Date: Sun, 27 Mar 2022 15:06:50 +0000 From: Rick Macklem <rmacklem@uoguelph.ca> To: Yoshihiro Ota <ota@j.email.ne.jp> Cc: mike tancsa <mike@sentex.net>, freebsd-stable <freebsd-stable@freebsd.org> Subject: Re: nfsd becomes slow when machine CPU usage is at or over 100% on STABLE/13 Message-ID: <YT2PR01MB9730B3A7BB007F2B0B865019DD1C9@YT2PR01MB9730.CANPRD01.PROD.OUTLOOK.COM> In-Reply-To: <20220323001002.35f7a7d660654757178f0518@j.email.ne.jp> References: <20220309034601.ea3135e31aec3ffb2623f145@j.email.ne.jp> <YT2PR01MB9730D7B51D325258AAA29828DD0A9@YT2PR01MB9730.CANPRD01.PROD.OUTLOOK.COM> <20220318211850.67b77d43b3a02043c3819bf3@j.email.ne.jp> <9b2935cd-5f70-d2a2-a328-004fa3e7b3f8@sentex.net> <9aee6d0f-e1c9-8c98-7352-7a0ff3e3d147@sentex.net> <YT2PR01MB9730287F7C97ED81BA270DB5DD159@YT2PR01MB9730.CANPRD01.PROD.OUTLOOK.COM> <20220323001002.35f7a7d660654757178f0518@j.email.ne.jp>
next in thread | previous in thread | raw e-mail | index | archive | help
Yoshihiro Ota <ota@j.email.ne.jp> wrote:=0A=
> I've been building default kernel, that is GENERIC for releng and stable=
=0A=
> branches.=0A=
>=0A=
> I see GENERIC-NODEBUG on head but don't seem to find it on releng/=0A=
> nor stable/=0A=
> I grepped /usr/src/* for std.nodebug but don't seem have a match...=0A=
> I wonder how nodebug kicks in releng.=0A=
I do not know what magic is used so that the releng kernels build=0A=
without debugging.=0A=
=0A=
However, if you see this line:=0A=
WARNING: WITNESS option enabled, expect reduced performance.=0A=
in the first screen when booting, you are running a kernel with debugging=
=0A=
built into it.=0A=
=0A=
rick=0A=
=0A=
I added "enable DDB" to i386 kernel but didn't touch amd64 kernel.=0A=
Both amd64 and i386 looks okay with releng/13.1.=0A=
=0A=
Hiro=0A=
=0A=
On Sun, 20 Mar 2022 20:45:30 +0000=0A=
Rick Macklem <rmacklem@uoguelph.ca> wrote:=0A=
=0A=
> mike tancsa <mike@sentex.net> wrote:=0A=
> > On 3/20/2022 7:43 AM, mike tancsa wrote:=0A=
> >> On 3/18/2022 9:18 PM, Yoshihiro Ota wrote:=0A=
> >>> I had built several versions between releng/13.0 branch point to=0A=
> >>> stable/13 (before releng/13.1 was created) and all of them had such=
=0A=
> >> performance degrade.=0A=
> >>>=0A=
> >>> I started suspecting stable debug options and thus built releng/13.1=
=0A=
> >>> and tested.=0A=
> >>> I don't see NFS slowdown unlike stable/13.=0A=
> >>> releng/13.0 and releng/12.2 were also fine.=0A=
> >>=0A=
> >> Hi,=0A=
> >>=0A=
> >> I would think there is very little difference (if any) between=0A=
> >> releng/13.1 and stable/13 right now. Are you sure stable/13 suffers=
=0A=
> >> from this issue you are seeing ?=0A=
> The sources may be almost the same, but the build is not.=0A=
> See /usr/src/sys/conf/std.nodebug.=0A=
>=0A=
> I'm assuming his releng/13.1 build created a non-debug kernel.=0A=
> Debug kernels do spit out "expect reduced performance" if i recall=0A=
> correctly. It sounds like he found an example of this.=0A=
>=0A=
> rick=0A=
>=0A=
> >=0A=
> >=0A=
> These look to be the only files touched below.=0A=
>=0A=
> 0{cage}% git diff remotes/origin/releng/13.1..remotes/origin/stable/13 |=
=0A=
> grep '^\-\-'=0A=
> --- a/contrib/tzcode/stdtime/ctime.3=0A=
> --- a/lib/libc/gen/time.3=0A=
> --- a/lib/libcasper/services/cap_net/cap_net.c=0A=
> --- a/lib/libpfctl/libpfctl.c=0A=
> --- a/lib/libpfctl/libpfctl.h=0A=
> --- a/libexec/rc/rc.d/dumpon=0A=
> --- a/release/pkg_repos/release-dvd.conf=0A=
> --- a/sbin/devd/devd.conf=0A=
> --- a/sbin/ipf/common/ipf.h=0A=
> --- a/sbin/ipf/libipf/printactivenat.c=0A=
> --- a/sbin/ipf/libipf/printstate.c=0A=
> --- a/sbin/pfctl/pfctl.c=0A=
> --- a/sbin/pfctl/pfctl_optimize.c=0A=
> --- a/share/man/man4/Makefile=0A=
> --- a/share/man/man4/netmap.4=0A=
> --- /dev/null=0A=
> --- a/share/man/man4/vale.4=0A=
> --- a/share/man/man9/crypto_buffer.9=0A=
> --- a/stand/efi/libefi/efi_console.c=0A=
> --- a/stand/i386/libi386/vidconsole.c=0A=
> --- a/sys/arm64/include/pcpu.h=0A=
> --- a/sys/cddl/contrib/opensolaris/uts/common/dtrace/fasttrap.c=0A=
> --- a/sys/cddl/contrib/opensolaris/uts/intel/dtrace/fasttrap_isa.c=0A=
> --- a/sys/conf/newvers.sh=0A=
> --- a/sys/crypto/armv8/armv8_crypto.c=0A=
> --- a/sys/crypto/armv8/armv8_crypto.h=0A=
> --- a/sys/crypto/armv8/armv8_crypto_wrap.c=0A=
> --- a/sys/dev/netmap/netmap.c=0A=
> --- a/sys/dev/netmap/netmap_bdg.c=0A=
> --- a/sys/dev/netmap/netmap_kern.h=0A=
> --- a/sys/dev/netmap/netmap_vale.c=0A=
> --- a/sys/i386/i386/machdep.c=0A=
> --- a/sys/kern/kern_rmlock.c=0A=
> --- a/sys/kern/sys_process.c=0A=
> --- a/sys/kern/vfs_cache.c=0A=
> --- a/sys/kern/vfs_subr.c=0A=
> --- a/sys/modules/if_epair/Makefile=0A=
> --- a/sys/modules/linuxkpi/Makefile=0A=
> --- a/sys/net/if_epair.c=0A=
> --- a/sys/opencrypto/cryptodev.h=0A=
> --- a/sys/riscv/include/cpufunc.h=0A=
> --- a/sys/riscv/include/pmap.h=0A=
> --- a/sys/riscv/include/pte.h=0A=
> --- a/sys/riscv/include/riscvreg.h=0A=
> --- a/sys/riscv/include/vmparam.h=0A=
> --- a/sys/riscv/riscv/elf_machdep.c=0A=
> --- a/sys/riscv/riscv/locore.S=0A=
> --- a/sys/riscv/riscv/pmap.c=0A=
> --- a/sys/sys/param.h=0A=
> --- a/sys/x86/x86/mp_x86.c=0A=
> --- a/usr.bin/diff/pr.c=0A=
> --- a/usr.bin/touch/touch.c=0A=
> 0{cage}%=0A=
>=0A=
>=0A=
>=0A=
=0A=
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?YT2PR01MB9730B3A7BB007F2B0B865019DD1C9>