Re: SVN commit 259045 breaks -CURRENT
Am 15.12.2013 17:27, schrieb Steven Hartland: > > - Original Message - From: "Stefan Esser" > To: "Konstantin Belousov" ; "Steve Kargl" > > Cc: > Sent: Sunday, December 15, 2013 1:25 PM > Subject: Re: SVN commit 259045 breaks -CURRENT > > >> Am 15.12.2013 06:47, schrieb Konstantin Belousov: >>> On Sat, Dec 14, 2013 at 02:16:27PM -0800, Steve Kargl wrote: >>>> On Sat, Dec 14, 2013 at 11:11:15PM +0100, Stefan Esser wrote: >>>>> Am 14.12.2013 22:59, schrieb Steve Kargl: >>>>>> On Sat, Dec 14, 2013 at 10:44:10PM +0100, Stefan Esser >>>>>> wrote: >>>>>>> >>>>>>> 2) SSH logins are very slow, many seconds of delay between >>>>>>> connect and password prompt, several seconds after password >>>>>>> entry until a command prompt appears (normally >>>>>>> instantaneous) >>>>>>> >>>>>> >>>>>> Ah, so that explains the behavior I'm see. Just updated a >>>>>> circa Aug 3rd i386 FreeBSD to top-of-tree. My ssh logins to >>>>>> my work system take 30+ seconds now. :( >>>>> >>>>> You may want to test the attached patch, which reverts the >>>>> above mentioned commit. >>>>> >>>> >>>> I probably won't get to it until tomorrow, because I had started a >>>> dog-food system purge including re-installing all ports. The laptop >>>> takes a bit a time to recompile everything. >>>> >>> >>> Are you all running i386, compiled with gcc ? >> >> I'm on -CURRENT, CLANG, amd64. But since the problem has also been >> reported for i386 compiled with GCC, there seems to be some problem >> in common kernel code, that has been uncovered by your change, >> >> BTW: I remember seeing two wait channels being reported when I type >> ^T during multi-user startup (sa-spamd, which needs 140 seconds with >> the broken kernel: >> >> nanosleep >> kqueue (I do not remember whether this name is exact or abbreviated) >> >> I've been assuming that the problem might actually be in nanosleep(), >> since this is a timing related function and we are seeing huge >> delays, but eventually the delayed action succeeds. >> >> But a kernel with only kern_conf.c compiled with -fno-strict-overflow >> did not show the delays. I do not have time for further tests, today. > > Delay in ssh login for ~30 is typical if its failing to resolve > the connecting IP. The update didn't break your resolver in anyway > did it? No, I see a delay of some 3 to 5 seconds between start of the client and password prompt and another delay of 3 to 10 seconds after password entry and login prompt. There is no observable delay with a non-broken kernel. It was not me who reported 30s+ delays for SSH logins. And the delays exist even if connecting to the local SMTP port (e.g. a telnet localhost 25 takes a few seconds to establish the connection). (I had to disable VerifyHostKeyDNS in my SSH client, to prevent a 30s delay on outgoing connections. But this is completely unrelated.) Regards, STefan ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: SVN commit 259045 breaks -CURRENT
On 15 Dec, Konstantin Belousov wrote: > On Sun, Dec 15, 2013 at 05:28:14PM +0100, Marc UBM wrote: >> On Sun, 15 Dec 2013 08:43:22 +0200 >> Konstantin Belousov wrote: >> >> > On Sat, Dec 14, 2013 at 09:56:04PM -0800, Steve Kargl wrote: >> > > On Sun, Dec 15, 2013 at 07:47:22AM +0200, Konstantin Belousov wrote: >> > > > On Sat, Dec 14, 2013 at 02:16:27PM -0800, Steve Kargl wrote: >> > > > > On Sat, Dec 14, 2013 at 11:11:15PM +0100, Stefan Esser wrote: >> > > > > > Am 14.12.2013 22:59, schrieb Steve Kargl: >> > > > > > > On Sat, Dec 14, 2013 at 10:44:10PM +0100, Stefan Esser wrote: >> > > > > > >> >> > > > > > >> 2) SSH logins are very slow, many seconds of delay between >> > > > > > >> connect >> > > > > > >>and password prompt, several seconds after password entry >> > > > > > >> until >> > > > > > >>a command prompt appears (normally instantaneous) >> > > > > > >> >> > > > > > > >> > > > > > > Ah, so that explains the behavior I'm see. Just updated a circa >> > > > > > > Aug 3rd >> > > > > > > i386 FreeBSD to top-of-tree. My ssh logins to my work system >> > > > > > > take 30+ >> > > > > > > seconds now. :( >> > > > > > >> > > > > > You may want to test the attached patch, which reverts the above >> > > > > > mentioned commit. >> > > > > > >> > > > > >> > > > > I probably won't get to it until tomorrow, because I had started >> > > > > a dog-food system purge including re-installing all ports. The >> > > > > laptop takes a bit a time to recompile everything. >> > > > > >> > > > >> > > > Are you all running i386, compiled with gcc ? >> > > >> > > The Aug 3rd system was built with gcc. The system upgrade I >> > > did this morning is using clang. I rebuilt everything and >> > > delete old things with delete-old and delete-old-libs. >> >> A kernel built with the commit reverted has not exhibited any similar >> behavior during the whole day. I'll recompile again with the overflow >> option enabled to see if the issue returns. > > I am interested in a different experiment. Compile the same version > of the kernel with gcc with -fno-strict-overflow enabled. Report if > the same problem persist. > > I am going to revert the option in several minutes, but do want see > the result of the above test. I have not seen the problem here: FreeBSD scratch.catspoiler.org 11.0-CURRENT FreeBSD 11.0-CURRENT #65 r259394M: Sat Dec 14 16:16:30 PST 2013 d...@scratch.catspoiler.org:/usr/obj/usr/src/sys/GENERICSMB i386 compiled with clang. ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: SVN commit 259045 breaks -CURRENT
On Sun, Dec 15, 2013 at 05:28:14PM +0100, Marc UBM wrote: > On Sun, 15 Dec 2013 08:43:22 +0200 > Konstantin Belousov wrote: > > > On Sat, Dec 14, 2013 at 09:56:04PM -0800, Steve Kargl wrote: > > > On Sun, Dec 15, 2013 at 07:47:22AM +0200, Konstantin Belousov wrote: > > > > On Sat, Dec 14, 2013 at 02:16:27PM -0800, Steve Kargl wrote: > > > > > On Sat, Dec 14, 2013 at 11:11:15PM +0100, Stefan Esser wrote: > > > > > > Am 14.12.2013 22:59, schrieb Steve Kargl: > > > > > > > On Sat, Dec 14, 2013 at 10:44:10PM +0100, Stefan Esser wrote: > > > > > > >> > > > > > > >> 2) SSH logins are very slow, many seconds of delay between > > > > > > >> connect > > > > > > >>and password prompt, several seconds after password entry > > > > > > >> until > > > > > > >>a command prompt appears (normally instantaneous) > > > > > > >> > > > > > > > > > > > > > > Ah, so that explains the behavior I'm see. Just updated a circa > > > > > > > Aug 3rd > > > > > > > i386 FreeBSD to top-of-tree. My ssh logins to my work system > > > > > > > take 30+ > > > > > > > seconds now. :( > > > > > > > > > > > > You may want to test the attached patch, which reverts the above > > > > > > mentioned commit. > > > > > > > > > > > > > > > > I probably won't get to it until tomorrow, because I had started > > > > > a dog-food system purge including re-installing all ports. The > > > > > laptop takes a bit a time to recompile everything. > > > > > > > > > > > > > Are you all running i386, compiled with gcc ? > > > > > > The Aug 3rd system was built with gcc. The system upgrade I > > > did this morning is using clang. I rebuilt everything and > > > delete old things with delete-old and delete-old-libs. > > A kernel built with the commit reverted has not exhibited any similar > behavior during the whole day. I'll recompile again with the overflow > option enabled to see if the issue returns. I am interested in a different experiment. Compile the same version of the kernel with gcc with -fno-strict-overflow enabled. Report if the same problem persist. I am going to revert the option in several minutes, but do want see the result of the above test. pgpB77zzeAzwQ.pgp Description: PGP signature
Re: SVN commit 259045 breaks -CURRENT
On Sun, 15 Dec 2013 08:43:22 +0200 Konstantin Belousov wrote: > On Sat, Dec 14, 2013 at 09:56:04PM -0800, Steve Kargl wrote: > > On Sun, Dec 15, 2013 at 07:47:22AM +0200, Konstantin Belousov wrote: > > > On Sat, Dec 14, 2013 at 02:16:27PM -0800, Steve Kargl wrote: > > > > On Sat, Dec 14, 2013 at 11:11:15PM +0100, Stefan Esser wrote: > > > > > Am 14.12.2013 22:59, schrieb Steve Kargl: > > > > > > On Sat, Dec 14, 2013 at 10:44:10PM +0100, Stefan Esser wrote: > > > > > >> > > > > > >> 2) SSH logins are very slow, many seconds of delay between connect > > > > > >>and password prompt, several seconds after password entry until > > > > > >>a command prompt appears (normally instantaneous) > > > > > >> > > > > > > > > > > > > Ah, so that explains the behavior I'm see. Just updated a circa > > > > > > Aug 3rd > > > > > > i386 FreeBSD to top-of-tree. My ssh logins to my work system take > > > > > > 30+ > > > > > > seconds now. :( > > > > > > > > > > You may want to test the attached patch, which reverts the above > > > > > mentioned commit. > > > > > > > > > > > > > I probably won't get to it until tomorrow, because I had started > > > > a dog-food system purge including re-installing all ports. The > > > > laptop takes a bit a time to recompile everything. > > > > > > > > > > Are you all running i386, compiled with gcc ? > > > > The Aug 3rd system was built with gcc. The system upgrade I > > did this morning is using clang. I rebuilt everything and > > delete old things with delete-old and delete-old-libs. A kernel built with the commit reverted has not exhibited any similar behavior during the whole day. I'll recompile again with the overflow option enabled to see if the issue returns. Bye Marc ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: SVN commit 259045 breaks -CURRENT
- Original Message - From: "Stefan Esser" To: "Konstantin Belousov" ; "Steve Kargl" Cc: Sent: Sunday, December 15, 2013 1:25 PM Subject: Re: SVN commit 259045 breaks -CURRENT Am 15.12.2013 06:47, schrieb Konstantin Belousov: On Sat, Dec 14, 2013 at 02:16:27PM -0800, Steve Kargl wrote: On Sat, Dec 14, 2013 at 11:11:15PM +0100, Stefan Esser wrote: Am 14.12.2013 22:59, schrieb Steve Kargl: On Sat, Dec 14, 2013 at 10:44:10PM +0100, Stefan Esser wrote: 2) SSH logins are very slow, many seconds of delay between connect and password prompt, several seconds after password entry until a command prompt appears (normally instantaneous) Ah, so that explains the behavior I'm see. Just updated a circa Aug 3rd i386 FreeBSD to top-of-tree. My ssh logins to my work system take 30+ seconds now. :( You may want to test the attached patch, which reverts the above mentioned commit. I probably won't get to it until tomorrow, because I had started a dog-food system purge including re-installing all ports. The laptop takes a bit a time to recompile everything. Are you all running i386, compiled with gcc ? I'm on -CURRENT, CLANG, amd64. But since the problem has also been reported for i386 compiled with GCC, there seems to be some problem in common kernel code, that has been uncovered by your change, BTW: I remember seeing two wait channels being reported when I type ^T during multi-user startup (sa-spamd, which needs 140 seconds with the broken kernel: nanosleep kqueue (I do not remember whether this name is exact or abbreviated) I've been assuming that the problem might actually be in nanosleep(), since this is a timing related function and we are seeing huge delays, but eventually the delayed action succeeds. But a kernel with only kern_conf.c compiled with -fno-strict-overflow did not show the delays. I do not have time for further tests, today. Delay in ssh login for ~30 is typical if its failing to resolve the connecting IP. The update didn't break your resolver in anyway did it? Regards Steve This e.mail is private and confidential between Multiplay (UK) Ltd. and the person or entity to whom it is addressed. In the event of misdirection, the recipient is prohibited from using, copying, printing or otherwise disseminating it or any information contained in it. In the event of misdirection, illegible or incomplete transmission please telephone +44 845 868 1337 or return the E.mail to postmas...@multiplay.co.uk. ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: SVN commit 259045 breaks -CURRENT
Am 15.12.2013 06:47, schrieb Konstantin Belousov: > On Sat, Dec 14, 2013 at 02:16:27PM -0800, Steve Kargl wrote: >> On Sat, Dec 14, 2013 at 11:11:15PM +0100, Stefan Esser wrote: >>> Am 14.12.2013 22:59, schrieb Steve Kargl: On Sat, Dec 14, 2013 at 10:44:10PM +0100, Stefan Esser wrote: > > 2) SSH logins are very slow, many seconds of delay between > connect and password prompt, several seconds after password > entry until a command prompt appears (normally > instantaneous) > Ah, so that explains the behavior I'm see. Just updated a circa Aug 3rd i386 FreeBSD to top-of-tree. My ssh logins to my work system take 30+ seconds now. :( >>> >>> You may want to test the attached patch, which reverts the >>> above mentioned commit. >>> >> >> I probably won't get to it until tomorrow, because I had started >> a dog-food system purge including re-installing all ports. The >> laptop takes a bit a time to recompile everything. >> > > Are you all running i386, compiled with gcc ? I'm on -CURRENT, CLANG, amd64. But since the problem has also been reported for i386 compiled with GCC, there seems to be some problem in common kernel code, that has been uncovered by your change, BTW: I remember seeing two wait channels being reported when I type ^T during multi-user startup (sa-spamd, which needs 140 seconds with the broken kernel: nanosleep kqueue (I do not remember whether this name is exact or abbreviated) I've been assuming that the problem might actually be in nanosleep(), since this is a timing related function and we are seeing huge delays, but eventually the delayed action succeeds. But a kernel with only kern_conf.c compiled with -fno-strict-overflow did not show the delays. I do not have time for further tests, today. Regards, STefan ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: SVN commit 259045 breaks -CURRENT
On Sun, 15 Dec 2013 07:47:22 +0200 Konstantin Belousov wrote: > On Sat, Dec 14, 2013 at 02:16:27PM -0800, Steve Kargl wrote: > > On Sat, Dec 14, 2013 at 11:11:15PM +0100, Stefan Esser wrote: > > > Am 14.12.2013 22:59, schrieb Steve Kargl: > > > > On Sat, Dec 14, 2013 at 10:44:10PM +0100, Stefan Esser wrote: > > > >> > > > >> 2) SSH logins are very slow, many seconds of delay between connect > > > >>and password prompt, several seconds after password entry until > > > >>a command prompt appears (normally instantaneous) > > > >> > > > > > > > > Ah, so that explains the behavior I'm see. Just updated a circa Aug 3rd > > > > i386 FreeBSD to top-of-tree. My ssh logins to my work system take 30+ > > > > seconds now. :( > > > > > > You may want to test the attached patch, which reverts the above > > > mentioned commit. > > > > > > > I probably won't get to it until tomorrow, because I had started > > a dog-food system purge including re-installing all ports. The > > laptop takes a bit a time to recompile everything. > > > > Are you all running i386, compiled with gcc ? I'm running amd64 compiled with clang; uname -a: FreeBSD xxx 11.0-CURRENT FreeBSD 11.0-CURRENT #15 r258254:259095M: Sun Dec 8 12:11:33 CET 2013 xxx:/usr/obj/usr/src/sys/SUBMARINE_SMP amd64 I'm recompiling right now to see if maybe I'm having a different issue. Bye Marc ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: SVN commit 259045 breaks -CURRENT
On Sun, Dec 15, 2013 at 08:43:22AM +0200, Konstantin Belousov wrote: > On Sat, Dec 14, 2013 at 09:56:04PM -0800, Steve Kargl wrote: > > > > The Aug 3rd system was built with gcc. The system upgrade I > > did this morning is using clang. I rebuilt everything and > > delete old things with delete-old and delete-old-libs. > > And, the clang-built kernel times out the connections ? On i386 ? With a clang built kernel, it takes 30+ seconds for a ssh to connect to my work system. Prior to the upgrade, connection speed was unnoticable. I won't be able to test Stefan suggested fix until tomorrow. I'm currently chasing down iconv issues. -- Steve ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: SVN commit 259045 breaks -CURRENT
On Sat, Dec 14, 2013 at 09:56:04PM -0800, Steve Kargl wrote: > On Sun, Dec 15, 2013 at 07:47:22AM +0200, Konstantin Belousov wrote: > > On Sat, Dec 14, 2013 at 02:16:27PM -0800, Steve Kargl wrote: > > > On Sat, Dec 14, 2013 at 11:11:15PM +0100, Stefan Esser wrote: > > > > Am 14.12.2013 22:59, schrieb Steve Kargl: > > > > > On Sat, Dec 14, 2013 at 10:44:10PM +0100, Stefan Esser wrote: > > > > >> > > > > >> 2) SSH logins are very slow, many seconds of delay between connect > > > > >>and password prompt, several seconds after password entry until > > > > >>a command prompt appears (normally instantaneous) > > > > >> > > > > > > > > > > Ah, so that explains the behavior I'm see. Just updated a circa Aug > > > > > 3rd > > > > > i386 FreeBSD to top-of-tree. My ssh logins to my work system take 30+ > > > > > seconds now. :( > > > > > > > > You may want to test the attached patch, which reverts the above > > > > mentioned commit. > > > > > > > > > > I probably won't get to it until tomorrow, because I had started > > > a dog-food system purge including re-installing all ports. The > > > laptop takes a bit a time to recompile everything. > > > > > > > Are you all running i386, compiled with gcc ? > > The Aug 3rd system was built with gcc. The system upgrade I > did this morning is using clang. I rebuilt everything and > delete old things with delete-old and delete-old-libs. And, the clang-built kernel times out the connections ? On i386 ? pgp7zUKeqeQtW.pgp Description: PGP signature
Re: SVN commit 259045 breaks -CURRENT
On Sun, Dec 15, 2013 at 07:47:22AM +0200, Konstantin Belousov wrote: > On Sat, Dec 14, 2013 at 02:16:27PM -0800, Steve Kargl wrote: > > On Sat, Dec 14, 2013 at 11:11:15PM +0100, Stefan Esser wrote: > > > Am 14.12.2013 22:59, schrieb Steve Kargl: > > > > On Sat, Dec 14, 2013 at 10:44:10PM +0100, Stefan Esser wrote: > > > >> > > > >> 2) SSH logins are very slow, many seconds of delay between connect > > > >>and password prompt, several seconds after password entry until > > > >>a command prompt appears (normally instantaneous) > > > >> > > > > > > > > Ah, so that explains the behavior I'm see. Just updated a circa Aug 3rd > > > > i386 FreeBSD to top-of-tree. My ssh logins to my work system take 30+ > > > > seconds now. :( > > > > > > You may want to test the attached patch, which reverts the above > > > mentioned commit. > > > > > > > I probably won't get to it until tomorrow, because I had started > > a dog-food system purge including re-installing all ports. The > > laptop takes a bit a time to recompile everything. > > > > Are you all running i386, compiled with gcc ? The Aug 3rd system was built with gcc. The system upgrade I did this morning is using clang. I rebuilt everything and delete old things with delete-old and delete-old-libs. -- Steve ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: SVN commit 259045 breaks -CURRENT
On Sat, Dec 14, 2013 at 02:16:27PM -0800, Steve Kargl wrote: > On Sat, Dec 14, 2013 at 11:11:15PM +0100, Stefan Esser wrote: > > Am 14.12.2013 22:59, schrieb Steve Kargl: > > > On Sat, Dec 14, 2013 at 10:44:10PM +0100, Stefan Esser wrote: > > >> > > >> 2) SSH logins are very slow, many seconds of delay between connect > > >>and password prompt, several seconds after password entry until > > >>a command prompt appears (normally instantaneous) > > >> > > > > > > Ah, so that explains the behavior I'm see. Just updated a circa Aug 3rd > > > i386 FreeBSD to top-of-tree. My ssh logins to my work system take 30+ > > > seconds now. :( > > > > You may want to test the attached patch, which reverts the above > > mentioned commit. > > > > I probably won't get to it until tomorrow, because I had started > a dog-food system purge including re-installing all ports. The > laptop takes a bit a time to recompile everything. > Are you all running i386, compiled with gcc ? pgp1bGAwXfMjJ.pgp Description: PGP signature
Re: SVN commit 259045 breaks -CURRENT
On Sat, 14 Dec 2013 13:59:04 -0800 Steve Kargl wrote: > On Sat, Dec 14, 2013 at 10:44:10PM +0100, Stefan Esser wrote: > > I noticed a severe slowdown and network problems on my amd64 -CURRENT > > system. By bisecting SVN revisions I identified the following commit > > to be responsible: > > > > -- > > r259045 | kib | 2013-12-06 22:44:13 +0100 (Fri, 06 Dec 2013) | 9 lines > > > > Disallow optimizations which potentially remove boundary checks > > for signed values due to a compiler authors considering integer > > overflow as impossible. > > > > The change follows suit of other projects taking the same measure. > > -- > > > > This commit added the following line to /sys/conf/kern.mk: > > > > CFLAGS+= -fno-strict-overflow > > > > > > The most obvious symptoms of the problem on my system are: > > > > 1) sa-spamd needs > 140 seconds to start > >(instead of a few seconds) > > > > 2) SSH logins are very slow, many seconds of delay between connect > >and password prompt, several seconds after password entry until > >a command prompt appears (normally instantaneous) > > > > Ah, so that explains the behavior I'm see. Just updated a circa Aug 3rd > i386 FreeBSD to top-of-tree. My ssh logins to my work system take 30+ > seconds now. :( I observe dnsmasq causing extremely high network latency without any visible reason - though that may not be related. I'll remove -fno-strict-overflow, recompile kernel and see if anything changes. Bye Marc ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: SVN commit 259045 breaks -CURRENT
On Sat, Dec 14, 2013 at 11:11:15PM +0100, Stefan Esser wrote: > Am 14.12.2013 22:59, schrieb Steve Kargl: > > On Sat, Dec 14, 2013 at 10:44:10PM +0100, Stefan Esser wrote: > >> > >> 2) SSH logins are very slow, many seconds of delay between connect > >>and password prompt, several seconds after password entry until > >>a command prompt appears (normally instantaneous) > >> > > > > Ah, so that explains the behavior I'm see. Just updated a circa Aug 3rd > > i386 FreeBSD to top-of-tree. My ssh logins to my work system take 30+ > > seconds now. :( > > You may want to test the attached patch, which reverts the above > mentioned commit. > I probably won't get to it until tomorrow, because I had started a dog-food system purge including re-installing all ports. The laptop takes a bit a time to recompile everything. -- Steve ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: SVN commit 259045 breaks -CURRENT
Am 14.12.2013 22:59, schrieb Steve Kargl:
> On Sat, Dec 14, 2013 at 10:44:10PM +0100, Stefan Esser wrote:
>> I noticed a severe slowdown and network problems on my amd64 -CURRENT
>> system. By bisecting SVN revisions I identified the following commit
>> to be responsible:
>>
>> --
>> r259045 | kib | 2013-12-06 22:44:13 +0100 (Fri, 06 Dec 2013) | 9 lines
>>
>> Disallow optimizations which potentially remove boundary checks
>> for signed values due to a compiler authors considering integer
>> overflow as impossible.
>>
>> The change follows suit of other projects taking the same measure.
>> --
>>
>> This commit added the following line to /sys/conf/kern.mk:
>>
>> CFLAGS+= -fno-strict-overflow
>>
>>
>> The most obvious symptoms of the problem on my system are:
>>
>> 1) sa-spamd needs > 140 seconds to start
>>(instead of a few seconds)
>>
>> 2) SSH logins are very slow, many seconds of delay between connect
>>and password prompt, several seconds after password entry until
>>a command prompt appears (normally instantaneous)
>>
>
> Ah, so that explains the behavior I'm see. Just updated a circa Aug 3rd
> i386 FreeBSD to top-of-tree. My ssh logins to my work system take 30+
> seconds now. :(
You may want to test the attached patch, which reverts the above
mentioned commit.
Regards, STefan
Index: /sys/conf/kern.mk
===
--- /sys/conf/kern.mk (revision 259396)
+++ /sys/conf/kern.mk (working copy)
@@ -148,12 +148,6 @@
CFLAGS+= -ffreestanding
#
-# Do not allow a compiler to optimize out overflow checks for signed
-# types.
-#
-CFLAGS+= -fno-strict-overflow
-
-#
# GCC SSP support
#
.if ${MK_SSP} != "no" && ${MACHINE_CPUARCH} != "ia64" && \
___
freebsd-current@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: SVN commit 259045 breaks -CURRENT
On Sat, Dec 14, 2013 at 10:44:10PM +0100, Stefan Esser wrote: > I noticed a severe slowdown and network problems on my amd64 -CURRENT > system. By bisecting SVN revisions I identified the following commit > to be responsible: > > -- > r259045 | kib | 2013-12-06 22:44:13 +0100 (Fri, 06 Dec 2013) | 9 lines > > Disallow optimizations which potentially remove boundary checks > for signed values due to a compiler authors considering integer > overflow as impossible. > > The change follows suit of other projects taking the same measure. > -- > > This commit added the following line to /sys/conf/kern.mk: > > CFLAGS+= -fno-strict-overflow > > > The most obvious symptoms of the problem on my system are: > > 1) sa-spamd needs > 140 seconds to start >(instead of a few seconds) > > 2) SSH logins are very slow, many seconds of delay between connect >and password prompt, several seconds after password entry until >a command prompt appears (normally instantaneous) > Ah, so that explains the behavior I'm see. Just updated a circa Aug 3rd i386 FreeBSD to top-of-tree. My ssh logins to my work system take 30+ seconds now. :( -- steve -- Steve ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
SVN commit 259045 breaks -CURRENT
I noticed a severe slowdown and network problems on my amd64 -CURRENT system. By bisecting SVN revisions I identified the following commit to be responsible: -- r259045 | kib | 2013-12-06 22:44:13 +0100 (Fri, 06 Dec 2013) | 9 lines Disallow optimizations which potentially remove boundary checks for signed values due to a compiler authors considering integer overflow as impossible. The change follows suit of other projects taking the same measure. -- This commit added the following line to /sys/conf/kern.mk: CFLAGS+= -fno-strict-overflow The most obvious symptoms of the problem on my system are: 1) sa-spamd needs > 140 seconds to start (instead of a few seconds) 2) SSH logins are very slow, many seconds of delay between connect and password prompt, several seconds after password entry until a command prompt appears (normally instantaneous) In general it takes many seconds to open a TCP socket, even to localhost. I can perform further tests on this system, but it will be a lot of work to locate the source files that are mis-compiled with -fno-strict-overflow. I'm surprised that nobody else seems to be affected by this problem, since it is very obvious on my system and clearly caused by the above mentioned commit. My kernel configuration is a stripped down GENERIC plus ZFS, IPFW and LINUX emulation. I can provide full details and a kernel that exposes the problem on request. Regards, STefan ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
