Disclaimer: 

Please note: Although OBSOLETED patches are available on SunSolve, Sun recommends using the most recent patches and the most recent revision of those patches. OBSOLETED patches do not include the latest bug fixes and/or product enhancements, and may require the installation of additional patches as a corrective measure.

Status: OBSOLETE

Patch Id: 105181-19

Summary: SunOS 5.6: Kernel update patch

Date:  Feb/04/00

Installation Requirements:

Solaris Release: 2.6

Sun OS Release: 5.6

Unbundled Product: 

Unbundled Release: 

Xref: This patch available for x86 as patch 105182

Topic: 

SunOS 5.6: kernel update patch

	NOTE:	Refer to Special Install Instructions section for 
                IMPORTANT specific information on this patch.

Relevant Architecture: sparc

BugId's fixed with this patch:

1237009 1266113 1267506 4009567 4011225 4024106 4034675 4043763 4044980 4049222 4056222 4061229 4062572 4063932 4065248 4065762 4066897 4067568 4067641 4070714 4070968 4073684 4076062 4076118 4077343 4079028 4080160 4080714 4086957 4087112 4089451 4089777 4089974 4090929 4092838 4093708 4094313 4095093 4095650 4097064 4097082 4097233 4098645 4098732 4099437 4099656 4099664 4100666 4101647 4102334 4102546 4104625 4105822 4106093 4107724 4108806 4110503 4110763 4114176 4115010 4115793 4115951 4117624 4117875 4117898 4118425 4119193 4119498 4119745 4120985 4122292 4122408 4122617

Changes incorporated in this version:

4110503 4136829 4140683 4164648 4177905 4243937 4248024 4270284 4278536 4293406 4188064 4266736 4274565 4283719

Patches accumulated and obsoleted by this patch:

105214-01 105636-01 105776-01 106031-02 106308-01 106048-07

Patches which conflict with this patch: 

Required Patches:

Obsoleted by:

Files Included in this Patch: 

/kernel/sys/shmsys
/kernel/fs/procfs
/kernel/fs/sockfs
/kernel/genunix
/kernel/misc/consconfig
/kernel/misc/klmmod
/kernel/misc/krtld
/kernel/sys/doorfs
/kernel/sys/kaio
/kernel/exec/elfexec
/platform/SUNW,Ultra-4/kernel/misc/platmod
/platform/SUNW,Ultra-Enterprise/kernel/misc/platmod
/platform/sun4c/kadb
/platform/sun4c/kernel/unix
/platform/sun4d/kadb
/platform/sun4d/kernel/unix
/platform/sun4m/kadb
/platform/sun4m/kernel/unix
/platform/sun4u/kadb
/platform/sun4u/kernel/cpu/SUNW,UltraSPARC
/platform/sun4u/kernel/genunix
/platform/sun4u/kernel/misc/consconfig
/platform/sun4u/kernel/misc/platmod
/platform/sun4u/kernel/unix
/platform/sun4m/kernel/cpu/Ross,RT625
/platform/sun4m/kernel/cpu/TI,TMS390Z55
/usr/platform/sun4u/lib/libc_psr.so.1
/usr/bin/adb
/usr/include/sys/kobj.h
/usr/include/sys/proc.h
/usr/include/sys/proc/prdata.h
/usr/include/sys/rce.h
/usr/include/sys/sunddi.h
/usr/include/sys/vnode.h
/usr/kernel/drv/ksyms
/usr/lib/adb/thread
/usr/lib/adb/shmid
/usr/platform/sun4u/include/sys/iommu.h
/usr/platform/sun4u/include/sys/machsystm.h
/usr/platform/sun4u/include/sys/prom_plat.h

/platform/SUNW,Ultra-Enterprise-10000/kernel/misc/platmod
/platform/SUNW,Ultra-Enterprise-10000/kernel/unix
/platform/sun4u1/kadb
/platform/sun4u1/kernel/cpu/SUNW,UltraSPARC
/platform/sun4u1/kernel/drv/dr
/platform/sun4u1/kernel/genunix
/platform/sun4u1/kernel/misc/consconfig
/platform/sun4u1/kernel/misc/hswp
/platform/sun4u1/kernel/misc/platmod
/usr/platform/sun4u1/lib/libc_psr.so.1

Problem Description: 

4110503 as_setprot heuristic gave my process a wedgie
4136829 overflow in p_lckcnt/p_cowcnt leads to panic
4140683 Cleanup of p_lckcnt and p_cowcnt stuff
4164648 Software clock has no prevention for possible broken tod
4270284 tod_version is not needed
4177905 sun4d cannot boot with Solaris 2.6, FDDI/S 4.0, FC-AL SBus X 5 and VxVM2.5
4243937 kpreempt() doesn't finish
4248024 Starfire panic's in lwp_sigredirect
4278536 Solaris 2.5.1 HW 3 and Solaris 2.6 HW 3 panics on boot net with 256M
4293406 System hang/panic in pollwakeup_safe

4188064 getting a hang running the mem31 test from mixstress on 2.6 -10 KU patch
4266736 recursive mutex in kmem_cache_kstat_update when cpu_offline.
4274565 System could hard hang in dr_relocate_ismpage
4283719 large kernel allocations can end up waiting forever for their pages

(from 105181-18)

4228409 rmalloc(9F) sloth and fragmentation kill Veritas performance
4229608 SS20 running 5.7 hangs, deadlock between page_lock and entersq
4261612 profil not disabled on exec*() as indicated in man page

4251398 DR memory drain stalls.  Logic to relocate ISM pages needs to be improved.
4265862 SyMON causes hang (panic1) of domain after a DR Detach
4281208 page_get_mnode_freelist() bug can cause hang or panic

(from 105181-17)

4061229 NFS server crashes in flk_delete_active_lock
4125102 ufs_itrunc()/top_end_async() deadlock
4202680 thread_create() can return at different interrupt level
4205026 panic in sfmmu_pageunload - large page, 4 Meg page sizes
4208677 5.5.1 panics when getpgid() is called with PID of a process just being
4211474 Open call hangs on nfs mount point with stale file handle
4220295 System panics and then reboots even with OBP settings=false
4222164 KAIO wait broken under 2.6 returns ERR #22 EINVAL
4229582 lwpchan cache corruption
4232237 bug in the POSIX Message Queue
4242754 rsh hang caused by KJP 105181-13
4246455 ddi_umem_alloc() panic in s27
4246501 E4000 panic with xc_attention after installing 103640-27

4248738 panic with freemem dropping to zero
4254715 VerExec error occurs when 2.6 domain is rebooted from OBP
4255955 With enable_grp_ism=1 on E10000, 5.6 -15 KJP, oracle 7.3.4 crashes
4265527 The abort of Memor Drain resulted in panic of domain
4267105 Kernel cage shrink problem.  domain hangs when DR enabled.

(from 105181-16)

4220949 dis_err_panic1() does not save the AFSR value during panic.
4205009 system rebooted sfmmu_tsb_miss trap 31
4200217 program hangs with resource temporarily unavailable using lio_listio library
4195726 page_list_sub: large page on cachelist panic
4087112 panic in putq due to null q_last value
4049222 excessive xcalls when processes with very large shared memory segment exit

4221587 system hang: race condition between fuser and clock thread
4228891 Possible hang in dr_relocate_ismpage
4247010 Enabling the DR cage can sometimes cause unnecessary context switching.
4247028 Potential deadlock in DR code that causes domain to hang.

(from 105181-15)

4214611 hot kernel lock limiting scalability on Solaris 2.6+ or greater
4164856 pmap -x reports bogus values in the private column
4115010 dump times out on large memory systems
4206401 panic when linking active stream below mux
4128397 kernel crash dump after system panics causes another panic in bio.c::getblk()
4099437 segvn_pagelock must not softlock COW pages;
4226963 A Memory Leak of pthread program is due to schedctl doesn't free user address
4184291 kernel crash dumping problems when kvm usage > 2GB
4120985 panic in free_page on multiprocessor machines with fix bug 4026411
4086957 System crashes accessing 512k or 4M page on sun4u
4094313 sfmmu (sun4u) hat_unload(HAT_UNLOAD_UNMAP) does not work as expected

4218687 page_create_wait can be called before pageout exists in large configs
4227197 2.6 OS with 105181-13 KU patch took hostreset dump running drstress
4240380 DR memory drain stalls -- ism_page_relocation logic unsuccessful
4244523 Data corruption in ISM shared memory segs with heavy load/multi-threaded apps.

(from 105181-14)

4192227 recursive rw_enter on as->a_lock acquired initially in procfs
4175350 longjmp see NULL value with jmp_buf causes csh dump core on SS20 hyperSPARC MP
4178298 race condition in pr_getprot()
4151266 400 mhz cpu with 8 meg cache fails to boot - panic bad mutex

4194506 E10k hangs in biowait
4196459 page locks causing DR drain and detach ioctl failures
4226189 domain panic with dr_mem: dr_relocate_page: !sfmmu_ism

(from 105181-13)

4198942 solaris ntp_adjtime broken: clock.c; useless for PPS sync of the system clock
4191544 System panic "blocked on mutex" in streams
4190117 networked application cause a hard hang
4184430 panic: recursive mutex_enter on kmem_cache_lock
4176940 Panic() when parsing driver.conf greater than 8KB
4169509 memory delete encounters a page in impossible state
4132290 can't dump core when deadman drops into debugger
4110763 panic in ts_update_list when accessing t_schedctl

4131387 the test "lwp03" from mixstress test suite panics on domain xf3 with 2.6hw3
4213437 System Panic'd when page_sub hit upon a null page

(from 105181-12)

4104625 UE4000's 2.5.1 panic: xc_attention() timeout, recv_cpuset 0x0, xc_cpuset 0x1
4066897 Calling listen() more than once on socket causes EPROTO error
4179269 giant core images cause filesystem corruption
4102546 assertion failed: pp != NULL, file: ../../common/vm/vm_page.c, line: 566

4134487 system hung - cage grow blocked by shared locked pages
4181969 domain panic with dr_mem: dr_relocate_page: !sfmmu_ism
4187680 System panic with assertion failed: MUTEX_HELD(&ctx_lock)
4192891 system panics with dr_relocate_page: could not page_hashin().

(from 105181-11)

4179342 solaris needs to provide a resource management solution for big servers
4056222 sema_p_sig is broken

(from 105181-10)

4170500 solaris ntp_adjtime broken, useless for PPS sync of the system clock
4151480 under Solaris 2.6, adb reports wrong information for o registers for v9 arch
4147079 stubs mechanism for modules is faulty 
4139770 fcntl() returns EINVAL error in BCP mode when NFS file is read
4131439 deadlock_panic from pi_willto
4118425 sfmmu_tsb_miss() may get a recursive mutex panic
4117624 if lockd is restarted, clients receiving signals have problems with locks
4108806 rename of automounted directory results in panic

4169916 Excessive ECC errors
4174959 System hard hangs w/ oracle causing sigbcmd or hostint not to break to 'ok'

(from 105181-09)

4162055 invalid socket return error code: ECONNRESET should be ECONNREFUSED
4151212 system crashes in page table steal
4148073 successful fork() sometimes does not return zero in mt/multi lwp child proc
4141788 system hangs due to pagefault loop in shared memory
4122617 device driver providing devmap not unloadable because of leaked hold count
4122292 multithreaded httpd process deadlock during cfork()
4119745 realitexpire() algorithm is too slow when system time is changed
4107724 implement workarounds for spitfire errata 32 and 54
4065248 UFS caching can adversely affect application performance

(from 105181-08)

4144929 kernel patch 105181-05 causes dbx on setuid programs to get EBUSY
4127499 SunFire should not be as verbose in printing CE ECC messages
4098732 recursive mutex enter in kstrgetmsg()
4043763 in MT applications, close() blocks if fd is in use by another thread

(from 105181-07)

4132927 open system call does not audit if creat bit set.
4125580 system panics in cstat64 with type prvnodeops vnode
4122408 Backup performance with Netbackup 3.0 is far below expectations.
4119498 HSI/P - Performance problem upto 200 m/s interframe delay.
4115951 Diskless Ultra-1s unable to perform system crash dump across network

(from 105181-06)

4080160 tickint_clnt_add miscalculates interval between handler calls
4089777 processes can hang or crash while forking with ISM on sun4u
4098645 setcontext() uses >25% of the stack & segkp_fault: accessing redzone panic.
4102334 sunfire PDB node panics with xc_one() timeout,  no core
4119193 ASSERT() panic due to race condition in /proc-supported watchpoints

4134357 availrmem not being reduced during Starfire memory detach
4136544 getting "flusher thread" hang during dr drain
4137584 CE reporting incorrect P numbers

(from 105181-05)

1237009 users umask modifies ACL's of new files under default ACL's
1266113 due to memory corruption in the OS, Xsun crashes randomly on IPX
4070714 system hangs/resets running OLTP benchmark on Veritas Raid-5 volumes on plutos
4073684 "mkdir -p dir" and "mkdir dir" work differently in the presence of default ACLs
4076118 pf_set_dmacapable() needs to support ffb detach/attach
4080714 panic if pci device drivers fails to detach.
4093708 process/streams modules hung on rwlocks
4095093 IOMMU tsb is preallocated.  We need to preallocation some for DR.
4095650 sunddi.c needs more framework for nexus hotplug
4097064 hotplug daemon asynchronous attach collision with detach requests.
4097082 _lwp_sigredirect() if called from a non-mt process panics kernel.
4097233 TSB for segspt provides up to 15% boost in OLTP performance.
4099656 httpd process hangs and can't be killed
4099664 hp_op_make_new_children uses wrong name for child alloc
4101647 loop-back mounts of /proc (for chroot(1M)) are insecure and broken
4106093 system hangs during high I/O loads (bread/bflush deadlock)
4114176 system panic due to one lwp closing another lwp's /proc file descriptor
4115793 /proc security too restrictive; a process should be able to open itself
4117875 adb threadlist causes panic:  peek_fault: bad nofault data
4117898 panic in procfs due to unprotected reference count

4093052 Starfire 2.6 needs shuffle lest memory hot-spotting occurs
4098576 cannot see the ssa after dr attaching the board with the soc
4099466 "DR: driver (vge) not known to support DDI_DETACH" while trying to detach
4109198 pcf_acquire_all argument error in page_reclaim (Starfire only)
4110996 oninit informix processes hang
4115526 assertion fail removing interrupt threads from detaching proc
4116141 DR panics machine with BAD TRAP
4116349 panic during cpu detach- interrupt thread detected after cpu_offline
4116351 While draining board: Error draining resources: ioctl failed.
4116396 domain hang during drstress - locked thread on deathrow
4117010 Deadlock during Starfire DR memory attach
4117862 send_mondo panics during drstress on Starfire
4117925 hotplug_daemon and Starfire DR driver deadlock when walking devtree
4119285 re-activate hotplug_daemon with suitable synchronization for DR
4119751 deadlock between DR and pageout daemon
4128193 maxusers calculation still based on dr-max-mem in s297hw3
4129902 Oracle panics on startup with s297hw3 & KU-05 on sun4u1 platform

(from 105181-04)

4067641 changing acl's on a UFS fs mounted readonly causes machine to panic
4089451 doors/scheduler activation broken in the execed process.

(from 105181-03)

4070968 a synch object that spans pages can cause deadlock

(from 105181-02)

4009567 permanent cache can grow very large leaving the system hungry for kmem
4062572 syncing page gives up during shutdown
4070968 synch object that spans pages can cause deadlock
4076062 sigtimedwait() returns immediately
4077343 sun4u systems incorrectly report "sync - giving up" while halting
4089974 new ddi utility routines for nexus drivers 

(from 105181-01)

4044980 software trap #6 (ST_FIX_ALIGN) does not work in a threaded application

(from 105636-01)

4079028 ECC and other errors not handled correctly on Ultra-5/Ultra-10 due to 
	Sabre spec change

(from 105776-01)

4090929 system panics due to infinite recursion in prrealvp

(from 105214-01)

4065762 recursive mutex_enter panic in sockfs
4067568 assertion failure in sock_poll(): *reventsp != 0    *phpp != NULL

(from 106031-02)

4092838 kadb cannot reliably set breakpoints in loadable modules
4011225 kadb does not always set breakpoints as expected
4034675 kadb can't store to kernel memory on sun4u
1267506 kadb breakpoints don't work
4024106 kadb :c often causes panics
4100666 kadb data addresses not properly locked in kernel mmu

(from 106031-01)

4105822 In Solaris 2.6, adb reports value 0 for unmapped addresses.

(from 106308-01)

4063932 orphan lock problem caused by sigalrm/sigintr & large packet loss

(from 106048-07)

4251398 DR memory drain stalls.  Logic to relocate ISM pages needs to be improved.
4265862 SyMON causes hang (panic1) of domain after a DR Detach

(from 106048-06)

4227197 2.6 OS with 105181-13 KU patch took hostreset dump running drstress

(from 106048-05)

4151266 400 mhz cpu with 8 meg cache fails to boot - panic bad mutex
4207064 System board detach failed with mistaken low free memory condition.
4225351 dr driver's page_reloc_map needs to be initialized

(from 106048-04)

4028250 DR memory detaches can cause system to hang waiting waiting for memory
4129429 system resources exhausted during board detach causing panic
4133607 dr operations with mem_board_interleave_ok problems
4139521 Resource panic during complete_attach on domain xf3 running drstress
4144996 dr is reporting mem-unit already attached during brd attaches
4169694 Unable to get memory drain info - stop drain
4174097 patch to 2.6 dynamic reconfiguration needed for gem card driver

(from 106048-03)

4129729 panic with solaris 2.6  s297hw3_07 and pci/vge during dr attach

(from 106048-02)

4129031 Use of PC_IDLE in DR is flawed

(from 106048-01)

4032766 A deadlock occurred in dr_drv_detach during a DR detach.
4052382 Detach of a board with memory interleaving results into an Arbstop.
4099466 "DR: driver (vge) not known to support DDI_DETACH" while trying to detach
4116141 DR panics machine with BAD TRAP
4116351 While draining board: Error draining resources: ioctl failed
4117010 Deadlock during Starfire DR memory attach
4119285 re-activate hotplug_daemon with suitable synchronization for DR
4119751 deadlock between DR and pageout daemon
4119852 DR must update post2obp structure after copy/rename for reboot to work
4121978 arbstop after failure to quiesce
4129031 DR should use PC PAUSE/FLUSH instead of PC IDLE

Revision History: 

105181-10 105181-08 105181-13 105214-01 105181-14 105181-03 105181-17 105636-01 105181-06 105181-15 105181-04 105181-11 105181-07 105776-01 106031-02 105181-09 105181-05 105181-16 105181-12

Patch Installation Instructions: 

-------------------------------- 
Refer to the Install.info file within the patch for instructions on
using the generic 'installpatch' and 'backoutpatch' scripts provided
with each patch.  Any other special or non-generic installation
instructions should be described below.

Special Install Instructions: 

----------------------------- 
Reboot the system after patch installation.

	NOTE 1:  To get the complete fix for bug 4090929 (system panics due to
       	         infinite recursion in prrealvp) we recommend installation of
                 rev 01 (or newer revisions) of the following patches:

                	105778-01  /kernel/fs/specfs
                	105780-01  /kernel/fs/fifofs

	NOTE 2:  To get the complete fix for 1237009 (users umask modifies
                 ACL's of new files under default ACL's) and 4073684 (mkdir
                 -p dir" and "mkdir dir" work differently in the presence
                 of default ACLs), we recommend installation of rev 01
                 (or newer revisions) of the /usr/bin/mkdir patch:

			106141-01  /usr/bin/mkdir

	         As a result of this fix for 1237009, the umask/ACL interaction  
		 has changed.  

                 Please refer to the following paragraph for a revision to the
		 manpage for the command 'setfacl' which defines the new umask/ACL
                 interaction.

                 ************************************************************

                 A directory may contain default ACL entries.  If a  file
                 is created in a directory that contains default ACL entries,
                 the newly created file will have permissions generated
                 according to the intersection of the default ACL entries,
                 and the permissions requested at creation time.

                 The umask(1) will not be applied if the directory contains
                 default ACL entries. If a default ACL is specified for a
                 specific user (or users), the file will have a regular ACL
                 created;  otherwise, only the mode bits will be initialized
                 according to the intersection described above.  The default
                 ACL should be thought of as the maximum discretionary access
                 permissions that may be granted.

                *************************************************************

	NOTE 3: For users running Java or other multithreaded applications.
                We highly recommend installing the linker patch,   
                105490-04 (or newer).

        NOTE 4: To get the complete fix for bug 4232237 (bug in the POSIX
                Message Queue) we recommend installation of 106797-03
                (or newer revisions). 

        NOTE 5: For the users of sun4u/sunfire platform, To get the complete
		fix for bug 4164648 (Software clock has no prevention for 
		possible broken tod) we recommend installation of 105651-11
		(or newer revisions). 

	NOTE 6: After installing this patch on a Sun Enterprise 10000 using
		Dynamic Reconfiguration, you should also install the dr
		daemon patch 106284-02 or newer.

	NOTE 7: Dynamic Reconfiguration (DR) on E10000. These problem should
    		be fixed in 105181-20 or newer, when available.

		bug 4297641:
		A system contains a board which has a device on it for which 
		the domain does not have the corresponding driver, that board
		will not be able to be detached from the system.  The following
		messages will appear on the console when this happens:

		WARNING: DR: driver (<null string>) not known to support 
		DDI_DETACH Error detaching board: ioctl failed.
		dr_daemon[292]: ...I/O error

		At this point the user would have to abort the detach and keep 
		that board in their domain.  The only way to remove that board
		from the domain would be to bring down the OS, and remove the 
		board by using the domain_remove and domain_create commands on 
		the SSP. Other boards can still be attached and detached from 
		their system this has no other effect on the system.

		bug 4304993:
		Due to a bug introduced in KU-19, if a user plans on using DR
		to attach 2 or more new boards to a domain, it is possible
		that the kernel cage could prematurely grow onto one of these
		boards. This could prevent a subsequent detach of a board from
		the domain.  If these conditions apply to the user, they should
		look at the workaround for bug 4304993 for further details.
		If a user DRs a board out, and then DRs the same board back in,
		this bug will not occur.