Difference between revisions of "Download/kernel/rhel6-testing/042stab080.3/changes"
< Download | kernel | rhel6-testing | 042stab080.3
(explain CVE-2013-2224) |
m (explain rtnl_lock release lost fix) |
||
(2 intermediate revisions by the same user not shown) | |||
Line 5: | Line 5: | ||
* [x86] ARAT feature bit to be set on AMD CPUs: performance boost on AMD Opteron 62xx-based systems | * [x86] ARAT feature bit to be set on AMD CPUs: performance boost on AMD Opteron 62xx-based systems | ||
* [nfsd] return ext4 file system superblock time granularity on FSINFO request => performance boost because inodes are not revalidated most of time (PCLIN-31863) | * [nfsd] return ext4 file system superblock time granularity on FSINFO request => performance boost because inodes are not revalidated most of time (PCLIN-31863) | ||
− | * [net] | + | * [net] fix freeing RCU-protected IP-options (same fix as in 042stab078.27; CVE-2013-2224, {{B|2648}}) |
* [relay] consequent execution of blktrace failure has been fixed (PSBM-20247) | * [relay] consequent execution of blktrace failure has been fixed (PSBM-20247) | ||
− | * [net] rtnl_lock release lost on error path (PSBM-21008) | + | * [net] rtnl_lock release lost on error path (same fix as in 042stab078.28; PSBM-21008) |
* [ubc/net] tcpsndbuf race between poll and send cause caused counters skew and taints kernel on bc release attempt (PCLIN-31931) | * [ubc/net] tcpsndbuf race between poll and send cause caused counters skew and taints kernel on bc release attempt (PCLIN-31931) | ||
* [stat] replaces global counters with per-cpu structures to avoid kstat_glb_lock congestion => performance increases | * [stat] replaces global counters with per-cpu structures to avoid kstat_glb_lock congestion => performance increases | ||
Line 13: | Line 13: | ||
* [vedev] if devices is removed from CT, it's sysfs part should be removed too (PSBM-19097) | * [vedev] if devices is removed from CT, it's sysfs part should be removed too (PSBM-19097) | ||
* [pfcache] do not crash and do not fail mounting new filesystems if cache are not found (PSBM-19581) | * [pfcache] do not crash and do not fail mounting new filesystems if cache are not found (PSBM-19581) | ||
− | * [nfs/quota] NFS local quota (vzquota in case /vz | + | * [nfs/quota] NFS local quota (vzquota in case /vz over NFS) significantly reworked => no more crashes / stats skews are expected (PCLIN-31747) |
* [ploop] kaio must merge bios for aio_kernel_submit => should avoid problem with massive 4k FUSE requests (PSBM-19207) | * [ploop] kaio must merge bios for aio_kernel_submit => should avoid problem with massive 4k FUSE requests (PSBM-19207) | ||
* [sysctl] add global sysctl 'fs.ve-mount-nr' (4096 by default) to customize per-CT limit for mount points | * [sysctl] add global sysctl 'fs.ve-mount-nr' (4096 by default) to customize per-CT limit for mount points |
Latest revision as of 00:19, 12 July 2013
Changes[edit]
Since 042stab080.2:
- [gcc] compilation with gcc 4.5 fixed
- [x86] ARAT feature bit to be set on AMD CPUs: performance boost on AMD Opteron 62xx-based systems
- [nfsd] return ext4 file system superblock time granularity on FSINFO request => performance boost because inodes are not revalidated most of time (PCLIN-31863)
- [net] fix freeing RCU-protected IP-options (same fix as in 042stab078.27; CVE-2013-2224, #2648)
- [relay] consequent execution of blktrace failure has been fixed (PSBM-20247)
- [net] rtnl_lock release lost on error path (same fix as in 042stab078.28; PSBM-21008)
- [ubc/net] tcpsndbuf race between poll and send cause caused counters skew and taints kernel on bc release attempt (PCLIN-31931)
- [stat] replaces global counters with per-cpu structures to avoid kstat_glb_lock congestion => performance increases
- [cgroups] allow to restrict CT permissions for devices inside a CT (PSBM-19097)
- [vedev] if devices is removed from CT, it's sysfs part should be removed too (PSBM-19097)
- [pfcache] do not crash and do not fail mounting new filesystems if cache are not found (PSBM-19581)
- [nfs/quota] NFS local quota (vzquota in case /vz over NFS) significantly reworked => no more crashes / stats skews are expected (PCLIN-31747)
- [ploop] kaio must merge bios for aio_kernel_submit => should avoid problem with massive 4k FUSE requests (PSBM-19207)
- [sysctl] add global sysctl 'fs.ve-mount-nr' (4096 by default) to customize per-CT limit for mount points