OBSOLETE Patch-ID# 150400-09

	***********************************************************
	NOTE: This patch may contain one or more OEM-specific platform ports.
	      See the appropriate OEM_NOTES file within the patch for
	      information specific to these platforms.
	      DO NOT INSTALL this patch on an OEM system if a corresponding
	      OEM_NOTES file is not present (or is present, but instructs not
	      to install the patch), unless the OEM vendor directs otherwise.
	***********************************************************

17628036 race between pollhead_delete() and strclose->pollhead_clean causes panic
 
(from 150400-08)
 
15414382 in.iked should exit with a non-zero code when checking against an invalid ike.config file
15449671 assertion failure in sd_unit_attach due to kstat_create failing
15566297 faulty PF_KEY replies should not cause in.iked to halt
15566298 PF_KEY diagnostics need to be more specific
15567469 get_ipsa_pair() does not always follow bucket lock entry rules, could potentially deadlock
15570951 ipsecalgs(1m) does not cope when there are no algorithms registered
15612419 ipsecconf -F failed with exit status 1
15716439 times() system call is not scalable
15794263 pmap -L does not scale
15809288 'newtask -p default -c $$' hangs when run within a S9 Branded Zone with cpu-shares
15816639 clock slowness issues with Fair Share Scheduling
15818472 threads favoring cpu0 tend to slow clock tick processing
16354357 clock scalability issues with CPU-capping enabled
16672077 clock tick processing should be optimized for SYSDC scheduling class
17075733 crypto framework should return mechanism not supported instead of bad arguments
17075773 in.iked(1M) outputs a message of CKR_ARGUMENTS_BAD in sun4v platform
17289900 regression in times() for multithreaded
17361787 ikev1 certutils should print out human-readable PKCS#11 errors
17816539 idle CPU's are failing to steal runnable threads from other CPU's
17974760 IKEv1 needs to give more insight into its certificate operations
17986274 IKEv1 cert chain validation with cert payloads needs work
 
(from 150400-07)
 
15041583 thread walker can loop if pointers change under its feet
15556414 self-encapsulated inbound ICMP fails upon reflection
15613053 implement ATA TRIM, SCSI Write Same / UNMAP , Thin Reclamation
15647156 ZFS should support unmap for ssd devices
15735442 support Logical Block Provisioning in Solaris SD driver
15767303 sd_get_lbp_attribute alignment error panic
15786808 deadlock between zfs_ioc_recv and zfs_ioc_rollback
15795037 sd_send_scsi_INQUIRY truncates pages
15795327 some zpool commands would hang when there is a suspended pool
15796962 zpool destroy hangs when performing 'unmap' on SSD
15804599 massive write slowdown on random SCSI write workloads due to SCSI unmap overhead on pool data devices
15819443 regression in pool version 32 with multiple slog devices and large sync writes
16068162 zfs_ioc_vdev_set_state should not POOL_CHECK_SUSPENDED before spa_open_policy()
17082774 Solaris 10 Kernel Patch 148888-01 V880 panic in sfmmu_mlspl_enter() recursive mutex_enter
17510986 panic at cu_pg_reconfig+0x88 on Athena system at suspend/resume test
17609380 double-free of PF_KEY mblk possible in low memory situations
17615534 64KB free page search occupies 100% CPU load with Solaris 10
17798090 recursive mutex_enter panic from dnlc_dir_add_entry()
17799335 ZFS increasing volsize causes panic in VirtualBox VM
 
(from 150400-06)
 
15706409 mpt should handle failing disks more intelligently
15714128 zpool import should not abort with overlapping partitions
15786729 Ozone: failed to cap power on Solaris 10 and 11 if hard cap is enabled before OS handoff
15808754 swap on zvol is slow under memory pressure as compared to raw swap impacting applications instantly
15816323 zpool import with zvols takes longer with each import
15875298 mpt does not recover from "Disconnected command timeout" with failing disks
15892821 ndi_devi_offline fails when SAS PHY is offlined forcibly
16245585 "Disconnected command timeout for Target" SCSI warning
16329143 'divide by zero' panic during 'zfs send'
17366564 problem with kernel (syscall)
 
(from 150400-05)
 
15710775 vdc driver should support DKIOCGMEDIAINFOEXT
15780493 tx_sync_thread's stack is sometimes not large enough
15800486 kstat's mixing of scaled/unscaled hrtime sacrifices monotonicity
15824910 add support to configure vdisk physical block size in vdc.conf
15824925 zvol should support DKIOCGMEDIAINFOEXT
15980058 FWARC/2012/285 VIO protocol changes
15992306 using sizeof(struct dk_minfo_ext) in kernel may corrupt memory of 32-bit user app
16250054 Live Migration failed if klmmod was not loaded
16633265 vdev_check event for a missing vdev causes syseventd:zfs_mod to crash
16635844 z_hold_mtx[] locks need to be on their own cache line
16716278 probe io:::done does not fire for asynchronous I/O
17034106 LGRP_MEM_POLICY_RANDOM on M10 is not random for Solaris 10
17361762 incorrect anon_cache buf_inuse value reported on Solaris 10
 
(from 150400-04)
 
15665972 hypervisor call 0x21 returned an unexpected error 2 due to invalid 2nd sfmmu_info
15704632 unable to remove huge objects
16091474 sd_xbuf scsi_pkt = 0, leaving cmd stuck in ssd driver with no callback assigned
16519289 Solaris e1000g needs to support Intel Clarkville
16713450 SD_MUTEX should not be dropped around call to timeout(9f)
17220973 zone_locked_mem gets negative when DISM memory is locked
 
(from 150400-03)
 
15511003 can't bind to socket from kernel thread because of NET_MAC_AWARE privilege
15518895 swap zvols need safeguards when changing their size
15722845 rebooting zones several times leaves behind lot of dormant ipsecesp kstats
15723155 cannot receive new filesystem stream: invalid backup stream
15740941 increasing volsize of swap zvol causes panic
15746589 ::vdev is not pipable
15746994 resizing swap with no dump device set causes panic
15751898 swap should prohibit using a subrange of a zvol
15756032 memory leaks in swap_mod
15766940 swapping to large extent causes panic
15775996 zone defrouter option is not working properly with link down
15787467 decreasing volsize of swap zvol causes panic
15796390 zio/arc join: buffer frees wrong cache
15823090 head panicked while ZFS evicting cache buffer no longer found in the hash
15825650 data missing from ZFS file system after a panic
16013601 replay of partial block writes can extend EOF to end of block
16038166 transient user data corruption due to anon_map_privatepages() race
16205576 problem with cpc driver
16532828 zvol_swap_004_pos fails on Solaris 10 patch bits
17214105 pgstat -v cap is not fixed if the load is over the maximum listed speed of CPU
17255714 bug needed for separate RTI for forthdebug deliverable related to RTI 383701
 
(from 150400-02)
 
15666664 allow partially mapped segments to share non-leaf page tables on x86
15689118 pgstat(1M) should report absolute maximum capacity of a PG
15720191 panic sfmmu_asm: valid SCD with no 3rd scd TSB
15758461 DISM segments should track page locking using bitmap
15764451 seg_spt should align inner and outer mapping VAs
15791822 RFE: enable hotplug service by default on sun4v platforms
15794950 system hang, threads blocked in DISM code
15812816 hotplug disable operation on a 16 CMU system can take 14-18 minutes
15816587 cputrack prints incorrect %tick
15819674 dynamic removal of root complex and slot with qlc/emlxs card fails due to a reference held by fmd
15819978 DDIO: add-io of a slot back to primary fails: ldma_iov_add: iov_add(500,@1/@0/@1) failed
15821644 kernel panics in pghw_cap_init when there are no hw groups
15823088 processes sharing the same ptables can race in x86 hat_share()
15826749 assertion failed: status == DDI_SUCCESS, file: ../../common/io/busra.c, line: 1342
15833828 I/O domain panic when dynamic removing a VF from this domain in primary
15833855 instance number of hxge/sxge changes after FEM is added back to primary domain
15857135 dynamic removal of root complex and slot can fail due to reference held by hald
15890500 BAD TRAP panic on a T4-1B due to NULL pointer dereference
15892217 panic assertion failed: !IN_VA_HOLE(va), file: ../../i86pc/vm/htable.c
15898382 ldm add-io VF to primary failed with ERROR: hardware or driver specific failure
15903493 prtdiag displays "SUNW,pf-emulation" entries in guest LDOM following VF add
15954351 systems hang in cv_wait() from prbarrier()
15994235 "hotplug poweroff" failed
16099913 ndi_devi_config_one can leak a node reference if NDI_CONFIG is used
16293098 qlc card has issues being added to an I/O domain
16311421 extra message seen with ldm add-io and rm-io VF
16420072 create slot-names properties for PCIe hotpluggable slots
16426940 M5/T5: ldm unbind of an SDIO Domain hangs
16493596 devices added to guest domain with add-io are not immediately visible
16503776 network devices under RCs loaned out to I/O domains have incorrect maxvfs limit
16562469 Oracle with a large SGA on T5/M5 startup&shutdown takes quite a long time
16608837 "ABORT: Unimplemented Fcode" messsage seen when adding pc_1 to a domain on T5-1B
16741159 ldmad core dumps found on ldoms split domain w/IB-QDR
16760866 backport softlso and packet hashing functions to Solaris 10
 
(from 150400-01)
 
15704883 capture ftrace_state during panic
15707492 enabling FTRACE on 32-bit x86 causes panic from within intr_thread()
15730133 pm_start() on an attaching node vulnerable to pm_powerup from framework calls
15741367 ftrace.h definition of caller needs to be local to FTRACE_x macros
15750574 anon_array_lock - high contention
15760160 RFE: add dcmd to mdb to extract ftrace information from crash dumps
15760769 FTRACE probe in i86pc trap() doesn't include addr
15794086 kstat assertion panics after LDOM migration
15817071 system panic in pollhead_delete when using devpoll on /proc file system
15823255 LDoms CPU DR of available CPUs causes lpl_topo_verify fail -5 panic
16039539 response of SYN/ACK is delayed when connected from 600 clients per second
16361843 panic[cpu74]/thread=1000a22317a0: Deadlock: cycle in blocking chain
16362919 assertion failed: cu_cpu_info, file: ../../common/os/cap_util.c, line: 1771
16425981 panic with freeb() in sctp_process_recvq()
16771544 integration of IB-SRIOV into Solaris 10 requires rejuvenating KU 148888-04, 148889-04
 
(from 148377-01)
 
        This revision accumulates generic Sustaining patch 148174-01
        into Solaris S10U11 update.
 
(from 148174-01)
 
7085913 system is panicking in pollhead_delete() even with the fix for 7008672 in place
 
(from 150108-01)
 
15781929 ino_ipil_cntr should be reinitialized after spurious MSI
15797792 T4-2 device interrupt to CPU binding is poor using half of CPUs and causing performance issues
15824758 ino_unclaimed_intrs should be reinitialized in px_msiq_intr()
15825402 7202764 breaks 7176502 resulting in panic
 
(from 150300-01)
 
15593115 e1000g driver_alias entries for unsupported device IDs
15780385 "ESB2 receiver disabled" warning message should be only printed in debug binary
15785440 network lost connection when doing NFS stress test with heavy load on HP Z620
15997791 VNICs don't work when you run out of H/W MAC addresses on 82579 chipset
 
(from 150532-01)
 
16908436 forthdebug is a deliverable for Bug 16164869
 
(from 149502-01)
 
6856693 sadb_update_sa() checks for duplicate SADB_UPDATE messages in the wrong place
 
(from 149616-01)
 
7178504 header ordering issue with tunnel encap limit in IPv6 tunnel mode ESP
7182033 fix for CR 7178504 fails for IPv4 over IPv6 tunnels
 
(from 150125-04)
 
17075297 deliver usr/lib/mdb/kvm/sparcv9/zfs.so separate from RTI 383705
 
(from 150125-03)
 
15662359 activate ZPOOL_CONFIG_FRU in ON
15715215 assertion failed: REF_INACTIVE(ref), file: ../../common/fs/zfs/arc.c, line: 2643
15715549 arc_p is inflated
15718032 73% vdb407_nvSeqReadBs128kFs1g_zfsraidz3_9 performance regression x86
15750994 performance pathology when manipulating files with large numbers of empty blocks (arc_trim_refs)
15752435 need steady state arc_size changes compatible with rearc
15764044 rw_exit: not owner, system panic during unlink from running consumer test util
15780200 assertion failure !BUF_L2CACHED(buf), function arc_destroy_buf
15785248 system hang due to physical memory exhaustion seen when major shift in workload
15789239 arc data shrinks continuously after arc grew to reach its steady state
15792971 4GB system with 4 zones hangs while trying to upgrade; kernel runs out of memory
15794791 'zinject -c all' panic, #pf Page fault Bad kernel fault at addr=0x30
15797481 small block workloads can foil steady-state ARC
15797710 ARC: bytes evicted should be tracked in a kstat
15798684 arc_read is trying to drop a lock it does not own, panics system
15799935 dism_stress tests are causing x86 and SPARC to hang
15801366 arc_no_grow forcing unnecessary shrink in arc_size
15801413 arc state eviction list size needs to account for "HOLE" buffers
15802661 l2arc write leaks checksum
15803852 arc_get_data_buf() waiting for shrink, but has lock that arc_reclaim_thread() needs to effect shrink
15804591 userland ::arc no longer works
15804687 frequent 32-bit ztest core dumps due to being out of memory
15810974 panic assertion failed: rc->rc_count >= number, file: ../../common/fs/zfs/refcount.c
15811520 arc_reclaim_thread missing a ptob call on the redzone
15813348 SMB performance test shows severe degradation
15817594 arc_anonymize_ref() invokes bcopy() on invalid structures
15819541 arc_evict_buf() returns wrong value for HOLE buffers
15844219 group with write_attributes ACL permission cannot update timestamp
15867625 arc_hold(buf, cb) not arc_rele'd on L2 error path
15987135 l2arc_evict() l2buf == buf->b_cookie assertion failure
15998183 arc_memory_throttle can stall TXG pipeline when system is paging
16776160 unable to bind the LDoms on zpool volumes after reboot
 
(from 150125-02)
 
15807421 load_nvlist() should handle read errors
15811225 zfs_ioc_pool_get_props can suspend holding spa_props_lock provoking a deadlock
15916736 zpool iostat -T option does not print the timestamp
 
(from 150125-01)
 
15494128 assertion failed: zvol_get_stats(os, nv) == 0, when EIO encountered
15710534 advance rotor before blocking on space_map_load
15771768 memleaks: arc_read/arc_hole_refi, vdev_alloc/metaslab_group_create
15793416 lock contention related to ZFS root lock
15809921 metadata rewrites should coordinate with l2arc
15822306 zfs panic at zfsvfs_teardown
15823992 IMPLY and EQUIV in libzpool need to honor aok
16395538 need to withdraw 149636-01 due to wrong deliverables
 
(from 150115-04)
 
15572296 do plumb/unplumb test with USB flash disk will cause system panic
 
(from 150115-03)
 
15602062 hang in sd while importing a ZFS pool
 
(from 150115-02)
 
15555874 sd setting incorrect buffer length on mode select
15696163 I/O is not retried to another path after one port is masked out
 
(from 150115-01)
 
15588762 format -e to does not update device softstate
 
(from 150629-02)
 
17336846 under error conditions mpt may call mpt_init_chip() recursively and panic
 
(from 150629-01)
 
15775014 mpt_check_scsi_io_error should sync the DMA handle before using it
 
(from 147697-01)
 
6925293 ipsecconf(1M) needs a receive-buffer increase for SPD_DUMP
7040324 ipsecconf -l fails with dump read: message corruption, 39184 len exceeds 8192 boundary
7042201 ipsecconf core dumps around send_pf_pol_message on certain rules
 
(from 148338-07)
 
15707174 s8_boot script has not been prepared to perform a p2v of an s8 32-bit system
 
(from 148338-06)
 
16845490 patch 148338-06 needs to add hard requirement for KU 148888-04
 
(from 148338-05)
 
16249123 /usr/bin/ps with "-L" hangs in Solaris 9 branded zone, when zombie LWP exists
16709548 backport 15788521 to Solaris 10 (s9_brand delivery)
 
(from 148338-04)
 
16002784 Kstat Perl module failed assertion caused by vminfo_t size mismatch in s9 zone
 
(from 148338-03)
 
15825188 Solaris 9 branded zone spews in.mpathd related errors to messages
 
(from 148338-02)
 
        This revision accumulates generic Sustaining patch 148161-02
        into Solaris S10U11 update.
 
(from 148338-01)
 
        This revision accumulates generic Sustaining patch 148161-01
        into Solaris S10U11 update.
 
(from 148161-02)
 
6597202 Solaris Auditing does not work in S8 containers
6958780 Solaris Auditing does not work in S9 containers
7108366 flar/cpio archive files created with absolute paths can cause GZ corruption whileBranded NGZ install
 
(from 148161-01)
 
7055689 emulation function for aio (s9_kaio()) fails to handle AIOWAITN
 
(from 139944-01)
 
        This revision accumulates generic Sustaining patch 139510-01
        into Solaris S10U7 update.
 
(from 139510-01)
 
6723187 /devices directory should be removed upon S8/S9 zone creation
6725716 sys-unconfig hangs when installing Solaris 8 Containers (Etude) from a Flash Archive
 
(from 148557-04)
 
17065864 panic in n2cp_move_req_off_page due to NULL reqp
17454726 panic in n2cp_move_req_off_page due to NULL nr_context
 
(from 148557-03)
 
17071025 panic in n2cp_clean_blockctx
17371338 double free in n2cp_ulcwq_freebuf and null dereference in n2cp_clean_hmacctx
 
(from 148557-02)
 
15521496 cryptoadm enable/disable all mechanisms inconsistently handles random
15686082 fix to 6769595 inadvertently changes behavior of cryptoadm disable provider=<prov> all
15765104 digest fails on files between 2MB and 1GB, gives wrong sha1 sum on files > 1GB on SPARC T3 machines
15766446 n2cp should not practice recursion
16340257 'pkg update' may fail or get hung when n2cp digest operation is enabled
 
(from 148557-01)
 
        This revision accumulates generic Sustaining patch 148231-01
        into Solaris S10U11 update.
 
(from 148231-01)
 
6273805 C_Sign call with NULL_PTR for data buffer returns CKR_FUNCTION_FAILED
 
(from 150627-01)
 
15781192 double free in kernelSlottable.c kernel_slottable_init()
16894659 fix typos in PKCS11 source file comments
 
(from 149729-03)
 
16397136 ikecert certlocal usage for PKCS#8 export is wrong
16886354 IKEv1 should understand certificates signed with SHA-2
17165615 IKEv1 pkcs#11 on-token key and certificate generation broken
 
(from 149729-02)
 
15618846 in.iked does not explicitly flag missing certs
15685991 memory leaks in in.iked
15783297 need a way to turn off OCSP checking in IKEv1
15814090 uninitialized variable in libike certlib_load function
16218907 minor error in fix for 15618846 -- error message loop should use 'j'
 
(from 149729-01)
 
        This revision accumulates generic Sustaining patch 149718-01
        into Solaris S10U11 update.
 
(from 149718-01)
 
6885375 check_rule() checks p2_lifetime_kb against p2_softlife_kb in wrong units
6885809 p2_lifetime_kb value conversions lead to integer overflows for p2_lifetime_secs > 4096
6885833 IPsec utilities should print lifetimes in human-readable format
6887320 check_rule() fails to enforce lower limit for SOFT lifetime in case of small HARD lifetime values
6889082 SOFT kbyte lifetime can be insane
6889086 ikeadm reports kilobyte lifetimes with wrong units
6889474 check_rule() should convert to different philosophy
6890960 ikeadm reports wrong value of system phase 2 soft byte lifetime
6892388 some ACQUIRE-specified lifetimes are still ignored by in.iked
6892846 RESPONDER-LIFETIME notify message handling should scrutinize kilobyte lifetimes as well
6893246 RESPONDER-LIFETIME handler should respond to invalid notification messages
6894689 iked should send kilobyte lifetimes in RESPONDER-LIFETIME messages
6895232 ikeadm get defaults does not work correctly if only p2_lifetime_kb is set
6897711 iked debug output should be less confusing for average sysadmin
6898492 iked should enforce lower maximum values for lifetimes
6900609 p1_lifetime_secs parsing should use the concept of least surprise too
6905264 iked debug claims Phase 2 notification failed as a result of delete
6908382 reflect the IPsec aging change in SOFT expire handling
 
(from 150841-01)
 
17031640 T4 server panics in page_vpsub

--------------------------------
 
Please refer to the man pages for instructions on using 'patchadd'
and 'patchrm' commands provided with Solaris.
 
The following example installs a patch to a standalone machine:
 
       example# patchadd /var/spool/patch/123456-07
 
The following example removes a patch from a standalone system:
 
       example# patchrm 123456-07
 
For additional examples please see the appropriate man pages. Any
other special or non-generic installation instructions should be
described below as special instructions.

-----------------------------
 
NOTE 1:  After installing this patch, all zones have to be rebooted.
 
NOTE 2:  With the integration of bug 15689118 (pgstat(1M) should report
         absolute maximum capacity of a PG), pgstat(1M) will report
         fixed capacity reporting for hardware utilization by default
         but leave the ability to use the learned capacities by
         specifying the -l option.
 
         -l
             Use the learned capacity in hardware utilization calcu-
             lations.  The learned  capacity is the maximum observed
             hardware utilization since boot.
 
         For OUTPUT HEADINGS, UTIL and CAP output printed is updated.
         The updated meaning of these headings is as below:
 
         UTIL
 
             Hardware utilization of PG's shared hardware component
             normalized per second.  This can be a large number, so K,
             M, B, and T are used for denoting  thousand, million,
             billion, and trillion, respectively.
 
         CAP
 
             Approximate maximum possible utilization for PG's shared
             hardware component per second.  This value is fixed by
             default at a best guess for the given platform, but can
             be changed to a learned value using the -l option. This
             can be a large number, so K, M, B, and T are used for
             denoting thousand, million, billion, and trillion,
             respectively.
 
NOTE 3:  On removal of this patch, the ps binary is not restored automatically.
         It needs to be manually restored using the following command at
         zone prompt:
 
         # cp /usr/bin/ps.pre_p2v /usr/bin/ps
 
NOTE 4:  This patch requires a reboot to restart the in.iked(1M) daemon.
         Until the daemon is restarted it will continue to run the older
         version of the software.
 
         The in.iked daemon may be restarted manually when convenient by
         assuming a role with the necessary privileges and performing the
         following command:
 
         # svcadm restart svc:/network/ipsec/ike
 
NOTE 5:  If ignore_ocsp keyword is present and uncommented in the
         ike.config(4) file,  in.iked(1M) ignores OCSP information in
         certificate X.509v3 extensions and does not query Certificate
         Authorities via OCSP.
 
NOTE 6:  With the integration of bug 16886354 (IKEv1 should understand
         certificates signed with SHA-2), support has been added for
         SHA-2 signed certificates to ikecert(1M). Now '-t' option of
         ikecert(1M) supports following key types in addition to existing
         key types:
 
             rsa-sha256, rsa-sha384, rsa-sha512 and dsa-sha256.

NOTE 7:  Systems with this patch applied may experience a rare race
         condition resulting in a bad trap panic in sysdc_tick().  To
         work around this issue, disable SYSdc (SDC) scheduling class.
         This can be done by adding the following entry in '/etc/system'
         and then rebooting the system to activate the change:

         exclude: sched/SDC

         Please see Service Alert Document 1965270.1:

         https://support.oracle.com/rs?type=doc&id=1965270.1

         for full details.

NOTE 8:  ZFSSA may experience a system panic with:

         NULL pointer dereference in l2arc_read()

         Please see Service Alert Document 1918169.1:

         https://support.oracle.com/rs?type=doc&id=1918169.1

         for more details.  This issue is resolved in 150400-10/150401-10
         or later.

NOTE 9:  Systems Utilizing Dynamic Intimate Shared Memory (DISM) may
         experience a BAD TRAP system panic in spt_cache_pps_task().
         For more details, see Service Alert Document 1676474.1:

         https://support.oracle.com/rs?type=doc&id=1676474.1

NOTE 10: To avoid an issue where processes may become unkillable, please
         apply revision -10 (or greater) of this patch.  For more details,
         see Service Alert Document 1641745.1:

         https://support.oracle.com/rs?type=doc&id=1641745.1