Commit by bruno :: r205168 /stable/8/sys/kern/kern_thr.c: (link)

MFC r204519:
Deliver siginfo when signal is generated by thr_kill(2) (SI_USER with properly
filled si_uid and si_pid).

PR: 141956

Commit by bruno :: r204519 /head/sys/kern/kern_thr.c: (link)

Deliver siginfo when signal is generated by thr_kill(2) (SI_USER with properly
filled si_uid and si_pid).

Reported by: Joel Bertrand <joel.bertrand systella fr>
PR: 141956
Reviewed by: kib
MFC after: 2 weeks

Commit by bruno :: r2577 /branches/2.2.9/mindi/mindi:

• First try to improve OVM support with ocfs2 and Xen modules added

Commit by bruno :: r2576 /branches/2.2.10/ (3 files in 2 dirs):

svn merge -r 2567:2574 /mondorescue/branches/2.2.9
(At parity with 2.2.9.2)

• Improve device exclusion for LVM by adding support for symlinks and mapper systematically at all levels (PV, VG, LV). Should really provvide the best support possible for LVM exclusion.
• Also use the same exclusion technique for MINDI_EXCLUDE_DEVS in mindi itself just in case.
• ums-cypress also added to module list following user feedback
• Fix a bug in the vgcreate command generation where too many PVs were mentioned, due to the extension of the list. We now use a $current_PVs variable which is limited to what i needed
• setfacl doesn't need the -h option to restore physical paths
• THis version of mondo uses another interface for the mr_asprintf function than the one used just previously which was coredumping
• Adds support for sata_svw for Proliant DL 385 G5p. Fixes #391.

Commit by bruno :: r2575 /tags/2.2.9.2:

Tag official version 2.2.9.2

Commit by bruno :: r2574 /branches/2.2.9/mindi/mindi:

• Adds support for sata_svw for Porliant DL 385 G5p. Fixes #391.

Commit by bruno :: r2573 /branches/2.2.9/mondo/src/common/libmondo-filelist.c:

• THis version of mondo uses another interface for the mr_asprintf function than the one used just previously which was coredumping

Commit by bruno :: r2572 /branches/2.2.9/mondo/src/common/libmondo-filelist.c:

setfacl doesn't need the -h option to restore physical paths

Commit by bruno :: r2571 /branches/2.2.9/mindi/analyze-my-lvm:

• Fix a bug in the vgcreate command generation where too many PVs were mentioned, due to the extension of the list. We now use a $current_PVs variable which is limited to what i needed

Commit by bruno :: r2570 /branches/2.2.9/mindi/mindi:

ums-cypress also added to module list following user feedback

Commit by bruno :: r2569 /branches/2.2.10/ (8 files in 4 dirs):

svn merge -r 2545:2567 /mondorescue/branches/2.2.9

• Improve device exclusion for LVM by adding support for symlinks and mapper systematically at all levels (PV, VG, LV). Shoul d really provvide the best support possible for LVM exclusion.
• Now supports also exclusion of LVs directly
• Improve mindi by handling excluded devices earlier in the loop.
• Fix LV exclusion when VGs are excluded (the excludedvgs file has to be removed at start and end of analyze-my-lvm not in fu nctions, as they are called multiple times)
• Rename some variables exported to avoid accidental conflict (EXCLUDE_DEVS => MINDI_EXCLUDE_DEVS and ADDITIONAL_BOOT_PARAMS => MINDI_ADDITIONAL_BOOT_PARAMS)
• setfattr also needs the -h flag to restore attributes on symlinks, not on the target file (which may not exist BTW). Should solve #388.
• Using forcemods="mod1 mod2" will now do something and load those modules first
• Update web page on distributions to download
• Add support for ums_cypress module in mindi
• Fix acl backup which was completely wrong.
• Fix a bug where LV were not created anymore if no exclusion !!
• In analyze-my-lvm only use pv.tmp2 when it exists to avoid error msgs -Adds support for links to modules (.ko pointing to .o) such as with VMWare extensions
• Adds i-want-my-lvm content in mindi log file
• Adds vmxnet3 driver support to mindi for some VMWare versions
• Fix a typo in the way device mapper files where excluded
• use option -h of getfattr to *not* follow symlinks, which breaks RHEL 5.4 as reported in #388. However, getfacl doesn't provide such an option.
• Adds some logs for filelist creation
• Remove option -P from getfattr which may also skip symlinks - Attempt to solve #388
• Exclude rpc_pipefs type of filesystems from find
• Suppress unused mode_of_file function

Commit by bruno :: r2568 /branches/2.2.9/mindi/mindi:

• Also use the same exclusion technique for MINDI_EXCLUDE_DEVS in mindi itself just in case.

Commit by bruno :: r2567 /branches/2.2.9/mindi/analyze-my-lvm:

• Improve device exclusion for LVM by adding support for symlinks and mapper systematically at all levels (PV, VG, LV). Should really provvide the best support possible for LVM exclusion.

Commit by bruno :: r2566 /branches/2.2.9/mindi/ (mindi analyze-my-lvm):

• Now supports also excludion of LVs directly
• Improve mindi by handling excluded devices earlier in the loop.

Commit by bruno :: r2565 /branches/2.2.9/mindi/analyze-my-lvm:

• Fix LV exclusion when VGs are excluded (the excludedvgs file has to be removed at start and end of analyze-my-lvm not in functions, as they are called multiple times)

Commit by bruno :: r2564 /branches/2.2.9/mindi/ (rootfs/sbin/init mindi analyze-my-lvm):

• Rename some variables exported to avoid accidental conflict (EXCLUDE_DEVS => MINDI_EXCLUDE_DEVS and ADDITIONAL_BOOT_PARAMS => MINDI_ADDITIONAL_BOOT_PARAMS)

Commit by bruno :: r2563 /branches/2.2.9/mondo/src/common/libmondo-filelist.c:

• setfattr also needs the -h flag to restore attributes on symlinks, not on the target file (which may not exist BTW). Should solve #388.

Commit by bruno :: r2562 /branches/2.2.9/mindi/rootfs/sbin/init:

• Using forcemods="mod1 mod2" will now do something and load those modules first

Commit by bruno :: r2561 /branches/2.2.9/website/downloads.shtml:

• Update web page on distributions to download

Commit by bruno :: r2560 /branches/2.2.9/mindi/mindi:

• Add support for ums_cypress module in mindi