[Bug 230491] stat(1): Improve performance with getpwuid() and getgrgid() caching
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=230491 Conrad Meyer changed: What|Removed |Added CC||c...@freebsd.org --- Comment #1 from Conrad Meyer --- Patch looks functionally ok to me (works because stat is a single-threaded program); has some style issues. It's unclear to me if we want to add caching to every individual tool accessing these databases, or if the caching should just happen in libc. One other concern that the thought of caching in libc raises is: how do we handle cache invalidation? Obviously we don't care if a user/group change races a running but ultimately bounded program like stat(1) across many files, but we probably do care if a daemon never sees renamed or numbered ids. Btw, it seems such caching is already implemented in libc via the NS_CACHING option. -- You are receiving this mail because: You are the assignee for the bug. ___ freebsd-bugs@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-bugs To unsubscribe, send any mail to "freebsd-bugs-unsubscr...@freebsd.org"
[Bug 230491] stat(1): Improve performance with getpwuid() and getgrgid() caching
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=230491 Kubilay Kocak changed: What|Removed |Added Summary|[patch] stat(1): Improve|stat(1): Improve |performance with getpwuid() |performance with getpwuid() |and getgrgid() caching |and getgrgid() caching Status|New |Open Keywords||needs-qa, performance Flags||mfc-stable10?, ||mfc-stable11? -- You are receiving this mail because: You are the assignee for the bug. ___ freebsd-bugs@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-bugs To unsubscribe, send any mail to "freebsd-bugs-unsubscr...@freebsd.org"
[Bug 230492] make: -C option doesn't work as explained in man page about value of ${.CURDIR} when logical and physical directory are different.
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=230492 Yasuhiro KIMURA changed: What|Removed |Added Summary|make: -C options doesn't|make: -C option doesn't |work as explained in man|work as explained in man |page about value of |page about value of |${.CURDIR} when logical and |${.CURDIR} when logical and |physical directory are |physical directory are |different. |different. -- You are receiving this mail because: You are the assignee for the bug. ___ freebsd-bugs@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-bugs To unsubscribe, send any mail to "freebsd-bugs-unsubscr...@freebsd.org"
[Bug 230492] make: -C options doesn't work as explained in man page about value of ${.CURDIR} when logical and physical directory are different.
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=230492 Bug ID: 230492 Summary: make: -C options doesn't work as explained in man page about value of ${.CURDIR} when logical and physical directory are different. Product: Base System Version: CURRENT Hardware: Any OS: Any Status: New Severity: Affects Some People Priority: --- Component: bin Assignee: b...@freebsd.org Reporter: y...@utahime.org In the man page of make(1) '-C' options is explained as following. -- -C directory Change to directory before reading the makefiles or doing anything else. If multiple -C options are specified, each is interpreted relative to the previous one: -C / -C etc is equivalent to -C /etc. -- This means 'make -C /foo/bar/baz' works same as 'cd /foo/bar/baz; make'. But there is an case that make doesn't behaves so. As explained in summary it is about value of ${.CURDIR} and happens when logical and physical directory are different. Let me explain it by providing an example. -- yasu@rolling-vm-freebsd1[2018]% pwd /home/yasu/tmp yasu@rolling-vm-freebsd1[2019]% ls -l total 1 lrwxr-xr-x 1 yasu user 3 8月 10 08:21 bar@ -> foo drwxr-xr-x 2 yasu user 3 8月 10 08:20 foo/ -- /home/yasu/tmp/foo is real directory and /home/yasu/tmp/bar is symbolic link to it. So logical and physical expression of /home/yasu/tmp/bar is different. -- yasu@rolling-vm-freebsd1[2020]% cd /home/yasu/tmp/bar yasu@rolling-vm-freebsd1[2021]% /bin/pwd -L /home/yasu/tmp/bar yasu@rolling-vm-freebsd1[2022]% /bin/pwd -P /home/yasu/tmp/foo -- And there is following Makefile under /home/yasu/tmp/bar. -- yasu@rolling-vm-freebsd1[2023]% cat Makefile ~/tmp/bar all: @echo ${.CURDIR} -- It simply displays value of ${.CURDIR}. Now let's execute make without any arguments. It causes following result. -- yasu@rolling-vm-freebsd1[2024]% make ~/tmp/bar /home/yasu/tmp/foo -- Physical expression of current working directory is displayed. Next, let's execute make with -C options and logical expression of current working directory. In this case it result as following. -- yasu@rolling-vm-freebsd1[2025]% make -C /home/yasu/tmp/bar ~/tmp/bar /home/yasu/tmp/bar -- If '-C' options works exactly as explained in man page, same value as previous execution, that is, physical expression of current working directory should be displayed. Actually, however, what is displayed is logical expression of current working directory. This means '-C' option doesn't work as explained in man page. I confirmed it happens on both 11.2-RELEASE and 12.0-CURRENT (r337160). -- You are receiving this mail because: You are the assignee for the bug. ___ freebsd-bugs@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-bugs To unsubscribe, send any mail to "freebsd-bugs-unsubscr...@freebsd.org"
[Bug 230491] [patch] stat(1): Improve performance with getpwuid() and getgrgid() caching
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=230491 Bug ID: 230491 Summary: [patch] stat(1): Improve performance with getpwuid() and getgrgid() caching Product: Base System Version: 11.2-STABLE Hardware: Any OS: Any Status: New Severity: Affects Only Me Priority: --- Component: bin Assignee: b...@freebsd.org Reporter: t...@hur.st Created attachment 196041 --> https://bugs.freebsd.org/bugzilla/attachment.cgi?id=196041=edit stat(1) caching (applies to STABLE and CURRENT) This patch for stat(1) adds very simple caching to its calls to getpwuid() and getgrgid(), for a substantial performance boost when user/group is displayed. With this test command on a cached ZFS dataset: time (find /usr/src -type f -print0 | xargs -0 stat >/dev/null) I see run time reduce from 22 seconds to 8.5 seconds. On the OpenBSD CVS repository with files owned by a normal user I see an even more pronounced difference: 55 seconds reduced to 14 - I guess these functions have runtime proportional to how far down the user/group lists they look? sysutils/coreutils gnustat shows identical performance to stock FreeBSD stat(1), so could probably benefit from similar changes. Thanks to Vall on FreeNode #freebsd for reporting the performance problem. -- You are receiving this mail because: You are the assignee for the bug. ___ freebsd-bugs@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-bugs To unsubscribe, send any mail to "freebsd-bugs-unsubscr...@freebsd.org"
[Bug 200109] [PATCH] Minor EFI boot image enhancements
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=200109 --- Comment #5 from Pedro F. Giffuni --- (In reply to Kyle Evans from comment #4) Heh ... and as I wrote .. "I won't feel offended if the patch is rejected" :). -- You are receiving this mail because: You are the assignee for the bug. ___ freebsd-bugs@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-bugs To unsubscribe, send any mail to "freebsd-bugs-unsubscr...@freebsd.org"
[Bug 229000] update and replace old rc daemons by google_network_daemon in relase/tools/gce.conf
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=229000 Glen Barber changed: What|Removed |Added Status|New |In Progress --- Comment #10 from Glen Barber --- Committed, thanks. Will MFC in 3 days. Sorry for the delay. -- You are receiving this mail because: You are the assignee for the bug. ___ freebsd-bugs@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-bugs To unsubscribe, send any mail to "freebsd-bugs-unsubscr...@freebsd.org"
[Bug 229000] update and replace old rc daemons by google_network_daemon in relase/tools/gce.conf
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=229000 --- Comment #9 from commit-h...@freebsd.org --- A commit references this bug: Author: gjb Date: Thu Aug 9 23:31:18 UTC 2018 New revision: 337555 URL: https://svnweb.freebsd.org/changeset/base/337555 Log: Update and replace old rc daemons for GCE images. PR: 229000 Submitted by: helen.ko...@collabora.com MFC after:3 days Sponsored by: The FreeBSD Foundation Changes: head/release/tools/gce.conf -- You are receiving this mail because: You are the assignee for the bug. ___ freebsd-bugs@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-bugs To unsubscribe, send any mail to "freebsd-bugs-unsubscr...@freebsd.org"
[Bug 230489] [patch] bin/cat compiled with -DNO_UDOM_SUPPORT will fail to compile as it is missing #include
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=230489 will.s...@gmail.com changed: What|Removed |Added Attachment #196038|0 |1 is obsolete|| --- Comment #1 from will.s...@gmail.com --- Created attachment 196039 --> https://bugs.freebsd.org/bugzilla/attachment.cgi?id=196039=edit patch to cat.c to include errno.h for all paths got old and new reversed on prior attachment, apologies - first bug report... ever. -- You are receiving this mail because: You are the assignee for the bug. ___ freebsd-bugs@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-bugs To unsubscribe, send any mail to "freebsd-bugs-unsubscr...@freebsd.org"
[Bug 230489] [patch] bin/cat compiled with -DNO_UDOM_SUPPORT will fail to compile as it is missing #include
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=230489 Bug ID: 230489 Summary: [patch] bin/cat compiled with -DNO_UDOM_SUPPORT will fail to compile as it is missing #include Product: Base System Version: 11.2-RELEASE Hardware: Any OS: Any Status: New Severity: Affects Some People Priority: --- Component: bin Assignee: b...@freebsd.org Reporter: will.s...@gmail.com Created attachment 196038 --> https://bugs.freebsd.org/bugzilla/attachment.cgi?id=196038=edit patch to cat.c to include errno.h for all paths #include is required by all conditions. -- You are receiving this mail because: You are the assignee for the bug. ___ freebsd-bugs@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-bugs To unsubscribe, send any mail to "freebsd-bugs-unsubscr...@freebsd.org"
[Bug 229000] update and replace old rc daemons by google_network_daemon in relase/tools/gce.conf
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=229000 --- Comment #8 from Helen Koike --- ping? -- You are receiving this mail because: You are the assignee for the bug. ___ freebsd-bugs@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-bugs To unsubscribe, send any mail to "freebsd-bugs-unsubscr...@freebsd.org"
[Bug 230487] jail -m works incorrectly in some cases
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=230487 Bug ID: 230487 Summary: jail -m works incorrectly in some cases Product: Base System Version: 11.2-RELEASE Hardware: amd64 OS: Any Status: New Severity: Affects Only Me Priority: --- Component: bin Assignee: b...@freebsd.org Reporter: jasonma...@gmail.com It looks like a `jail -m` is not correctly handling jail.conf and then sets incorrect values to a running jail. Example, here the `php` jail is getting the wrong hostname, ip6 address, and devfs ruleset. root# jls -a JID IP Address Hostname Path 1 localhost /zroot/webserver/jails/mysql 8 129.174.130.141 www.ccsa.gmu.edu /zroot/webserver/jails/php root# jail -mv mysql: jail_set(JAIL_UPDATE) jid=1 name=mysql securelevel=2 host.hostname=localhost devfs_ruleset=5 mysql: updated php: jail_set(JAIL_UPDATE) jid=8 name=php securelevel=2 host.hostname="" ip4.addr=129.174.130.141 ip6.addr="" devfs_ruleset=0 php: updated root# jls -a JID IP Address Hostname Path 1 localhost /zroot/webserver/jails/mysql 8 129.174.130.141 /zroot/webserver/jails/php /etc/jail.conf: mysql { securelevel=2; host.hostname="localhost"; ip4 = disable; ip6 = disable; path = "/var/jails/mysql"; devfs_ruleset=5; mount.devfs; exec.start = "/usr/sbin/daemon -c -f /usr/local/bin/mysqld_safe --user=mysql --datadir=/var/db --pid-file=/var/run/mysql.pid --explicit_defaults_for_timestamp=true"; exec.stop = "/bin/pkill -TERM -F /var/run/mysql.pid"; } php { securelevel=2; host.hostname="www.ccsa.gmu.edu"; ip4.addr=129.174.130.141; ip6.addr=2620:10e:6024:f004::141; path = "/var/jails/php"; devfs_ruleset=5; mount.devfs; mount.fstab = "/var/jails/php/etc/fstab"; exec.start = "/usr/sbin/php-fpm"; exec.stop = "/bin/pkill -QUIT -F /var/run/php-fpm.pid"; } -- You are receiving this mail because: You are the assignee for the bug. ___ freebsd-bugs@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-bugs To unsubscribe, send any mail to "freebsd-bugs-unsubscr...@freebsd.org"
[Bug 230008] [panic] [fdescfs] Page fault in vn_finished_write+0x13
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=230008 Mark Johnston changed: What|Removed |Added CC||ma...@freebsd.org --- Comment #1 from Mark Johnston --- Looks like this occurred because mp->mnt_op == NULL (though it's non-null in the kernel core). MNTK_REFEXPIRE|MNTK_UNMOUNTF is set on the mount, so it seems this was a race with a free of the mountpoint. -- You are receiving this mail because: You are the assignee for the bug. ___ freebsd-bugs@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-bugs To unsubscribe, send any mail to "freebsd-bugs-unsubscr...@freebsd.org"
[Bug 220246] syslogd does not send RFC3164-conformant messages [PATCH]
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=220246 --- Comment #2 from Ed Schouten --- Hi there, Thanks for the patch! The syslogd code has been refactored a lot lately to support the RFC 5424 message format. Looking at the code, the issue still applies. That said, I'm a bit hesitant to change anything for the RFC 3164 format support. People expect that it works in a certain way and changing that seems like a bad idea. Could you please give the RFC 5424 support that's present in 11-STABLE and HEAD a try? Just add "-O rfc5424" to syslogd_flags in rc.conf. Please let me know whether that addresses the issue for you sufficiently. -- You are receiving this mail because: You are the assignee for the bug. ___ freebsd-bugs@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-bugs To unsubscribe, send any mail to "freebsd-bugs-unsubscr...@freebsd.org"
[Bug 215143] FreeBSD 11 install cd does not boot on macpro1,1 (32-bit efi)
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=215143 --- Comment #12 from Jason W. Bacon --- The patch works like a charm, thanks. I commented out these lines and followed the instructions in the release man page. I few hours later, I had a FreeBSD-11.1-RELEASE-p11-amd64-disc1.iso that boots on my old iMac. Next question: Is there any chance freebsd-update will break a system installed from the modified ISO by "restoring" the missing EFI files? -- You are receiving this mail because: You are the assignee for the bug. ___ freebsd-bugs@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-bugs To unsubscribe, send any mail to "freebsd-bugs-unsubscr...@freebsd.org"
[Bug 225932] uplcom Prolific PL2302 doesn't support high baud rates
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=225932 --- Comment #5 from Gabor Simon --- (In reply to Kyle Evans from comment #4) Hi, Here's the revision link: https://reviews.freebsd.org/D16639 I hope I did it according to the process. I haven't used Phabricator before, so if I missed something, please give a hint how to do it right :). Regards, Gabor Simon -- You are receiving this mail because: You are the assignee for the bug. ___ freebsd-bugs@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-bugs To unsubscribe, send any mail to "freebsd-bugs-unsubscr...@freebsd.org"
[Bug 210537] [patch] [feature request] set MIME type in cron-generated e-mails
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=210537 --- Comment #24 from Eugene Grosbein --- (In reply to Mikhail Teterin from comment #23) While I do not object against the feature and understand if could be useful for some part of user base, I still do not think it is cron's job to check and/or correct MIME format of job's output. The feature could be implemented as external filter processing an output using a pipe, if needed. And if it is embedded into the cron anyway, it should impose no regressions on correctness and performance for general case. I'm not going to develop that myself but I would commit the patch in form of correct and low-overhead optional feature. -- You are receiving this mail because: You are the assignee for the bug. ___ freebsd-bugs@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-bugs To unsubscribe, send any mail to "freebsd-bugs-unsubscr...@freebsd.org"