIN and similarly be reinstalled. Hi, Thanks for trying to help - as you say, it's quite poorly documented. I'd document it if I had more of as clue what I was doing. Do you know how to get the process to respect MAKE_JOBS_NUMBER=32 ? (I mean I'm happy to have to put everything on one line rather than setting variables in src.conf or make.conf, it's just that it used to work and now does not, so am interested in *The Right Way*(tm) to do what I'm trying to do). Many thanks, -- J. ___ freebsd-stable@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
Re: make buildkernel does not respect KERNCONF or JOBS in /etc/make.conf
Kevin Oberman skrev: > > On Mon, Dec 12, 2016 at 4:20 PM, Herbert J. Skuhra> wrote: > >> Kevin Oberman skrev: >> > >> > Clearly the documentation is a bit behind the times. For some time people >> > have used KERNCONF to build multiple kernels, but that was a lucky things >> > that was not officially supported. It just happened to work. Then, with >> > 11.0, it no longer did in many cases sue to changes in the kernel build >> > system. When people complained, there did not seem to be a way to fix >> this >> > without blocking future goals in speeding up and enhancinghte standard >> > kernel build. >> > >> > The result was BUILDKERNELS. It was specifically designed to allow the >> > building of multiple kernels and the appropriate modules. This would >> always >> > take longer and use more disk space, but it would work reliably. Now, bit >> > building a single, local kernel, KERNCONF is the best way, though I >> suspect >> > that it make only a small difference until new capabilities are added >> later >> > in the life of 11. >> > >> > So, while it seems the man pages need to catch up, building multiple >> > kernels should be done with KERNCONF in either make.conf or src.conf and >> > multiple kernels with BUILDKERNELS. >> > >> > This is from my recollection of the discussion thread. I'll admit to >> being >> > too lazy to go find and read all of it. I suspect it was on current@, >> but >> > I'm not even sure of that. >> >> ??? >> >> From /usr/src/Makefile.inc1: >> >> 1137 .if ${TARGET_ARCH} == "powerpc64" >> 1138 KERNCONF?= GENERIC64 >> 1139 .else >> 1140 KERNCONF?= GENERIC >> 1141 .endif >> [...] >> 1149 BUILDKERNELS= >> 1150 INSTALLKERNEL= >> 1151 .if defined(NO_INSTALLKERNEL) >> 1152 # All of the BUILDKERNELS loops start at index 1. >> 1153 BUILDKERNELS+= dummy >> 1154 .endif >> 1155 .for _kernel in ${KERNCONF} >> 1156 .if exists(${KERNCONFDIR}/${_kernel}) >> 1157 BUILDKERNELS+= ${_kernel} >> 1158 .if empty(INSTALLKERNEL) && !defined(NO_INSTALLKERNEL) >> 1159 INSTALLKERNEL= ${_kernel} >> 1160 .endif >> 1161 .endif >> 1162 .endfor >> >> So setting BUILDKERNELS has no effect. >> >> -- >> Herbert >> > > I think you miss the significance. BUILDKERNELS only is used to build the > kernels. It is not used by installkernel as installing two kernels does not > make sense. It is ONLY used to build multiple kernels. KERNCONF is still > used to specify the kernel, with attendant modules, to be installed. > > I should also mention that, if you want to install a new kernel without > overwriting the old kernel and modules (/boot/kernel.old/), "make > reinstallkernel". It replaces the existing kernel instead of renaming the > kernel directory to kernel.old. I find this very handy, but it is poorly > documented. So, if you want to install GENERIC and not lose your last > working kernel, "make reinstallkernel KERNCONF=GENERIC". That will blow > away the bad kernel and modules and install GENERIC. Note that it does not > touch the /usr/obj directory that PUMPKIN is built in, so PUMPKIN and > similarly be reinstalled. > > The man page for src.conf is automatically generated and only lists those > values which are specific to src.conf (WITH_ and WITHOUT_) and describes > its use. It is used as input for system builds and installs but should not > be accessed for any other purpose. make.conf is always read by make with no > regard to what is being made. (N.B. I believe that some people have ignored > this in some ports stuff.) Anything that is put into make.conf may be > placed in src,conf if you only want to have it used when > building/installing the kernel and world. > > I'm probably forgetting something, but I hope this explains it a bit. Sorry, but why are you telling (me) all this? :) Can we return to the OP's issue? How does setting BUILDKERNELS (as suggested by Scott B.) resolves (t)his issue? His problem is obviously PORTS_MODULES (see my previous message). -- Herbert ___ freebsd-stable@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"