|
Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email hr@neohapsis.com |
From: IT Resource Center (support_feedback_at_us-support2-mail.external.hp.com)
Date: Sun Feb 23 2003 - 15:20:18 CST
HP Support Information Digests
===============================================================================
o IT Resource Center World Wide Web Service
---------------------------------------------------
If you subscribed through the IT Resource Center and would
like to be REMOVED from this mailing list, access the
IT Resource Center on the World Wide Web at:
Login using your IT Resource Center User ID and Password.
Then select Support Information Digests (located under
Maintenance and Support). You may then unsubscribe from the
appropriate digest.
To download a patch referenced below, access the
IT Resource Center on the World Wide Web at:
Login using your IT Resource Center User ID and Password.
Then select Individual Patches (under Maintenance and Support)
to access the patch. You may also download a patch via anonymous
ftp(1) from ftp.itrc.hp.com.
===============================================================================
Digest Name: weekly HP-UX series 700 11.X patch digest
Created: Sun Feb 23 6:05:13 EST 2003
Table of Contents:
Document ID Title
--------------- -----------
PHSS_28304 s700_800 11.22 FORTRAN I/O Library [libIO77]
PHSS_28303 s700_800 11.11 LIBCL patch
PHSS_28302 s700_800 11.00 LIBCL patch
PHKL_28602 s700_800 11.00 VxFS fsadm,MMF hang,mmap() perf improvement
PHCO_28413 s700_800 11.00 pax(1) cumulative patch
PHSS_28339 s700_800 11.X OV OB4.10 patch - SSEA packet
PHKL_28488 s700_800 11.00 copyin EFAULT, LDCD access type
The documents are listed below.
-------------------------------------------------------------------------------
Document ID: PHSS_28304
Date Loaded: 20030219
Title: s700_800 11.22 FORTRAN I/O Library [libIO77]
Patch Name: PHSS_28304
Patch Description: s700_800 11.22 FORTRAN I/O Library [libIO77]
Creation Date: 03/01/09
Post Date: 03/02/19
Hardware Platforms - OS Releases:
s700: 11.22
s800: 11.22
Products: N/A
Filesets:
COMPLIBS.LIBIO77-IA32,fr=B.11.22,fa=HP-UX_B.11.22_IA,v=HP
COMPLIBS.LIBIO77-IS32,fr=B.11.22,fa=HP-UX_B.11.22_IA,v=HP
COMPLIBS.LIBIO77-IA64,fr=B.11.22,fa=HP-UX_B.11.22_IA,v=HP
COMPLIBS.LIBIO77-IS64,fr=B.11.22,fa=HP-UX_B.11.22_IA,v=HP
Automatic Reboot?: No
Status: General Release
Critical: No
Category Tags:
defect_repair enhancement general_release
Path Name: /hp-ux_patches/s700_800/11.X/PHSS_28304
Symptoms:
PHSS_28304:
JAGae51914: Support for Dynamic Thread Local Storage to
enable shl_load to load this library.
JAGae45769: Abnormal output for 0.0 using P scaling edit
descriptor
JAGae50607: +FPD and underflow trap handling aborting
instead of setting the result to zero and
continuing as is done on PA-Risc machines
PHSS_27471:
JAGae31448: Support for I/O greater than 2 GB files.
Defect Description:
PHSS_28304:
JAGae51914: Support for Dynamic Thread Local Storage to
enable shl_load to load this library.
JAGae45769: Abnormal output for 0.0 using P scaling edit
descriptor
JAGae50607: +FPD and underflow trap handling aborting
instead of setting the result to zero and
continuing as is done on PA-Risc machines
PHSS_27471:
JAGae31448: Support for I/O greater than 2 GB files.
Enhancement:
No (superseded patches contained enhancements)
PHSS_27471:
JAGae31448: Support for I/O greater than 2 GB files.
SR:
8606267205 8606287981 8606281826 8606286668
Patch Files:
COMPLIBS.LIBIO77-IA32,fr=B.11.22,fa=HP-UX_B.11.22_IA,v=HP:
/usr/lib/hpux32/libIO77.a
COMPLIBS.LIBIO77-IS32,fr=B.11.22,fa=HP-UX_B.11.22_IA,v=HP:
/usr/lib/hpux32/libIO77.so.1
COMPLIBS.LIBIO77-IA64,fr=B.11.22,fa=HP-UX_B.11.22_IA,v=HP:
/usr/lib/hpux64/libIO77.a
COMPLIBS.LIBIO77-IS64,fr=B.11.22,fa=HP-UX_B.11.22_IA,v=HP:
/usr/lib/hpux64/libIO77.so.1
what(1) Output:
COMPLIBS.LIBIO77-IA32,fr=B.11.22,fa=HP-UX_B.11.22_IA,v=HP:
/usr/lib/hpux32/libIO77.a:
HP-UX libIO77 ILP32_NO B.11.23.%58 (Itanium-based ho
st and target) Wed Jan 8 13:33:00 2003 CST
COMPLIBS.LIBIO77-IS32,fr=B.11.22,fa=HP-UX_B.11.22_IA,v=HP:
/usr/lib/hpux32/libIO77.so.1:
HP-UX libIO77 ILP32_YES B.11.23.%58 (Itanium-based h
ost and target) Wed Jan 8 13:43:44 2003 CST
COMPLIBS.LIBIO77-IA64,fr=B.11.22,fa=HP-UX_B.11.22_IA,v=HP:
/usr/lib/hpux64/libIO77.a:
HP-UX libIO77 LP64_NO B.11.23.%58 (Itanium-based hos
t and target) Wed Jan 8 13:49:28 2003 CST
COMPLIBS.LIBIO77-IS64,fr=B.11.22,fa=HP-UX_B.11.22_IA,v=HP:
/usr/lib/hpux64/libIO77.so.1:
HP-UX libIO77 LP64_YES B.11.23.%58 (Itanium-based ho
st and target) Wed Jan 8 14:00:47 2003 CST
cksum(1) Output:
COMPLIBS.LIBIO77-IA32,fr=B.11.22,fa=HP-UX_B.11.22_IA,v=HP:
3079346885 1288544 /usr/lib/hpux32/libIO77.a
COMPLIBS.LIBIO77-IS32,fr=B.11.22,fa=HP-UX_B.11.22_IA,v=HP:
3962890573 1051200 /usr/lib/hpux32/libIO77.so.1
COMPLIBS.LIBIO77-IA64,fr=B.11.22,fa=HP-UX_B.11.22_IA,v=HP:
2976482156 1540420 /usr/lib/hpux64/libIO77.a
COMPLIBS.LIBIO77-IS64,fr=B.11.22,fa=HP-UX_B.11.22_IA,v=HP:
2641000335 1092392 /usr/lib/hpux64/libIO77.so.1
Patch Conflicts: None
Patch Dependencies: None
Hardware Dependencies: None
Other Dependencies: None
Supersedes:
PHSS_27471
Equivalent Patches:
PHSS_28301:
s700: 10.20
s800: 10.20
PHSS_28302:
s700: 11.00
s800: 11.00
PHSS_28303:
s700: 11.11
s800: 11.11
Patch Package Size: 4890 KBytes
Installation Instructions:
Please review all instructions and the Hewlett-Packard
SupportLine User Guide or your Hewlett-Packard support terms
and conditions for precautions, scope of license,
restrictions, and, limitation of liability and warranties,
before installing this patch.
------------------------------------------------------------
1. Back up your system before installing a patch.
2. Login as root.
3. Copy the patch to the /tmp directory.
4. Move to the /tmp directory and unshar the patch:
cd /tmp
sh PHSS_28304
5. Run swinstall to install the patch:
swinstall -x autoreboot=true -x patch_match_target=true \
-s /tmp/PHSS_28304.depot
By default swinstall will archive the original software in
/var/adm/sw/save/PHSS_28304. If you do not wish to retain a
copy of the original software, include the patch_save_files
option in the swinstall command above:
-x patch_save_files=false
WARNING: If patch_save_files is false when a patch is installed,
the patch cannot be deinstalled. Please be careful
when using this feature.
For future reference, the contents of the PHSS_28304.text file is
available in the product readme:
swlist -l product -a readme -d 
/tmp/PHSS_28304.depot
To put this patch on a magnetic tape and install from the
tape drive, use the command:
dd if=/tmp/PHSS_28304.depot of=/dev/rmt/0m bs=2k
Special Installation Instructions: None
-----End of Document ID: PHSS_28304------------------------------------------
Document ID: PHSS_28303
Date Loaded: 20030219
Title: s700_800 11.11 LIBCL patch
Patch Name: PHSS_28303
Patch Description: s700_800 11.11 LIBCL patch
Creation Date: 03/01/09
Post Date: 03/02/19
Hardware Platforms - OS Releases:
s700: 11.11
s800: 11.11
Products: N/A
Filesets:
OS-Core.CORE-SHLIBS,fr=B.11.11,fa=HP-UX_B.11.11_32/64,v=HP
OS-Core.CORE-64SLIB,fr=B.11.11,fa=HP-UX_B.11.11_32/64,v=HP
ProgSupport.LANG-64ALIB,fr=B.11.11,fa=HP-UX_B.11.11_32/64,v=HP
ProgSupport.LANG-MIN,fr=B.11.11,fa=HP-UX_B.11.11_32/64,v=HP
Automatic Reboot?: No
Status: General Release
Critical: No
Category Tags:
defect_repair enhancement general_release
Path Name: /hp-ux_patches/s700_800/11.X/PHSS_28303
Symptoms:
PHSS_28303:
JAGae48866: Enable 10.20 built applications to access
files > 2GB when run on 11.11
JAGae45769: Abnormal output for 0.0 using P scaling edit
descriptor
PHSS_27470:
JAGae32868: cosd nonzero in some cases.
PHSS_27107:
JAGab14202: U_STACK_TRACE leaks 1360 bytes of memory
JAGad90892: Not handling SIGXCPU signal
JAGad89129: Implicit write cannot open unit # > 99
JAGad92163: Bad Octal representation of numbers > 4294967295
JAGae03743: PROBER in unwind lib causes stack to grow
JAGae05973: Wrong conversion of hex read of 0x80000000
PHSS_25719:
JAGad75314: writing to unit ID 26843546 causes crash.
JAGad54112: Intrinsic function NINT produces incorrect
results.
JAGab19761: Defect passing CDABS and ZABS as a parameter to
a function using shared libcl
JAGad93863: cosd(90) should be 0.0, not 0.6123233996D-16
PHSS_24382:
1. JAGad69726: Implement M Edit descriptor for -f77 compat.
2. JAGad70835: Implement N Edit descriptor for -f77 compat.
3. JAGaa93357: shared_common memory not removed on EXIT.
4. Fixed trap unwind regression.
PHSS_23788:
1. JAGad25172: Signal 11 on throw in large 64 bit aC++
program
2. JAGad51631: unformatted OPEN on text file causes
FORTRAN I/O ERROR 913: OUT OF FREE SPACE
3. 'f90$ga' symbol not found in shared vesion.
Defect Description:
PHSS_28303:
JAGae48866: Enable 10.20 built applications to access
files > 2GB when run on 11.11
JAGae45769: Abnormal output for 0.0 using P scaling edit
descriptor
PHSS_27470:
JAGae32868: cosd in some cases returns non zero (but close)
when exactly zero is needed. This is an
extention of JAGad93863, and f90 flag +trigdacc
is needed to see these results.
PHSS_27107:
JAGab14202: U_STACK_TRACE leaks 1360 bytes of memory
JAGad90892: Not handling SIGXCPU signal
JAGad89129: Implicit write cannot open unit # > 99
JAGad92163: Bad Octal representation of numbers > 4294967295
JAGae03743: The stack is set up with some base address and
some maximum address. For the normal 32bit case
the max is the end of the data quadrant. The
size between the base and max is determined by
maxssiz, which is available through getrlimit(2)
(RLIMIT_STACK). All the addresses between the
base and the max are "valid". If we touch a
valid stack address beyond the current top of
the stack, we allocate physical and virtual
pages around that address. Virtual pages will
be allocated from the base address to the
highest touched address. It doesn't look like
physical pages are allocated for all those
addresses, though. You'll get a SIGSEGV if you
touch an address that isn't valid as defined
above.
JAGae05973: Wrong conversion of hex read of 0x80000000
PHSS_25719:
JAGad75314: Only works on large filesystems. Added a
clearer message to libIO77 when it fails at
exactly 2 GiG mark
JAGad54112: Increased precision of internal datatypes.
JAGab19761: Defect passing CDABS and ZABS as a parameter to
a function using shared libcl
JAGad93863: added new intrinsic routines for cosd(90)
sin(0), and tand(180). Default behavior is
old imprecise values. Use f90 flag
+trigdacc (trig degree accurate) to get new
alternate intrinsics. No other Source change
needed.
PHSS_24382:
1. JAGad04238: libc is real cause... implemented a temp fix.
2. JAGad69726: Implement M Edit descriptor for -f77 compat.
3. JAGad70835: Implement N Edit descriptor for -f77 compat.
4. JAGaa93357: cleaned up shared_common memory on EXIT.
5. Version of trap and unwind component regressed.
PHSS_23788:
1. JAGad25172: workaround bug in libcl by testing if offset
fits in 32 bits. Worked around in libCsup.
2. JAGad51631: re-enabled OPEN error check for unformatted
access to text file. New entry point f90$ckeckuf added
to support previous 11.X functionality. See +nocheckuf
option in F90 for more information.
3. 'f90$ga' added to symbol export list for shared libcl,
this symbol existed in archive verison since PHSS_21950.
Enhancement:
No (superseded patches contained enhancements)
PHSS_27470:
Enhancement to cosd.
SR:
5003415752 8606104417 8606107614 8606129759 8606131152
8606145506 8606158078 8606200550 8606201661 8606206139
8606184910 8606224775 8606219988 8606223059 8606234546
8606236924 8606219352 8606221758 8606284924 8606281826
Patch Files:
OS-Core.CORE-SHLIBS,fr=B.11.11,fa=HP-UX_B.11.11_32/64,v=HP:
/usr/lib/libcl.1
/usr/lib/libcl.2
/usr/lib/nls/msg/C/libcl.cat
OS-Core.CORE-64SLIB,fr=B.11.11,fa=HP-UX_B.11.11_32/64,v=HP:
/usr/lib/pa20_64/libcl.2
ProgSupport.LANG-64ALIB,fr=B.11.11,fa=HP-UX_B.11.11_32/64,
v=HP:
/usr/lib/pa20_64/libcl.a
ProgSupport.LANG-MIN,fr=B.11.11,fa=HP-UX_B.11.11_32/64,v=HP:
/usr/lib/libcl.a
what(1) Output:
OS-Core.CORE-SHLIBS,fr=B.11.11,fa=HP-UX_B.11.11_32/64,v=HP:
/usr/lib/libcl.1:
Trap Library version UX.10.20.15 - 99/12/13
Unwind Library version UX.10.20.16 - 02/09/02
libIO77 HP HPUX [ Release B.10.20.15 PA 32bit ]
(hp700:hp/ux) Jan 7 2003
Copyright (c) 2001 Hewlett Packard.
fs_amod.s $Revision: 1.9.1.1 $
libcl.a version B.10.29.18 - Jan 3, 2002
/usr/lib/libcl.2:
Trap Library version UX.11.01.06 - 02/04/16
Unwind Library version UX.11.01.05 - 00/08/15
libIO77 HP HPUX [ Release B.11.01.15 PA 32bit ]
(hp700:hp/ux) Jan 7 2003
Copyright (c) 2001 Hewlett Packard.
fs_amod.s $Revision: 1.9.1.1 $
libcl.sl version B.11.01.18 - Jan 3 2003
/usr/lib/nls/msg/C/libcl.cat:
None
OS-Core.CORE-64SLIB,fr=B.11.11,fa=HP-UX_B.11.11_32/64,v=HP:
/usr/lib/pa20_64/libcl.2:
libIO77 HP HPUX [ Release B.11.01.15 PA 64bit ]
(hp700:hp/ux) Jan 7 2003
Copyright (c) 2001 Hewlett Packard.
libcl.sl version B.11.01.18 - Jan 3 2003
Trap Library version UX.11.01.06 - 02/04/16
Unwind Library version UX.11.01.05 - 00/08/15
ProgSupport.LANG-64ALIB,fr=B.11.11,fa=HP-UX_B.11.11_32/64,
v=HP:
/usr/lib/pa20_64/libcl.a:
libcl.a version B.11.01.18 - Jan 3 2003
libIO77 HP HPUX [ Release B.11.01.15 PA 64bit ]
(hp700:hp/ux) Jan 7 2003
Copyright (c) 2001 Hewlett Packard.
Unwind Library version UX.11.01.05 - 00/08/15
Trap Library version UX.11.01.06 - 02/04/16
ProgSupport.LANG-MIN,fr=B.11.11,fa=HP-UX_B.11.11_32/64,v=HP:
/usr/lib/libcl.a:
libcl.a version B.11.01.18 - Jan 3 2003
fs_amod.s $Revision: 1.9.1.1 $
libIO77 HP HPUX [ Release B.11.01.15 PA 32bit ]
(hp700:hp/ux) Jan 7 2003
Copyright (c) 2001 Hewlett Packard.
Unwind Library version UX.11.01.05 - 00/08/15
Trap Library version UX.11.01.06 - 02/04/16
cksum(1) Output:
OS-Core.CORE-SHLIBS,fr=B.11.11,fa=HP-UX_B.11.11_32/64,v=HP:
2576109075 1431584 /usr/lib/libcl.1
3503178130 1261568 /usr/lib/libcl.2
1025839270 21393 /usr/lib/nls/msg/C/libcl.cat
OS-Core.CORE-64SLIB,fr=B.11.11,fa=HP-UX_B.11.11_32/64,v=HP:
4287015165 1072416 /usr/lib/pa20_64/libcl.2
ProgSupport.LANG-64ALIB,fr=B.11.11,fa=HP-UX_B.11.11_32/64,
v=HP:
3181378881 1856578 /usr/lib/pa20_64/libcl.a
ProgSupport.LANG-MIN,fr=B.11.11,fa=HP-UX_B.11.11_32/64,v=HP:
370291905 1475160 /usr/lib/libcl.a
Patch Conflicts: None
Patch Dependencies: None
Hardware Dependencies: None
Other Dependencies: None
Supersedes:
PHSS_23788 PHSS_24382 PHSS_25719 PHSS_27107 PHSS_27470
Equivalent Patches:
PHSS_28301:
s700: 10.20
s800: 10.20
PHSS_28302:
s700: 11.00
s800: 11.00
PHSS_28304:
s700: 11.22
s800: 11.22
PHSS_28305:
s700: 11.22
s800: 11.22
Patch Package Size: 6990 KBytes
Installation Instructions:
Please review all instructions and the Hewlett-Packard
SupportLine User Guide or your Hewlett-Packard support terms
and conditions for precautions, scope of license,
restrictions, and, limitation of liability and warranties,
before installing this patch.
------------------------------------------------------------
1. Back up your system before installing a patch.
2. Login as root.
3. Copy the patch to the /tmp directory.
4. Move to the /tmp directory and unshar the patch:
cd /tmp
sh PHSS_28303
5. Run swinstall to install the patch:
swinstall -x autoreboot=true -x patch_match_target=true \
-s /tmp/PHSS_28303.depot
By default swinstall will archive the original software in
/var/adm/sw/save/PHSS_28303. If you do not wish to retain a
copy of the original software, include the patch_save_files
option in the swinstall command above:
-x patch_save_files=false
WARNING: If patch_save_files is false when a patch is installed,
the patch cannot be deinstalled. Please be careful
when using this feature.
For future reference, the contents of the PHSS_28303.text file is
available in the product readme:
swlist -l product -a readme -d /tmp/PHSS_28303.depot
To put this patch on a magnetic tape and install from the
tape drive, use the command:
dd if=/tmp/PHSS_28303.depot of=/dev/rmt/0m bs=2k
Special Installation Instructions: None
-----End of Document ID: PHSS_28303------------------------------------------
Document ID: PHSS_28302
Date Loaded: 20030219
Title: s700_800 11.00 LIBCL patch
Patch Name: PHSS_28302
Patch Description: s700_800 11.00 LIBCL patch
Creation Date: 03/01/09
Post Date: 03/02/19
Hardware Platforms - OS Releases:
s700: 11.00
s800: 11.00
Products: N/A
Filesets:
OS-Core.CORE-SHLIBS,fr=B.11.00,fa=HP-UX_B.11.00_32/64,v=HP
OS-Core.CORE-64SLIB,fr=B.11.00,fa=HP-UX_B.11.00_32/64,v=HP
ProgSupport.LANG-64ALIB,fr=B.11.00,fa=HP-UX_B.11.00_32/64,v=HP
ProgSupport.LANG-MIN,fr=B.11.00,fa=HP-UX_B.11.00_32/64,v=HP
Automatic Reboot?: No
Status: General Release
Critical:
No (superseded patches were critical)
PHSS_21950: OTHER
JAGad14842: Supersedes Recalled patch PHSS_20578
which, causes existing fortran executables (linked
shared) to ignore it's first argument. Possibly
causing wrong answers.
Category Tags:
defect_repair enhancement general_release critical
Path Name: /hp-ux_patches/s700_800/11.X/PHSS_28302
Symptoms:
PHSS_28302:
JAGae48866: Enable 10.20 built applications to access
files > 2GB when run on 11.X
JAGae45769: Abnormal output for 0.0 using P scaling edit
descriptor
PHSS_27469:
JAGae32868: cosd nonzero in some cases.
PHSS_27106:
JAGab14202: U_STACK_TRACE leaks 1360 bytes of memory
JAGad90892: Not handling SIGXCPU signal
JAGad89129: Implicit write cannot open unit # > 99
JAGad92163: Bad Octal representation of numbers > 4294967295
JAGae03743: PROBER in unwind lib causes stack to grow
JAGae05973: Wrong conversion of hex read of 0x80000000
PHSS_25718:
JAGad75314: writing to unit ID 26843546 causes crash.
JAGad54112: Intrinsic function NINT produces incorrect
results.
JAGab19761: Defect passing CDABS and ZABS as a parameter to
a function using shared libcl
JAGad93863: cosd(90) should be 0.0, not 0.6123233996D-16
PHSS_24381:
1. JAGad69726: Implement M Edit descriptor for -f77 compat.
2. JAGad70835: Implement N Edit descriptor for -f77 compat.
3. JAGaa93357: shared_common memory not removed on EXIT.
4. Fixed trap unwind regression.
PHSS_23699:
1. JAGad25172: Signal 11 on throw in large 64 bit aC++
program
2. JAGad51631: unformatted OPEN on text file causes
FORTRAN I/O ERROR 913: OUT OF FREE SPACE
3. 'f90$ga' symbol not found in shared vesion.
PHSS_22653:
1. JAGaa93357: Calls to __F90_F_EXIT always had and exit
code of 0, even when not appropriate.
2. JAGad00306: large real constants not assigned properly
3. JAGad27408: Problem with numbers starting list
directed I/O streams.
PHSS_21950:
1. JAGad14842: Patch PHSS_20578 causes existing fortran
executables (linked shared) to ignore first argument.
2. JAGab77879: Memory leak will occur with Pascal or
Fortran dynamic strings/arrays. (libcl.2 only)
There would also be performance problems or hangs for
dynamic strings/arrays and Pascal's escapecode.
(libcl.1 only)
PHSS_20578:
1. JAGab21216: U_STACK_TRACE only unwinds the frames
up to the sigreturn call, and does not display the
frame of the routine that was running when the signal
was caught.
2. JAGab71918: A throw or escape out of a signal handler
in aC++ will likely cause an abort when used in the catch
clause.
3. JAGac88664: Heap corruption when calling U_STACK_TRACE
4. JAGaa95966: f90 getarg behaves differently than f77
PHSS_19956:
1. JAGab16660: large DATA segment
2. JAGaa68257: I/O error 910 writing to file using f77
NAMELIST and file opened implicitly.
3. JAGab16656: Quad libs consuming too much DATA space
4. Unable to write files > 2GB in 64 bit mode.
PHSS_16849:
1. CLLbs16318: 13 ulp error in LP64 log10
2. CLLbs16323: coshf not symmetric
3. CLLbs16322: tanhf not symmetric
4. CLLbs16310: FTN_QACOSH returns bogus values
5.CLLbs14378 The error message reported to the user when
an allocate frame is encountered by U_get_previous_frame()
is incorrect. "5613 Procedure entry sequence is too long
for Unwind. Contact HP Service." is displayed rather than
"5612 Old version data structures won't restore r3 and r4
for Alloca Unwind." Reminder: U_get_previous_frame_x()
is the new 10.20 interface which permits unwinding
alloca() stack frames.
6.CLLbs16294 When the unwind library is asked to unwind
the stack of code which was interrupted in the prolog of a
function which has non-zero stack size and saves no GR's
FR's, stack pointer return pointer the unwind library may
incorrectly analyze the stack for that function causing a
reference to a random memory location, and thus unexpected
and incorrect behavior. While most user code will never
hit this case, this corner case has been corrected in the
Unwind component of libcl.
PHSS_14583:
1. Support for Fortran90 Huge Data and for entry points
that were updated for better performance.
2.CLLbs13395: Support for C++ Exception Handling with
alloca() frames hangs.
3.CLLbs12177: U_STACK_TRACE() gets an error while unwinding
from a signal handler in an ANSI C++ program that uses
exception handling. If the try block is removed from the
test program the stack trace / unwind works correctly
4.CLLbs13789: Unwind fails on exception frame from a
function with 0 length entry
PHSS_13291:
1. Fortran 90 programs terminate due to segmentation
violations in the trap handling support routines
which are located in libcl.sl and libcl.a. This
symptom affects 64 bit Fortran 90 programs which
make use of the exception handling and stack
trace reporting features of the trap library
(for example, by specifying the
+fp_exception and +DA2.0W compiler command line
options)
2. The Stack Trace feature in "libcl" fails to unwind
and prints an error message referring to pre-release
code upon encountering a signal context stack frame.
The error message reads as follows:
"Interrupt frame marker detected in pa64 be1
work. Unwinding through interrupts not yet
implemented. [SIC]"
3. The stack unwind routines in "libcl" sometimes fail
to unwind a stack which contains calls from one shared
library to another shared library.
Defect Description:
PHSS_28302:
JAGae48866: Enable 10.20 built applications to access
files > 2GB when run on 11.X
JAGae45769: Abnormal output for 0.0 using P scaling edit
descriptor
PHSS_27469:
JAGae32868: cosd in some cases returns non zero (but close)
when exactly zero is needed. This is an
extention of JAGad93863, and f90 flag +trigdacc
is needed to see these results.
PHSS_27106:
JAGab14202: U_STACK_TRACE leaks 1360 bytes of memory
JAGad90892: Not handling SIGXCPU signal
JAGad89129: Implicit write cannot open unit # > 99
JAGad92163: Bad Octal representation of numbers > 4294967295
JAGae03743: The stack is set up with some base address and
some maximum address. For the normal 32bit case
the max is the end of the data quadrant. The
size between the base and max is determined by
maxssiz, which is available through getrlimit(2)
(RLIMIT_STACK). All the addresses between the
base and the max are "valid". If we touch a
valid stack address beyond the current top of
the stack, we allocate physical and virtual
pages around that address. Virtual pages will
be allocated from the base address to the
highest touched address. It doesn't look like
physical pages are allocated for all those
addresses, though. You'll get a SIGSEGV if you
touch an address that isn't valid as defined
above.
JAGae05973: Wrong conversion of hex read of 0x80000000
PHSS_25718:
JAGad75314: Only works on large filesystems. Added a
clearer message to libIO77 when it fails at
exactly 2 GiG mark
JAGad54112: Increased precision of internal datatypes.
JAGab19761: Defect passing CDABS and ZABS as a parameter to
a function using shared libcl
JAGad93863: added new intrinsic routines for cosd(90)
sin(0), and tand(180). Default behavior is
old imprecise values. Use f90 flag
+trigdacc (trig degree accurate) to get new
alternate intrinsics. No other Source change
needed.
PHSS_24381:
1. JAGad69726: Implement M Edit descriptor for -f77 compat.
2. JAGad70835: Implement N Edit descriptor for -f77 compat.
3. JAGaa93357: cleaned up shared_common memory on EXIT.
4. Version of trap and unwind component regressed.
PHSS_23699:
1. JAGad25172: workaround bug in libcl by testing if offset
fits in 32 bits. Worked around in libCsup.
2. JAGad51631: re-enabled OPEN error check for unformatted
access to text file. New entry point f90$ckeckuf added
to support previous 11.X functionality. See +nocheckuf
option in F90 for more information.
3. 'f90$ga' added to symbol export list for shared libcl,
this symbol existed in archive verison since PHSS_21950.
PHSS_22653:
1. JAGaa93357: Calls to __F90_F_EXIT always had and exit
code of 0, even when not appropriate.
2. JAGad00306: treat large real constants like F77.
3. JAGad27408: Problem with numbers starting list
directed I/O streams.
PHSS_21950:
1. JAGad14842: Supersedes Recalled patch PHSS_20578 which,
causes existing fortran executables (linked shared) to
ignore first argument. This patch replaces getargs() to
it's original functionality.
2. JAGab77879: Memory leak
PHSS_20578:
1. JAGab21216: Error in U_STACK_TRACE unwinding past
64 bit _sigreturn
2. JAGab71918: If a throw or escape is done out of a
signal handler that interrupted a system call the values
of the callee save registers (at least R3 and R4) may be
invalid
3. JAGac88664: Use of U_STACK_TRACE on 11.x may corrupt the
heap by one byte if a string ends just over 1Kb multiples.
4. JAGaa95966: f77's getarg is "zero" based, while F90's
is "one" based (the program name is indexed with "1" in
F90's getarg, with "0" in f77, C/C++, and competitors).
PHSS_19956:
1. JAGab16660: Declared constant data "const". Used +ESlit
2. JAGaa68257: Bad assumption when opening files implicitly
3. JAGab16656: Declared constant data "const". Used +ESlit
4. Bad I/O prototype assumptions.
PHSS_16849:
1. CLLbs16318: 13 ulp error in LP64 log10
2. CLLbs16323: coshf not symmetric
3. CLLbs16322: tanhf not symmetric
4. CLLbs16310: FTN_QACOSH returns bogus values
5. CLLbs14378 The message catalog for Unwind was missing
an entry. To reproduce this problem (and thus to
determine whether you need the patch on your system,)
use the following program. Note that this program
uses short cuts which are archive library specific.
It will not link shared. Just for reference, the
program also demonstrates use of the new
U_get_previous_frame_x interface for correct un-
winding through alloca frames.
#include <alloca.h>
#include <stdio.h>
typedef unsigned int address;
typedef unsigned int space;
main()
{
struct {
int curr_frame_size;
address curr_sp;
unsigned long curr_pcspace;
address curr_sp;
unsigned long curr_pcspace;
address curr_pcoffset;
address curr_dp;
address curr_rp;
address curr_mrp;
space curr_sr0, curr_sr4;
int r3;
address cur_r19; /* for PIC code */
int r4;
int reserved;
} cfi;
struct {
int prev_frame_size;
address prev_sp;
space prev_pcspace;
address prev_pcoffset;
int prev_dp;
unsigned int uw_descr[2];
address ustart;
address uend;
int uw_index;
address prev_r19; /* for PIC code */
int r3;
int r4;
} pfi;
#ifdef NEW_INTERFACE
#define UNWIND_STEP(cfi,pfi) \
U_get_previous_frame_x(&cfi,&pfi,sizeof(pfi));
#else
#define UNWIND_STEP(cfi,pfi) \
U_get_previous_frame(&cfi,&pfi);
#endif
void *mptr;
mptr = alloca(1000);
U_get_frame_info(&cfi);
UNWIND_STEP(cfi,pfi);
copy_frame_info(&cfi,&pfi);
UNWIND_STEP(cfi,pfi);
}
/* END */
Compile Line: cc -Ae test_alloca.c -Wl,-aarchive -lcl
$a.out will display, "Procedure entry sequence is too
long for Unwind. Contact HP Service." if patch
PHSS_16690 has not been installed on your system. It will
display "Old version data structures won't restore r3 and
r4 for Alloca Unwind." if the patch has been installed.
6.CLLbs16294 The unwind library was missing logic to
handle a request to "U_get_previous_frame_x" on a stack
frame of a function with an entry prolog of non-zero frame
size and no other saved registers. The following assembly
language program duplicates the condition. When linked
with the pascal program containing a handler which
exercises the unwind library using pascal's Try/Recover
construct a "pre-patch" libcl could possibly dump core.
.level 1.1
.space $TEXT$
.subspa $CODE$
stub
.proc
.callinfo frame=64
.enter
addito,tr -0xF,%r0,%r0
or %r0,%r0,%r0
bv %r0(%rp)
ldo -20(%sp),%sp
break
break 0xC,0
break
break 4,0
break
.leave
.procend
.export stub
.end
Pascal program with handler:
$literal_alias on, standard_level 'hp_modcal'$
program sigU(output);
{ do escape across interrupt marker for HPUX }
const
sigval = hex('FFFFFFFF');
var
oldval,
oldhndlr : integer;
procedure stub; external;
procedure xaritrap $alias 'XARITRAP'$(a,b : integer; \
var c,d : integer);
external;
procedure handle_it( instr, offset, space, \
error : integer );
begin
escape(100);
end;
begin
xaritrap( sigval, baddress(handle_it), oldval, oldhndlr);
try
stub;
recover
if escapecode = 100 then writeln('PASS') else \
writeln('fail');
end. {main}
Build commands:
as stub40.s
pc +DA1.1 stub40.esc0.p stub40.o
PHSS_14583:
1. The following is a list of new entry points:
FTN_KQNINT
IIDNINT
IININT
KIDNINT
KININT
The following is a list of new unsatisfied symbols
in libcl. All of these are resolved in the /usr/lib/milli
patch PHSS_14582.
$$atan2_20
$$exp_20
$$log_20
$$vcossin_20
_U_Qfcnvxf_quad_to_sgl
_U_qdfcnvfxt
__SYSTEM_ID
2. CLLbs13395
The following c++ program hangs when linked archive on an
HP_UX 10.20 system.
#include <stdio.h>
#include <signal.h>
extern "C" void U_STACK_TRACE();
void my_sigfunc(int sig)
{ printf("In my_sigfunc() with signal %d\n", sig);
U_STACK_TRACE();
printf(
"In my_sigfunc() with signal %d after U_STACK_TRACE.\n",
sig);}
int a()
{U_STACK_TRACE(); return 0;}
int main() {
signal(SIGABRT, my_sigfunc);
try { a(); raise(SIGABRT);}
catch(...) {printf("fail - not legal in aCC\n");}}
3.CLLbs12177 test case. this fails a signal 6
#include <stdio.h>
#include <signal.h>
extern "C" void U_STACK_TRACE();
void my_sigfunc(int sig)
{ printf("In my_sigfunc() with signal %d\n", sig);
U_STACK_TRACE();}
int main() {signal(SIGABRT, my_sigfunc);
try {raise(SIGABRT);}
catch(...) {printf("fail - not legal in aCC\n");}}
4.CLLbs13789:
Trap/Unwind libraries do not unwind correctly through a
trap frame generated by the first instruction of a
function which has a zero length entry preamble.
PHSS_13291:
1. The segmentation violation occurred when handling
floating point traps. The following Fortran 90, "x.f"
program reproduces the problem:
PROGRAM CM110
REAL*4 A,B
ON REAL*4 DIV 0 CALL TRAP
A = 0.0
B = 1.0 / A
PRINT *, B
END
SUBROUTINE TRAP(C)
REAL*4 C
PRINT *, 'trap'
C = 0.0
RETURN
END
$f90 -o a64.out -g +fp_exception +DA2.0W x.f
$a64.out
Signal 11: segmentation violation
PROCEDURE TRACEBACK:
Stack_Trace: error while unwinding stack
( 0) 0xc00000000015d0a0 _sigreturn [/lib/pa20_64/libc.2]
2. When displaying a stack trace after a floating point
exception, the unwind library does not display the
stack frames above the exception frame and displays
a "not yet implemented" message.v
The following Fortran 90, program, "test.f" reproduces the
problem:
DOUBLE PRECISION X, Y
X = 1.2345D0
X = X*1.1D0
Y = DLOG(0.0D0-X)
END
$ f90 +fp_exception +DA2.0W +_allow_wide_mode +Z test.f
$ a.out
PROGRAM ABORTED : IEEE invalid operation
PROCEDURE TRACEBACK:
Interrupt frame marker detected in pa64 be1 work.
Unwinding through interrupts not yet implemented.
3. Programs which attempt to display a stack trace which
crosses a shared library boundary (e.g. a function
in "libc.sl" which called a function in "libcl.sl"
fail to unwind through export stubs which are present
in HP_UX 32 bit programs.
The F90 program in item 2, "test.f" when compiled in
32 bit mode (+DA1.1 compiler option) reproduces the
problem:
$ f90 -g +DA1.1 +fp_exception test.f
$ a.out
PROGRAM ABORTED : IEEE invalid operation
PROCEDURE TRACEBACK:
Stack_Trace: error while unwinding stack
( 0) 0xc1055728 FTN_DLOG + 0x28 [/usr/lib//libcl.2]
Enhancement:
No (superseded patches contained enhancements)
PHSS_27469:
Enhancement to cosd.
SR:
5003415752 8606104417 8606129759 8606145506 8606107614
8606131152 8606158078 8606200550 8606201661 8606206139
8606184910 8606224775 8606219352 8606221758 8606219988
8606223059 8606234546 8606236924 8606284924 8606281826
Patch Files:
OS-Core.CORE-SHLIBS,fr=B.11.00,fa=HP-UX_B.11.00_32/64,v=HP:
/usr/lib/libcl.1
/usr/lib/libcl.2
/usr/lib/nls/msg/C/libcl.cat
OS-Core.CORE-64SLIB,fr=B.11.00,fa=HP-UX_B.11.00_32/64,v=HP:
/usr/lib/pa20_64/libcl.2
ProgSupport.LANG-64ALIB,fr=B.11.00,fa=HP-UX_B.11.00_32/64,
v=HP:
/usr/lib/pa20_64/libcl.a
ProgSupport.LANG-MIN,fr=B.11.00,fa=HP-UX_B.11.00_32/64,v=HP:
/usr/lib/libcl.a
what(1) Output:
OS-Core.CORE-SHLIBS,fr=B.11.00,fa=HP-UX_B.11.00_32/64,v=HP:
/usr/lib/libcl.1:
Trap Library version UX.10.20.15 - 99/12/13
Unwind Library version UX.10.20.16 - 02/09/02
libIO77 HP HPUX [ Release B.10.20.15 PA 32bit ]
(hp700:hp/ux) Jan 7 2003
Copyright (c) 2001 Hewlett Packard.
fs_amod.s $Revision: 1.9.1.1 $
libcl.a version B.10.29.18 - Jan 3, 2002
/usr/lib/libcl.2:
Trap Library version UX.11.01.06 - 02/04/16
Unwind Library version UX.11.01.05 - 00/08/15
libIO77 HP HPUX [ Release B.11.01.15 PA 32bit ]
(hp700:hp/ux) Jan 7 2003
Copyright (c) 2001 Hewlett Packard.
fs_amod.s $Revision: 1.9.1.1 $
libcl.sl version B.11.01.18 - Jan 3 2003
/usr/lib/nls/msg/C/libcl.cat:
None
OS-Core.CORE-64SLIB,fr=B.11.00,fa=HP-UX_B.11.00_32/64,v=HP:
/usr/lib/pa20_64/libcl.2:
libIO77 HP HPUX [ Release B.11.01.15 PA 64bit ]
(hp700:hp/ux) Jan 7 2003
Copyright (c) 2001 Hewlett Packard.
libcl.sl version B.11.01.18 - Jan 3 2003
Trap Library version UX.11.01.06 - 02/04/16
Unwind Library version UX.11.01.05 - 00/08/15
ProgSupport.LANG-64ALIB,fr=B.11.00,fa=HP-UX_B.11.00_32/64,
v=HP:
/usr/lib/pa20_64/libcl.a:
libcl.a version B.11.01.18 - Jan 3 2003
libIO77 HP HPUX [ Release B.11.01.15 PA 64bit ]
(hp700:hp/ux) Jan 7 2003
Copyright (c) 2001 Hewlett Packard.
Unwind Library version UX.11.01.05 - 00/08/15
Trap Library version UX.11.01.06 - 02/04/16
ProgSupport.LANG-MIN,fr=B.11.00,fa=HP-UX_B.11.00_32/64,v=HP:
/usr/lib/libcl.a:
libcl.a version B.11.01.18 - Jan 3 2003
fs_amod.s $Revision: 1.9.1.1 $
libIO77 HP HPUX [ Release B.11.01.15 PA 32bit ]
(hp700:hp/ux) Jan 7 2003
Copyright (c) 2001 Hewlett Packard.
Unwind Library version UX.11.01.05 - 00/08/15
Trap Library version UX.11.01.06 - 02/04/16
cksum(1) Output:
OS-Core.CORE-SHLIBS,fr=B.11.00,fa=HP-UX_B.11.00_32/64,v=HP:
2576109075 1431584 /usr/lib/libcl.1
3503178130 1261568 /usr/lib/libcl.2
1025839270 21393 /usr/lib/nls/msg/C/libcl.cat
OS-Core.CORE-64SLIB,fr=B.11.00,fa=HP-UX_B.11.00_32/64,v=HP:
4287015165 1072416 /usr/lib/pa20_64/libcl.2
ProgSupport.LANG-64ALIB,fr=B.11.00,fa=HP-UX_B.11.00_32/64,
v=HP:
3181378881 1856578 /usr/lib/pa20_64/libcl.a
ProgSupport.LANG-MIN,fr=B.11.00,fa=HP-UX_B.11.00_32/64,v=HP:
370291905 1475160 /usr/lib/libcl.a
Patch Conflicts: None
Patch Dependencies: None
Hardware Dependencies: None
Other Dependencies: None
Supersedes:
PHSS_13291 PHSS_14583 PHSS_16849 PHSS_19956 PHSS_20578 PHSS_21950
PHSS_22653 PHSS_23699 PHSS_24381 PHSS_25718 PHSS_27106 PHSS_27469
Equivalent Patches:
PHSS_28301:
s700: 10.20
s800: 10.20
PHSS_28303:
s700: 11.11
s800: 11.11
PHSS_28304:
s700: 11.22
s800: 11.22
PHSS_28305:
s700: 11.22
s800: 11.22
Patch Package Size: 7010 KBytes
Installation Instructions:
Please review all instructions and the Hewlett-Packard
SupportLine User Guide or your Hewlett-Packard support terms
and conditions for precautions, scope of license,
restrictions, and, limitation of liability and warranties,
before installing this patch.
------------------------------------------------------------
1. Back up your system before installing a patch.
2. Login as root.
3. Copy the patch to the /tmp directory.
4. Move to the /tmp directory and unshar the patch:
cd /tmp
sh PHSS_28302
5. Run swinstall to install the patch:
swinstall -x autoreboot=true -x patch_match_target=true \
-s /tmp/PHSS_28302.depot
By default swinstall will archive the original software in
/var/adm/sw/save/PHSS_28302. If you do not wish to retain a
copy of the original software, include the patch_save_files
option in the swinstall command above:
-x patch_save_files=false
WARNING: If patch_save_files is false when a patch is installed,
the patch cannot be deinstalled. Please be careful
when using this feature.
For future reference, the contents of the PHSS_28302.text file is
available in the product readme:
swlist -l product -a readme -d /tmp/PHSS_28302.depot
To put this patch on a magnetic tape and install from the
tape drive, use the command:
dd if=/tmp/PHSS_28302.depot of=/dev/rmt/0m bs=2k
Special Installation Instructions: None
-----End of Document ID: PHSS_28302------------------------------------------
Document ID: PHKL_28602
Date Loaded: 20030219
Title: s700_800 11.00 VxFS fsadm,MMF hang,mmap() perf improvement
Patch Name: PHKL_28602
Patch Description: s700_800 11.00 VxFS fsadm,MMF hang,mmap() perf improvement
Creation Date: 03/02/05
Post Date: 03/02/19
Hardware Platforms - OS Releases:
s700: 11.00
s800: 11.00
Products:
JFS 3.3;
Filesets:
JFS.JFS-BASE2-KRN,fr=3.3,fa=HP-UX_B.11.00_32,v=HP
JFS.JFS-BASE2-KRN,fr=3.3,fa=HP-UX_B.11.00_64,v=HP
Automatic Reboot?: Yes
Status: General Release
Critical:
Yes
PHKL_28602: HANG
PHKL_21499: HANG
PHKL_27212: HANG
PHKL_25021: PANIC HANG
PHKL_21774: HANG
PHKL_23773: HANG
PHKL_23192: PANIC HANG CORRUPTION
PHKL_22121: HANG
PHKL_21063: PANIC
PHKL_21181: PANIC
PHKL_21765: HANG
PHKL_21077: MEMORY_LEAK CORRUPTION
Category Tags:
defect_repair enhancement general_release critical panic
halts_system corruption memory_leak
Path Name: /hp-ux_patches/s700_800/11.X/PHKL_28602
Symptoms:
PHKL_28602:
( SR:8606280077 CR:JAGae44052 )
Filesystem commands that freeze filesystems (such as
fsadm resize) may deadlock with processes that use
F_WRLCK ioctl() to place locks on files in the filesystem.
A stack trace of the hung command will look similar to the
following:
_sleep+0x210
sleep_spinunlock+0x70
vx_event_wait+0xc0
vx_delay2+0x64
vx_freeze_level+0x278
vx_freeze+0x3c
vx_resize+0x120
vx_aioctl_full+0xf8
vx_aioctl_common+0x3b4
vx_aioctl+0xbc
vx_ioctl+0xc0
vno_ioctl+0x98
ioctl+0x120
syscall+0x750
$syscallrtn+0x0
Meanwhile, other threads will be sleeping with a stack trace
similar to the following:
_swtch+0xc4
_sleep+0x4cc
locked+0xd84
vx_rdwr+0x234
vno_rw+0x80
read+0x10c
syscall+0x204
$syscallrtn+0x0
_sleep+0x210
sleep_spinunlock+0x70
vx_event_wait+0xc0
vx_delay2+0x64
vx_active_common_flush+0xb8
vx_lockctl+0x3c
fcntl+0x2d4
syscall+0x750
$syscallrtn+0x0
PHKL_27503:
( SR:8606249539 CR:JAGae15929 )
Enhancement: Improve performance of VxFS processing
of mmap(2)ed files.
PHKL_21499:
( SR: 8606129559 DTS: JAGac87894 )
VxFS may hang if more than one process writes to a memory
mapped file. To reproduce the problem, map a file to memory
and keep writing to the pages. Start two other processes
which write to this file. VxFS hangs.
PHKL_27212:
( SR:8606236411 CR:JAGae05468 )
The system will hang when starting the Veritas HSM
Migration software due to a deadlock condition in
the DMAPI (see vxenablef(1M)).
( SR:8606236403 CR:JAGae05460 )
Attempting to mount the same filesystem on two
different directories may leave the corresponding
Logical Volume Manager (LVM) volume group in a state in
which it cannot be deactivated, even though its logical
volumes are not in use. The following steps would cause the
problem to be seen:
# mount /dev/vg100/lvol1 /tmp_mnt
# mount /dev/vg100/lvol1 /tmp_mnt2
vxfs mount: /dev/vg100/lvol1 is already mounted, /tmp_mnt2
is busy, or allowable number of mount points exceeded
# umount /tmp_mnt
# vgchange -a n vg100
vgchange: Couldn't deactivate volume group "vg100":
Device busy
PHKL_26470:
( SR:8606223056 CR:JAGad92160 )
After a file has been created with 000 permissions in a
directory with default ACLs on a VxFS filesystem using the
version 4 layout, all files created in that directory will
be created with 000 permissions until the filesystem that
directory is in is unmounted and remounted, or the system
is rebooted. The file which initiated this condition need
not have 000 permissions at the end of the process that
started its creation. For example, if a file is copied
into this directory using cp(1), this condition will be
triggered. cp(1) first creates a file with 000
permissions, and then sets the permissions as appropriate,
using the permissions of the copied file combined with the
user's umask and the parent directory's permission mask,
which is based upon the directory's default ACLs. Following
this, some commands such as cp(1) and mv(1) will behave as
documented since they explicitly set the permissions of
their target files, but others, such as touch(1) will
create files with 000 permissions since they do not
explicitly set the permissions of their targets, but
instead only use the user's umask and the parent
directory's mask.
PHKL_26670:
( SR:8606249754 CR:JAGae16140 )
PHKL_25021 introduced behavior that can cause VxFS 3.3
file system performance problems when sequential I/O
requests of less than 64KB are performed. This behavior
can affect backup utilities and other applications that
perform sequential I/O accesses to the file system.
PHKL_25021:
( SR:8606202113 CR:JAGad71287 )
VxFS 3.3 may hang a 64 bit system by zeroing out the
interrupt mask (cr15).
( SR:8606196201 CR:JAGad65404 )
Data Page Fault panic while using Hyperfabric network. The
stack trace may looks as follows:
panic+0x14
report_trap_or_int_and_panic+0x84
interrupt+0x1d4
$ihndlr_rtn+0x0
sendfile_rele+0x304
freeb_pullupmsg+0x238
freeb+0x7b4
CLIC_SEND+0x1ecc
clicdlpi_wput+0x140
putnext+0xcc
ip_wput_ire+0x454
ip_wput+0x470
putnext+0xcc
tcp_timer+0x334
tcp_wput+0x828
puthere+0x148
mi_timeout_exec+0x294
sw_service+0xb0
mp_ext_interrupt+0x150
ivti_patch_to_nop3+0x0
idle+0x81c
( SR:8606200313 CR:JAGad69497 )
Poor performance with VxFS 3.3 compared to VxFS 3.1 while
multiple threads/processes read from a file simultaneously.
( SR:8606203915 CR:JAGad73093 )
PHKL_23192 introduced behavior that can cause multiple
threads to deadlock on a specific file. The behavior only
occurs if the file is being flushed at the same time it is
being remotely accessed, such as via NFS or other networking
access. The behavior could eventually lead to hung
processes and subsystems, such as NFS. This problem happens
only if VxFS 3.3 fancy read ahead is turned on. Stacks of
deadlocked threads may look like as shown below:
_sleep+0x1fc
sleep_spinunlock+0x74
vx_event_wait+0xf0
vx_delay2+0x64
vx_vnode_flush+0x3bc
vx_do_putpage+0x11c
vx_idelxwri_flush+0xf8
vx_delxwri_flush+0x1dc
vx_worklist_process+0x1b0
vx_worklist_thread+0x4c
kthread_daemon_startup+0x24
_sleep+0x1fc
vx_rwsleep_lock+0x1f4
vx_iglock2+0x78
vx_iglock+0x28
vx_fancy_read_ahead+0x1c4
vx_read1+0x980
vx_vn_bread+0xfc
rfs_read+0x158
rfsexp_dispatch+0x210
svc_getreq+0x13c
svc_run+0x1e0
nfsexp_svc+0x4d4
nfs_stub_svc+0xa4
coerce_scall_args+0xcc
syscall+0x6f8
syscallinit+0x54c
( SR:8606181938 CR:JAGad51154 )
Data Page Fault in allocbuf1(). Stack of the panic thread
may look like:
allocbuf1+0xd0
allocbuf+0x30
vx_allocbuf+0x30
vx_async_shorten+0xc30
vx_fsync+0x120
rfs3_commit+0x3a0
rfsexp_dispatch+0x810
svc_getreq+0x500
svc_run+0x930
nfsexp_svc+0x860
nfs_stub_svc+0x390
coerce_scall_args+0x570
syscall+0xe70
PHKL_21774:
( SR: 8606139611 CR: JAGad08922 )
1. Executing an FSO (File Sharing Option) refresh of a file
causes subsequent executions of more(1) on that file to
hang. Any read-only node opening a refreshed file may
experience similar failures (hangs).
2. A filesystem with FSO enabled may hang while flushing
(writing memory contents to disk). If other filesystems
interact with this filesystem, the entire system may hang.
To recover the filesystem, the system must be rebooted.
Both of these problems are specific to FSO on top of
VxFS3.3. Without FSO installed and enabled, this patch will
have no impact on the system (the components delivered in
this patch, while part of VxFS3.3, are currently used only
by FSO).
PHKL_23773:
( SR: 8606178276 CR: JAGad49578 )
Backup utilities like fbackup take longer time to finish
when JFS 3.3 fancy read ahead is enabled.
( SR: 8606179211 CR: JAGad48435 )
sar -v and glance show ninode usage(count of active inodes
on the system) as 0.
( SR: 8606178276 CR: JAGad47503 )
System hang caused by someone going through the
buffer pool to flush out dirty buffers. The buffer
is marked dirty but also marked busy, and the process
or thread decide to wait for it. Because of this,
ServiceGuard commands like cmapplyconf, cmcheckconf.
hang. A typical stack of the hung thread is given
below.
_swtch+0x138
real_sleep+0x234
_sleep+0x14
syncip_flush_cache+0x190
vx_flushdev+0x10
vx_fsync+0x1cc
spec_fsync+0x17c
spec_inactive+0x14
vn_rele+0x1e8
vno_close+0x68
closef+0x68
close+0x30
syscall+0x75c
$syscallrtn+0x0
PHKL_23254:
(SR: 8606177460 DTS: JAGad46692)
Poor application performance with VxFS 3.3 compared to
VxFS 3.1, when the application does a lot of reads from
random offsets of a file.
PHKL_23192:
( SR: 8606175336 CR: JAGad44578 )
Poor system performance with VxFS vs HFS, when applications
read backward and forward through a file which cannot be
fully contained in the buffer cache.
( SR: 8606171316 CR: JAGad40579 )
System may panic or hang during heavy VxFS filesystem use.
Due to the nature of these problems, a panic/hang could
occur in any VxFS ('vx_') routine. There is no specific
stack trace for reference.
( SR: 8606168320 CR: JAGad37601 )
close(2) may cause data corruption by truncating VxFS files.
( SR: 8606162599 CR: JAGad31915 )
VxFS filesystems on disc devices with block size 4096 bytes
could not be mounted.
( SR: 8606152097 CR: JAGad21436 )
While umounting a VxFS snapshot filesystem, the filesystem
may hang. A reboot would be required to reaccess the
filesystem.
PHKL_22121:
( SR: 8606135462 CR: JAGad04596 )
VxFS 3.3 write(2) may return incorrect error value 61441 to
applications on error.
( SR: 8606138051 CR: JAGad07239 )
System hangs due to VxFS 3.3 deadlock during direct I/O. If
a system memory core dump is taken once the system is hung,
the stack trace will be similar to the following:
_swtch+0x1e4
1c_1d20_cl_real_sleep+0x1464
_sleep_one+0x16c
vx_rwsleep_lock+0x108
vx_iglock2+0x58
vx_iglock+0x28
vx_pagein+0x1a4
virtual_fault+0x7d8
vfault+0x274
trap+0x12f0
$RDB_trap_patch+0x30
lacc+0xdc
vx_dio_iovec+0xd0
vx_dio_rdwri+0x278
vx_write1+0x654
vx_rdwr+0x1c0
vno_rw+0xbc
rwuio+0x154
aio_rw_child_thread+0x204
kthread_daemon_startup+0x2c
kthread_daemon_startup+0x0
PHKL_21773:
( SR: 8606139352 CR: JAGad08645 )
VxFS 3.3 direct I/O returns 61441 to applications on error.
PHKL_21063:
( SR: 8606114160 CR: JAGac23138 )
A panic under heavy load on a multi-processor machine.
The stack trace might look like:
panic+0x14
panic+0x48
report_trap_or_int_and_panic+0x7c
trap+0x119c
nokgdb+0x8
vx_fancyra_predict+0x10
vx_read1+0x2a8
vx_rdwr_0x4c0
vn_rdwr+0x84
exec_file_read+0x40
get_aout_info+0xc0
load_dld+0x10c
hdl_load_process+0x300
getxfile+0x33c
execve+0x1bec
syscall+0x5fc
$syscallrtn
PHKL_21181:
( SR: 8606129265 DTS: JAGac86811 )
Data Page Fault in VxFS 3.3 vx_memunlock() while using
direct IO. The stack of the panic thread is given below.
panic+0x14
report_trap_or_int_and_panic+0x80
trap+0xdb8
nokgdb+0x8
vx_memunlock+0x80
vx_dio_iovec+0x74c
vx_dio_rdwri+0x194
vx_write1+0xc58
vx_rdwr+0x3dc
vno_rw+0x84
rwuio+0xe8
aio_rw_child_thread+0x80
kthread_daemon_startup+0x24
kthread_daemon_startup+0x0
vx_memunlock+0x20
PHKL_21938:
( SR: 8606142678 DTS: JAGad12033 )
Rcp is slow when file size exceeds 2GB.
PHKL_21765:
( SR: 8606137227 DTS: JAGad06345 )
Unmounting VxFS filesystems may hang.
PHKL_21077:
( SR: 8606114161 DTS: JAGac23139 )
System calls like open and stat cause memory leak,
while accessing VxFS 3.3 file system.
( SR: 8606113817 DTS: JAGac12337 )
ftruncate() corrupts the last page in a memory
mapped file.
Defect Description:
PHKL_28602:
( SR:8606280077 CR:JAGae44052 )
A thread trying to read a file is allowed to sleep
waiting for the read/write lock owned by another
thread while the filesystem active count is incremented.
The hang occurs when another thread tries to freeze the
filesystem before the read/write lock can be released,
and hangs in vx_freeze_level(). When the thread tries
to release the read/write lock, it hangs in
vx_active_common_flush() due to the pending freeze.
Resolution:
The active count is now decremented before calling locked()
and sleeping on the read/write lock. Once the lock is
obtained, the active count is incremented again.
PHKL_27503:
( SR:8606249539 CR:JAGae15929 )
VxFS handles faults on memory mapped pages one at a time and
does synchronous writes to the intent log for each one.
Resolution:
VxFS will read ahead up to 64 pages when processing a fault
on a memory mapped file and the associated intent log
processing will always be asynchronous.
PHKL_21499:
( SR: 8606129559 DTS: JAGac87894 )
If more than one process write to a memory mapped file,
a deadlock may occur between inode locks and buffer
cache, because of the incorrect ordering of the locks.
Resolution:
Deadlock is avoided by returning VX_ERETRY
if the buffer is busy, instead of waiting for it.
PHKL_27212:
( SR:8606236411 CR:JAGae05468 )
The DMAPI code holds the inode's read-write lock in
exclusive mode. Another thread then tries to acquire
the lock in shared mode for initializing ACL counts of an
inode. This causes the deadlock.
Resolution:
The code is changed so that the difference in locking
methods is expected so that deadlock does not occur.
( SR:8606236403 CR:JAGae05460 )
After the mount of a logical volume failed because the
volume was already in use (mounted), the volume group
device was left open as a result of the failed mount.
This prevents the subsequent deactivation of the volume
from completing successfully.
Resolution:
The error path now makes sure that the device is
appropriately closed.
PHKL_26470:
( SR:8606223056 CR:JAGad92160 )
The mode of a newly created file on a VxFS filesystem
using the version 4 layout (see vxupgrade(1M)) is
determined by the mode passed into open(2) or creat(2)
combined with the user's umask and the modemask of a
directory, determined by its default ACLs. When default
ACLs exist, the modemask of the directory would be set to
the mode passed into open(2) or creat(2). From the time
when a file is created with mode 000 on, all subsequently
created files in that directory with default ACLs would be
given 000 permissions. The permissions of affected files
can be corrected with chmod(1).
Resolution:
The modemasks of directories are no longer affected by the
modes of created files.
PHKL_26670:
( SR:8606249754 CR:JAGae16140 )
The readahead algorithm was incorrectly using
max_buf_data_size (default 8Kb) to calculate the
readahead size instead of read_pref_io (default
64kb), which resulted in a smaller readahead size.
Resolution:
Use read_pref_io to calculate the readahead size.
PHKL_25021:
( SR:8606202113 CR:JAGad71287 )
VxFS 3.3 was using a wrong data type (unsigned int) with
spinlock functions. Spinlock functions use "unsigned long".
Resolution:
Use the correct data type (unsigned long) with spinlock
functions.
( SR:8606196201 CR:JAGad65404 )
VxFS may free a vnode while a buffer associated with that
vnode is in use in sendfile(2). Later when the sendfile
code accesses the vnode through the buffer, the system
panics.
Resolution:
Set up a dummy vnode which is not freed and use that vnode
for buffers passed to sendfile(2) so that sendfile(2)
code will always be accessing a valid vnode.
( SR:8606200313 CR:JAGad69497 )
1. VxFS 3.3 read ahead code relies on the unexpected read
(buffer cache miss) count to determine if there is any
pressure on buffer cache. If there is pressure on buffer
cache, the amount of read ahead performed is reduced. If
the read ahead performed is less than the size of the read
size, there will be unexpected reads. This turns out into a
self fulfilling prophecy and read ahead is completely turned
off causing each read to read the data from the disk
synchronously degrading performance.
2. VxFS 3.3 uses the vxtunefs(1M) tunable "read_pref_io"
to calculate the intial and maximum read ahead region
length. But the read sizes could vary and this tunable may
not trigger enough read ahead for different read sizes.
Resolution:
1. Use the current read request size as the floor for the
read ahead region length.
2. Use the current read request size times the vxtunefs(1M)
"read_nstream" as the ceiling for the read ahead region
length.
( SR:8606203915 CR:JAGad73093 )
VxFS 3.3 read code was acquiring two locks in the wrong
order causing deadlock.
Resolution:
VxFS 3.3 read code is rewritten eliminating the need to
acquire the locks in the wrong order.
( SR:8606181938 CR:JAGad51154 )
allocbuf1() expects a non-zero value as its second
argument, but VxFS was not always passing a non-zero value.
Resolution:
Make sure that a non-zero value is always passed as the
second argument to allocbuf1() from VxFS code.
PHKL_21774:
( SR: 8606139611 CR: JAGad08922 )
1) The vnode counter was set incorrectly to 0, after which
another thread was waiting for the value to go to 1 forever.
Resolution:
Do not set the vnode counter to 0 if the current value is 1
for VxFS3.3.
2) The FSO flushing function tries to incorrectly reacquire
a lock which it already holds, causing the FSO filesystem to
hang.
Resolution:
The FSO flushing function should not try to acquire the
lock.
PHKL_23773:
( SR: 8606178276 CR: JAGad49578 )
VxFS fancy read ahead feature degrades read performance
when multiple process/threads read from a file
simultaneously. Because the VxFS fancy read ahead feature
was designed to handle only one reader for a file at a time,
when there are multiple readers for a file, its read ahead
pattern matching algorithm fails. This results in a lot of
unnecessary I/Os.
Resolution:
When there are multiple readers to a file, disable fancy
read feature and do sequential read ahead instead.
( SR: 8606179211 CR: JAGad48435 )
A function was decrementing the number of active inode
counts instead of incrementing when inodes become
active.
Resolution:
Increment active inode count when inodes become active.
( SR: 8606178276 CR: JAGad47503 )
Under some corner cases, some VxFS meta buffers(map buffers)
will be left in a busy and dirty state. This leads to hang
any process which is directly or indirectly doing a buffer
cache flush.
Resolution:
Make sure the VxFS meta buffers are flushed before doing
a buffer cache flush.
PHKL_23254:
(SR: 8606177460 DTS: JAGad46692)
A sanity check at the wrong place in a function disabled
read ahead for random reads to a file.
Resolution:
Move the sanity check to the correct place in the function
so that some read ahead could be done for random reads also.
PHKL_23192:
( SR: 8606175336 CR: JAGad44578 )
An incorrect sanity check in the VxFS read path turned off
both normal read ahead and fancy read ahead when the fancy
read ahead was enabled on the system. This caused large I/O
wait times for each read.
Resolution:
Corrected the logic error in the sanity check code such that
fancy read ahead is used when enabled.
( SR: 8606171316 CR: JAGad40579 )
Some routines processing VxFS inodes did not follow the
locking rules correctly, causing race conditions during
heavy stress to the inode cache which resulted in system
hangs or panics.
Resolution:
Enforced correct use of the locking protocol in VxFS icache
functions.
( SR: 8606168320 CR: JAGad37601 )
While closing (close(2)) a file, VxFS tried to free up extra
space (extents) allocated to the file. While doing this, it
incorrectly modified the actual file size, causing data
loss.
Resolution:
Corrected logic to prevent modification of file size while
freeing extra extents.
( SR: 8606162599 CR: JAGad31915 )
VxFS was not calculating the disc block size correctly.
This prevented filesystems on discs with block size of 4096
bytes from being mounted.
Resolution:
Using a different function (from LVM) to read the correct
disc block size.
( SR: 8606152097 CR: JAGad21436 )
Due to a race condition between two snapshot mounts, a lock
on the file system may never get released.
Resolution:
Make sure that the filesystem lock is released in all cases.
PHKL_22121:
( SR: 8606135462 CR: JAGad04596 )
VxFS 3.3 write() was returning an internal error value
VX_ERETRY(61441) to applications on certain conditions.
Resolution:
Reset error = 0 in VxFS 3.3 write function so that internal
errors won't be returned to applications.
( SR: 8606138051 CR: JAGad07239 )
During direct I/O VxFS 3.3 was holding a lock which was not
required. It may cause a deadlock if a page fault occurs
while holding the lock, since this lock is required while
resolving the page fault.
VxFS does not prevent direct I/O on memory mapped files.
Resolution:
Remove the unnecessary locks during direct I/O and skip
drect I/O if the file is memory mapped.
PHKL_21773:
( SR: 8606139352 CR: JAGad08645 )
Direct I/O was returning an internal error VX_ERETRY(61441)
to applications on certain conditions.
Resolution:
If direct I/O fails with VX_ERETRY, continue with
normal I/O.
PHKL_21063:
( SR: 8606114160 CR: JAGac23138 )
This is caused by checking a pointer for NON-NULL then
acquiring the lock that protects the pointer and then
using the pointer without re-checking it. While
acquiring the lock, the pointer could have been freed.
Resolution:
Re-check the pointer after acquiring the lock.
PHKL_21181:
( SR: 8606129265 DTS: JAGac86811 )
vx_memlock() sets length of actual memory locked incorrectly
if it could not lock the requested length of memory.
Resolution: mp->m_len is set correctly in vx_memlock().
PHKL_21938:
( SR: 8606142678 DTS: JAGad12033 )
When the file size is greater than 2GB, blocks of file are
read twice since the first read is invoked with wrong flags
and fails.
Resolution: Use the correct flags for the first read.
PHKL_21765:
( SR: 8606137227 DTS: JAGad06345 )
Empty buffers returned by vx_vn_bread() are not freed by
sendfile_rele() because of the B_BUSY flag set by
vx_geteblk(), which causes unmount to loop waiting
for the filesystem to become inactive.
Resolution:
Unset the B_BUSY flag of the buffers in vx_vn_brelse().
PHKL_21077:
( SR: 8606114161 DTS: JAGac23139 )
vx_real_readdir() called by vx_readdir() allocates
memory which is not freed always, causing memory
leak.
Resolution:
Removed the unnecessary VX_ZALLOC() call from
vx_real_readdir().
( SR: 8606113817 DTS: JAGac12337 )
There was an error in calculating the page offset in
vx_setattr().
Resolution:
Corrected the expression calculating the page offset.
Enhancement:
No (superseded patches contained enhancements)
PHKL_27503:
Improve performance of VxFS processing of mmap(2)ed
files.
SR:
8606113817 8606114160 8606114161 8606129265 8606129559
8606135462 8606137227 8606138051 8606139352 8606139611
8606142678 8606152097 8606162599 8606168320 8606171316
8606175336 8606177460 8606178276 8606179211 8606180357
8606181938 8606196201 8606200313 8606202113 8606203915
8606223056 8606236403 8606236411 8606249539 8606249754
8606280077
Patch Files:
JFS.JFS-BASE2-KRN,fr=3.3,fa=HP-UX_B.11.00_32,v=HP:
/usr/conf/lib/libvxfs.a(vx33_bio1.o)
/usr/conf/lib/libvxfs.a(vx33_iflush.o)
/usr/conf/lib/libvxfs.a(vx33_inode.o)
/usr/conf/lib/libvxfs.a(vx33_kdmi.o)
/usr/conf/lib/libvxfs.a(vx33_mount.o)
/usr/conf/lib/libvxfs.a(vx33_rdwri.o)
/usr/conf/lib/libvxfs.a(vx33_vm.o)
/usr/conf/lib/libvxfs.a(vx33_vnops.o)
/usr/conf/lib/libvxfs.a(vx_acl.o)
/usr/conf/lib/libvxfs.a(vx_kdmi_machdep.o)
/usr/conf/lib/libvxfs.a(vx_machdep.o)
JFS.JFS-BASE2-KRN,fr=3.3,fa=HP-UX_B.11.00_64,v=HP:
/usr/conf/lib/libvxfs.a(vx33_bio1.o)
/usr/conf/lib/libvxfs.a(vx33_iflush.o)
/usr/conf/lib/libvxfs.a(vx33_inode.o)
/usr/conf/lib/libvxfs.a(vx33_kdmi.o)
/usr/conf/lib/libvxfs.a(vx33_mount.o)
/usr/conf/lib/libvxfs.a(vx33_rdwri.o)
/usr/conf/lib/libvxfs.a(vx33_vm.o)
/usr/conf/lib/libvxfs.a(vx33_vnops.o)
/usr/conf/lib/libvxfs.a(vx_acl.o)
/usr/conf/lib/libvxfs.a(vx_kdmi_machdep.o)
/usr/conf/lib/libvxfs.a(vx_machdep.o)
what(1) Output:
JFS.JFS-BASE2-KRN,fr=3.3,fa=HP-UX_B.11.00_32,v=HP:
/usr/conf/lib/libvxfs.a(vx33_bio1.o):
vx33_bio1.c $Date: 2000/04/04 10:49:23 $Revision: r1
1ros/2 PATCH_11.00 (PHKL_21499)
/usr/conf/lib/libvxfs.a(vx33_iflush.o):
vx33_iflush.c $Date: 2001/08/21 12:43:47 $Revision:
r11ros/4 PATCH_11.00 (PHKL_25021)
/usr/conf/lib/libvxfs.a(vx33_inode.o):
vx33_inode.c $Date: 2001/08/21 12:43:47 $Revision: r
11ros/5 PATCH_11.00 (PHKL_25021)
/usr/conf/lib/libvxfs.a(vx33_kdmi.o):
vx33_kdmi.c $Date: 2002/05/30 13:21:58 $Revision: r1
1ros/2 PATCH_11.00 (PHKL_27212)
/usr/conf/lib/libvxfs.a(vx33_mount.o):
vx33_mount.c $Date: 2002/05/30 13:23:16 $Revision: r
11ros/3 PATCH_11.00 (PHKL_27212)
/usr/conf/lib/libvxfs.a(vx33_rdwri.o):
vx33_rdwri.c $Date: 2002/07/25 07:33:38 $Revision: r
11ros/11 PATCH_11.00 (PHKL_27503)
/usr/conf/lib/libvxfs.a(vx33_vm.o):
vx33_vm.c $Date: 2002/07/25 07:31:11 $Revision: r11r
os/3 PATCH_11.00 (PHKL_27503)
/usr/conf/lib/libvxfs.a(vx33_vnops.o):
vx33_vnops.c $Date: 2003/02/03 21:16:12 $Revision: r
11ros/8 PATCH_11.00 (PHKL_28602)
/usr/conf/lib/libvxfs.a(vx_acl.o):
vx_acl.c $Date: 2002/05/30 13:25:37 $Revision: r11ro
s/3 PATCH_11.00 (PHKL_27212)
/usr/conf/lib/libvxfs.a(vx_kdmi_machdep.o):
vx_kdmi_machdep.c $Date: 2001/08/21 12:43:47 $Revisi
on: r11ros/4 PATCH_11.00 (PHKL_25021)
/usr/conf/lib/libvxfs.a(vx_machdep.o):
vx_machdep.c $Date: 2001/01/19 07:56:53 $Revision: r
11ros/3 PATCH_11.00 (PHKL_23192)
JFS.JFS-BASE2-KRN,fr=3.3,fa=HP-UX_B.11.00_64,v=HP:
/usr/conf/lib/libvxfs.a(vx33_bio1.o):
vx33_bio1.c $Date: 2000/04/04 10:49:23 $Revision: r1
1ros/2 PATCH_11.00 (PHKL_21499)
/usr/conf/lib/libvxfs.a(vx33_iflush.o):
vx33_iflush.c $Date: 2001/08/21 12:43:47 $Revision:
r11ros/4 PATCH_11.00 (PHKL_25021)
/usr/conf/lib/libvxfs.a(vx33_inode.o):
vx33_inode.c $Date: 2001/08/21 12:43:47 $Revision: r
11ros/5 PATCH_11.00 (PHKL_25021)
/usr/conf/lib/libvxfs.a(vx33_kdmi.o):
vx33_kdmi.c $Date: 2002/05/30 13:21:58 $Revision: r1
1ros/2 PATCH_11.00 (PHKL_27212)
/usr/conf/lib/libvxfs.a(vx33_mount.o):
vx33_mount.c $Date: 2002/05/30 13:23:16 $Revision: r
11ros/3 PATCH_11.00 (PHKL_27212)
/usr/conf/lib/libvxfs.a(vx33_rdwri.o):
vx33_rdwri.c $Date: 2002/07/25 07:33:38 $Revision: r
11ros/11 PATCH_11.00 (PHKL_27503)
/usr/conf/lib/libvxfs.a(vx33_vm.o):
vx33_vm.c $Date: 2002/07/25 07:31:11 $Revision: r11r
os/3 PATCH_11.00 (PHKL_27503)
/usr/conf/lib/libvxfs.a(vx33_vnops.o):
vx33_vnops.c $Date: 2003/02/03 21:16:12 $Revision: r
11ros/8 PATCH_11.00 (PHKL_28602)
/usr/conf/lib/libvxfs.a(vx_acl.o):
vx_acl.c $Date: 2002/05/30 13:25:37 $Revision: r11ro
s/3 PATCH_11.00 (PHKL_27212)
/usr/conf/lib/libvxfs.a(vx_kdmi_machdep.o):
vx_kdmi_machdep.c $Date: 2001/08/21 12:43:47 $Revisi
on: r11ros/4 PATCH_11.00 (PHKL_25021)
/usr/conf/lib/libvxfs.a(vx_machdep.o):
vx_machdep.c $Date: 2001/01/19 07:56:53 $Revision: r
11ros/3 PATCH_11.00 (PHKL_23192)
cksum(1) Output:
JFS.JFS-BASE2-KRN,fr=3.3,fa=HP-UX_B.11.00_32,v=HP:
2741150537 8312 /usr/conf/lib/libvxfs.a(vx33_bio1.o)
3654757209 35408 /usr/conf/lib/libvxfs.a(vx33_iflush.o)
970546392 54492 /usr/conf/lib/libvxfs.a(vx33_inode.o)
4031971225 27080 /usr/conf/lib/libvxfs.a(vx33_kdmi.o)
800299073 37196 /usr/conf/lib/libvxfs.a(vx33_mount.o)
2183435764 45752 /usr/conf/lib/libvxfs.a(vx33_rdwri.o)
2275774202 14916 /usr/conf/lib/libvxfs.a(vx33_vm.o)
3973500888 42852 /usr/conf/lib/libvxfs.a(vx33_vnops.o)
4249138238 10932 /usr/conf/lib/libvxfs.a(vx_acl.o)
3537314736 7300 /usr/conf/lib/libvxfs.a(vx_kdmi_machdep.o)
1234978424 31748 /usr/conf/lib/libvxfs.a(vx_machdep.o)
JFS.JFS-BASE2-KRN,fr=3.3,fa=HP-UX_B.11.00_64,v=HP:
528274726 13104 /usr/conf/lib/libvxfs.a(vx33_bio1.o)
546572476 83176 /usr/conf/lib/libvxfs.a(vx33_iflush.o)
2681524324 124520 /usr/conf/lib/libvxfs.a(vx33_inode.o)
3876639993 56048 /usr/conf/lib/libvxfs.a(vx33_kdmi.o)
2406994812 76496 /usr/conf/lib/libvxfs.a(vx33_mount.o)
4044434194 69840 /usr/conf/lib/libvxfs.a(vx33_rdwri.o)
319977418 25000 /usr/conf/lib/libvxfs.a(vx33_vm.o)
625148030 80688 /usr/conf/lib/libvxfs.a(vx33_vnops.o)
1042676792 20928 /usr/conf/lib/libvxfs.a(vx_acl.o)
2331806707 12136 /usr/conf/lib/libvxfs.a(vx_kdmi_machdep.o)
1210492418 81752 /usr/conf/lib/libvxfs.a(vx_machdep.o)
Patch Conflicts: None
Patch Dependencies: None
Hardware Dependencies: None
Other Dependencies: None
Supersedes:
PHKL_27503 PHKL_27212 PHKL_26670 PHKL_26470 PHKL_25021 PHKL_23773
PHKL_23254 PHKL_23192 PHKL_22121 PHKL_21938 PHKL_21774 PHKL_21773
PHKL_21765 PHKL_21499 PHKL_21181 PHKL_21077 PHKL_21063
Equivalent Patches: None
Patch Package Size: 1010 KBytes
Installation Instructions:
Please review all instructions and the Hewlett-Packard
SupportLine User Guide or your Hewlett-Packard support terms
and conditions for precautions, scope of license,
restrictions, and, limitation of liability and warranties,
before installing this patch.
------------------------------------------------------------
1. Back up your system before installing a patch.
2. Login as root.
3. Copy the patch to the /tmp directory.
4. Move to the /tmp directory and unshar the patch:
cd /tmp
sh PHKL_28602
5. Run swinstall to install the patch:
swinstall -x autoreboot=true -x patch_match_target=true \
-s /tmp/PHKL_28602.depot
By default swinstall will archive the original software in
/var/adm/sw/save/PHKL_28602. If you do not wish to retain a
copy of the original software, include the patch_save_files
option in the swinstall command above:
-x patch_save_files=false
WARNING: If patch_save_files is false when a patch is installed,
the patch cannot be deinstalled. Please be careful
when using this feature.
For future reference, the contents of the PHKL_28602.text file is
available in the product readme:
swlist -l product -a readme -d /tmp/PHKL_28602.depot
To put this patch on a magnetic tape and install from the
tape drive, use the command:
dd if=/tmp/PHKL_28602.depot of=/dev/rmt/0m bs=2k
Special Installation Instructions: None
-----End of Document ID: PHKL_28602------------------------------------------
Document ID: PHCO_28413
Date Loaded: 20030219
Title: s700_800 11.00 pax(1) cumulative patch
Patch Name: PHCO_28413
Patch Description: s700_800 11.00 pax(1) cumulative patch
Creation Date: 03/01/30
Post Date: 03/02/19
Hardware Platforms - OS Releases:
s700: 11.00
s800: 11.00
Products: N/A
Filesets:
OS-Core.UX-CORE,fr=B.11.00,fa=HP-UX_B.11.00_32/64,v=HP
Automatic Reboot?: No
Status: General Release
Critical: No
Category Tags:
defect_repair enhancement general_release
Path Name: /hp-ux_patches/s700_800/11.X/PHCO_28413
Symptoms:
PHCO_28413:
( SR: 8606288039 CR: JAGae51972 )
When pax(1M) needs user attention it opens /dev/tty to
display messages. If this happens when the command is
run from a cron/at job the open fails because there's
no controlling terminal and the user gets a misleading
message saying /dev/tty is unavailable.
( SR: 8606292412 CR: JAGae56175 )
Enhancement Request: pax(1) is enhanced to support the
archival of files up to 8GB in USTAR format from 2GB.
PHCO_25418:
( SR: 8606216121 CR: JAGad85291 )
When using pax's copy mode, -rw, SUID/SGID bits are not
restored.
Example -- (current user is ville)
$chmod 4755 toto
$ll
total 2
-rwsr-xr-x 1 ville users 18 Aug 20 11:20 toto
$mkdir titi
$pax -pe -rw toto titi
$ll
total 2
drwx--x--x 2 ville users 96 Aug 20 11:21 titi
-rwsr-xr-x 1 ville users 18 Aug 20 11:20 toto
$cd titi
$ll
total 2
-rwx--x--x 1 ville users 18 Aug 20 11:20 toto
$
Note:
destination file toto (in directory titi) has lost his
suid bit, although -pe option was used.
(SR: 8606175067 CR: JAGad44311 )
Pax unable to restore UIG/GID > 60k.
Upon retrieving an archive using pax, the UID/GID greater
than 60K were replaced by the UID/GID of the current
process.
PHCO_20416:
For device special files with a concatenated
major/minor value greater than pax's maximum
small file size (0x7FFFFFFF - 2GB), pax fails
with the following error message:
/dev/XXXX > 2GB. Not Dumped.
PHCO_20027:
Enhancements for JFS 3.3 Access Control Lists.
PHCO_16885:
1. Pax does not handle soft/hard links properly
in ustar format if the file/link names
have a length >= 100 characters.
2. Pax cannot restore soft links with a length
> 100 characters in cpio format.
3. Pax does not restore hard links in cpio format.
Defect Description:
PHCO_28413:
( SR: 8606288039 CR: JAGae51972 )
pax(1M) doesn't display real failure reason when unable
to open /dev/tty.
Resolution:
A single line of code change which prints the real failure
reason on stderr.
( SR: 8606292412 CR: JAGae56175 )
Enhance pax(1M) to support archival of files up to 8GB
from the current limit of 2GB.
Resolution:
Modified pax(1M) to support archival of large files up to
8GB.
PHCO_25418:
( SR: 8606216121 CR: JAGad85291 )
Pax command fails to preserve the SUID/SGID bit when used
to move files, even though -pe option is used.
Resolution:
Restore SUID/SGID bits after the file is created.
( SR: 8606175067 CR: JAGad44311 )
The current pax command supports UID/GID only up to 60K.
Need to add support for large UID/GID.
Resolution:
Cpio format of pax supports UID/GID <= 262143
Tar format of pax supports UID/GID <= 2097151 if the uname/
gname doens't exist on the system. But if it does, tar
format of pax supports UID/GID up to 2G.
PHCO_20416:
The concatenated major/minor value for device special
files was incorrectly being range checked against
pax's maximum small file size (0x7FFFFFFF - 2GB).
Resolution:
The concatenated major/minor value for device special
file is not checked against 2GB.
PHCO_20027:
Changes to pax(1) to function correctly with Access
Control Lists (ACLs) feature introduced in JFS 3.3.
PHCO_16885:
1. In ustar format, pax will fail to create a
hard link, and the string "ustar" will be
appended to the link target of any symbolic
link where the target name is exactly 100
characters.
There is no error message output to indicate
to the user that the tar format of pax will not
backup symbolic/hard links if the link target
name is > 100 characters in length.
2. In cpio format, pax fill fail to create symbolic
links if the link target name is > 100
characters in length.
3. In cpio format, pax fails to restore hard links;
it creates regular files instead.
Enhancement:
Yes
PHCO_28413:
This product update makes pax(1M) large file aware.
SR:
8606288039 8606292412 8606216121 8606175067 4701406967
8606105686 5003311746 4701406967
Patch Files:
OS-Core.UX-CORE,fr=B.11.00,fa=HP-UX_B.11.00_32/64,v=HP:
/usr/lib/nls/msg/C/pax.cat
/sbin/pax
/usr/bin/pax
what(1) Output:
OS-Core.UX-CORE,fr=B.11.00,fa=HP-UX_B.11.00_32/64,v=HP:
/usr/lib/nls/msg/C/pax.cat:
None
OS-Core.UX-CORE,fr=B.11.00,fa=HP-UX_B.11.00_32/64,v=HP:
/sbin/pax:
$Revision: 82.13.1.9 $
$ PATCH/11.00:PHCO_27731 Oct 11 2002 12:59:13 $
PATCH_11_00: hpux_rel.o pax.o append.o buffer.o cpio
.o create.o extract.o fileio.o hash.o link.o
list.o mem.o namelist.o names.o pass.o path
name.o replace.o tar.o ttyio.o warn.o charma
p.o 03/01/30
OS-Core.UX-CORE,fr=B.11.00,fa=HP-UX_B.11.00_32/64,v=HP:
/usr/bin/pax:
$Revision: 82.13.1.9 $
PATCH_11_00: hpux_rel.o pax.o append.o buffer.o cpio
.o create.o extract.o fileio.o hash.o link.o
list.o mem.o namelist.o names.o pass.o path
name.o replace.o tar.o ttyio.o warn.o charma
p.o 03/01/30
cksum(1) Output:
OS-Core.UX-CORE,fr=B.11.00,fa=HP-UX_B.11.00_32/64,v=HP:
1299790370 6328 /usr/lib/nls/msg/C/pax.cat
OS-Core.UX-CORE,fr=B.11.00,fa=HP-UX_B.11.00_32/64,v=HP:
1493651740 405504 /sbin/pax
OS-Core.UX-CORE,fr=B.11.00,fa=HP-UX_B.11.00_32/64,v=HP:
3470529720 102400 /usr/bin/pax
Patch Conflicts: None
Patch Dependencies:
s700: 11.00: PHKL_28202
s800: 11.00: PHKL_28202
Hardware Dependencies: None
Other Dependencies: None
Supersedes:
PHCO_16885 PHCO_20027 PHCO_20416 PHCO_25418
Equivalent Patches:
PHCO_28414:
s700: 11.11
s800: 11.11
Patch Package Size: 530 KBytes
Installation Instructions:
Please review all instructions and the Hewlett-Packard
SupportLine User Guide or your Hewlett-Packard support terms
and conditions for precautions, scope of license,
restrictions, and, limitation of liability and warranties,
before installing this patch.
------------------------------------------------------------
1. Back up your system before installing a patch.
2. Login as root.
3. Copy the patch to the /tmp directory.
4. Move to the /tmp directory and unshar the patch:
cd /tmp
sh PHCO_28413
5. Run swinstall to install the patch:
swinstall -x autoreboot=true -x patch_match_target=true \
-s /tmp/PHCO_28413.depot
By default swinstall will archive the original software in
/var/adm/sw/save/PHCO_28413. If you do not wish to retain a
copy of the original software, include the patch_save_files
option in the swinstall command above:
-x patch_save_files=false
WARNING: If patch_save_files is false when a patch is installed,
the patch cannot be deinstalled. Please be careful
when using this feature.
For future reference, the contents of the PHCO_28413.text file is
available in the product readme:
swlist -l product -a readme -d /tmp/PHCO_28413.depot
To put this patch on a magnetic tape and install from the
tape drive, use the command:
dd if=/tmp/PHCO_28413.depot of=/dev/rmt/0m bs=2k
Special Installation Instructions: None
-----End of Document ID: PHCO_28413------------------------------------------
Document ID: PHSS_28339
Date Loaded: 20030218
Title: s700_800 11.X OV OB4.10 patch - SSEA packet
Patch Name: PHSS_28339
Patch Description: s700_800 11.X OV OB4.10 patch - SSEA packet
Creation Date: 03/02/14
Post Date: 03/02/18
Hardware Platforms - OS Releases:
s700: 11.00 11.11 11.20
s800: 11.00 11.11 11.20
Products:
OmniBackII A.04.10
Filesets:
OMNIBACK-II.OMNI-SSEA-P,fr=A.04.10,fa=HP-UX_B.11.00_32/64,v=HP
Automatic Reboot?: No
Status: General Release
Critical: No
Category Tags:
defect_repair general_release manual_dependencies
Path Name: /hp-ux_patches/s700_800/11.X/PHSS_28339
Symptoms:
PHSS_28339:
Id = HSLco38493 SR: H555009140
Aborting SSEA Integration backup session after the
mirror disks have been split, leaves backup system
enabled.
Id = HSLco39063 SR: H555009270
SSEA Integration backup occasionally fails with:
"|Critical| From: BSM<hostname> <barlist>
Time: <time>
None of the Disk Agents completed successfully.
Session has failed."
Id = HSLco39896 SR: H555009573
Instant Recovery fails reporting an invalid BC pairs
status.
PHSS_27594:
Id = HSLco36407 SR:H555008644
During an OmniBack II XP Integration restore
following error message is reported:
"[Major] From: SSEA<hostname> "" Time: <time>
Filesystem /dev/vx/dsk/<disk_group>/<lvol>
could not be dismounted from /<mountpoint>
(Details unknown.)".
Id = HSLco36669 SR:H555008685
In some cases SSEA takes a long time to resolve
all the objects on the backup side.
Id = HSLco37123 SR:H555008788
SSEA backup fails with:
"/usr/sbin/vgchange -q n -a y /dev/<vg_name>
[Major] From: SSEA<hostname> "" Time: <time>
Volume group /dev/<vg_name> could not be activated.".
Id = HSLco37703 SR:H555008886
SSEA agent abnormally ends on the application side
after Disk Agents are completed successfully.
Id = HSLco37704 SR:H555008887
The connection with the host running SSEA agent is
broken.
PHSS_26884:
Id = HSLco32243 SR: H555006991
OmniBack II XP split-mirror backup fails with:
"[Critical] From: SSEA<hostname> "" Time: <time>
Failed to attach to the command device <device>.
(Bad catalog access for message #10822 in set 12!)"
Id = HSLco32969 SR: H555007253
OmniBack II XP split-mirror backup fails with:
"[Major] From: SSEA<hostname> "" Time: <time>
Failed to split the BC pair of
LDEV <ldev#>, SEQ# <seq#>, MU# <mu#>."
Id = HSLco33105 SR: H555007298
SSEA agent dies ungracefully right after the last
medium is unloaded at the end of backup.
Id = HSLco33681 SR: H555007572
OmniBack II XP integration sessions fail with:
"[Major] From: SSEA<hostname> "" Time: <time>
Failed to split the <BC or CA> pair of
LDEV <ldev#>, SEQ# <seq#>, MU# <mu#>."
Id = HSLco33697 SR: H555007572
OmniBack II XP Integration backup reports:
"[Warning] From: SSEA<hostname> "" Time: <time>
File system <filesystem_path> could not be mounted
to <mountpoint>. ([27] File too large)."
Id = HSLco34841 SR: H555008141
Pre- and post-exec scripts are improperly executed under
shell or "sh -c" type.
Id = HSLco35984 SR: H555008573
During an OmniBack II XP Integration backup following
error message is reported:
"[Critical] From: SSEA<hostname> "" Time: <time>
Failed to attach to the command device <device>.
(Raid Manager LIB Error: [2] Can't attach to the
command device.)"
Id = HSLco36041 SR: H555008585
A warning:
"[Warning] From: SSEA<hostname> "" Time: <time>
Failed to get a BC pair for LDEV 0, MU# 0
in RAID <num>. (Details unknown.)"
is reported during an OmniBack II XP Integration
backup.
Id = JAGae22283 SR: 8606257980
OmniBack II XP Integration backup using split-mirror
fails with:
"[Major] From: SSEA<hostname> "" Time: <time>
File system for / object not mounted or local
disk is in backup."
Defect Description:
PHSS_28339:
Id = HSLco38493
Clean-up procedure is not executed after backup session
is aborted. Volume groups on backup host are left
activated with the filesystems mounted.
Resolution:
Proper clean-up procedure (deactivation of volume
groups, umounting the filesystems and re-synching the
mirror) is executed after backup session is aborted.
Part of SSPUX410_055.
Id = HSLco39063
Backup session fails because of incorrect information
sent to pairvolstat() API.
Resolution:
The data structure carrying the information to be
provided to pairvolstat() API is re-initialized before
executing new pairvolstat() call.
Id = HSLco39896
If only one BC pair target disk is configured for
Instant Recovery, then the session fails due to
incorrect checking for more target disks.
Resolution:
Status checking for target disks is improved in a
manner that it correctly recognizes what disks are
going to be used for Instant Recovery.
PHSS_27594:
Id = HSLco36407
If during the restore SSEA fails to unmount the
only mountpoint present in the session on the
application side, agent proceeds with synching
and splitting the disks although it should abort
the session.
Resolution:
If the only mountpoint present in the session fails
to be unmounted on the application side, then the
session fails immediately.
If there are multiple mountpoints in the session,
and some of the mountpoints fail to be unmounted
then SSEA skips the restore of the disks related
to those mountpoints, but proceeds the restore of
disks related to the mountpoints that were
successfully unmounted.
Id = HSLco36669
Resolving of objects from Business Copy XP on backup
host proceeds very slowly if some (out of backup)
objects are in PAIR state:
-CA PAIR are in PSUS/SUSU: resolving takes less
then <x> minutes,
-CA PAIR are in PAIR state: resolving takes more
then <6*x> minutes.
The resolving algorithm has a very bad performance
in this case.
Resolution:
The resolving algorithm is improved.
Part of SSPUX410_041.
Id = HSLco37123
If the application side is configured in a MC/SG
cluster then the volume groups on the backup side
can not be activated in the read/write mode with
"vgchange -q n -a y" command.
Resolution:
The volume group on the backup side is marked as
"not clustered" with the "vgchange -q n -c n" command
before the activation in the read/write mode happens.
To change this behavior set the SMB_CLUS_DEACT=1
environment variable in the .omnirc file on the R2 host.
Part of SSPUX410_044.
Id = HSLco37703
The allocated buffer size is too small.
Resolution:
The allocated buffer size is increased.
Id = HSLco37704
In environments where the firewall between Cell Server
and client(s) with firewall idle timeout of <x>
minutes is present, it can happen that an agent is in
idle state for the amount of time longer than mentioned
timeout. In such cases, firewall breaks the connection
between Cell Server and the client(s).
Resolution:
New environment variable OB2IPCKEEPALIVE is introduced.
Setting its value to any of YES, TRUE, ON or 1 in
/opt/omni/.omnirc file on Cell Server causes sending a
dummy package from time to time in order to keep the
connection open.
Part of SSPUX410_046.
PHSS_26884:
Id = HSLco32243
SSEA agent tries to attach to the first command device
it finds. In cases when first command device is busy,
SSEA fails to attach.
Resolution:
A list of command devices is created and if one of them
is busy, SSEA agent tries to attach to the next one.
Id = HSLco32969
SSEA agent collects physical volume names from
"vgdisplay" output. In environments where physical
volumes are configured their names are collected as
well although they are not needed for resolving
process. This causes backup to fail.
Resolution:
SSEA agent skips parsing of "vgdisplay" command as
soon as "PVG name" token is encountered.
Part of SSPUX410_010.
Id = HSLco33105
SSEA agent dies when updating XP database because of
buffering problem.
Resolution:
Buffer holding logical devices information is extended.
Part of SSPUX410_011.
Id = HSLco33681
A "Major" message is displayed instead of "Warning" is
environments with LVM mirroring on application side.
Resolution:
Message level is lowered to the level documented in
XP Integration Guide.
Id = HSLco33697
Some combinations of mount API options and ownerships
of disk groups cause failures when import mapfile from
Application to Backup host.
Resolution:
Command line mount is used instead of mount API.
Id = HSLco34841
Scripts are executed with ".%s" which is "sh" instead of
script name.
Resolution:
The dot is removed from function that executes the
scripts.
Id = HSLco35984
Attaching fails in environments where alternate links
are used because of incorrect instance number counting.
Resolution:
Instance numbers are properly counted if alternate links
are used.
Id = HSLco36041
SSEA agent improperly handles "LDEV 0" (unknown logical
device).
Resolution:
Logical device resolution is improved by adding
additional checking in cases where "LDEV 0" is
encountered.
Part of SSPUX410_032.
Id = JAGae22283
SSEA agent cannot handle "/" (root) mountpoint when
performing a split-mirror backup.
Resolution:
Mountpoint "/" (root) can be used in SSEA split-mirror
backup if following is satisfied:
1. Root mountpoint is located on an XP disk.
2. SSEA_PRESERVE_MOUNTPOINT environment variable is not
set to 1 on Backup host.
3. SSEA_BACKUP_ROOT_MOUNTPOINT environment variable is
set to 1 on Application side.
Note: Such backup cannot be used for Instant Recovery!
Enhancement:
No
SR:
H555009270 H555009140 H555009573 H555008685 H555008644
H555008886 H555008788 H555008887 H555006991 H555008585
H555007298 H555007748 8606257980 H555008141 H555007253
H555007572 H555007564 H555008573
Patch Files:
OMNIBACK-II.OMNI-SSEA-P,fr=A.04.10,fa=HP-UX_B.11.00_32/64,
v=HP:
/opt/omni/databases/vendor/ssea/hp/s800/hp-ux-1020/A.04.10/
packet.Z
/opt/omni/databases/vendor/ssea/hp/s800/hp-ux-11/A.04.10/
packet.Z
/opt/omni/databases/vendor/ssea/sun/sparc/solaris-7/A.04.10/
packet.Z
what(1) Output:
OMNIBACK-II.OMNI-SSEA-P,fr=A.04.10,fa=HP-UX_B.11.00_32/64,
v=HP:
/opt/omni/databases/vendor/ssea/hp/s800/hp-ux-1020/A.04.10/
packet.Z:
None
/opt/omni/databases/vendor/ssea/hp/s800/hp-ux-11/A.04.10/
packet.Z:
None
/opt/omni/databases/vendor/ssea/sun/sparc/solaris-7/A.04.10/
packet.Z:
None
cksum(1) Output:
OMNIBACK-II.OMNI-SSEA-P,fr=A.04.10,fa=HP-UX_B.11.00_32/64,
v=HP:
2213735778 523405 /opt/omni/databases/vendor/ssea/hp/s800/
hp-ux-1020/A.04.10/packet.Z
1430846738 539185 /opt/omni/databases/vendor/ssea/hp/s800/
hp-ux-11/A.04.10/packet.Z
3153860961 251805 /opt/omni/databases/vendor/ssea/sun/sparc/
solaris-7/A.04.10/packet.Z
Patch Conflicts: None
Patch Dependencies: None
Hardware Dependencies: None
Other Dependencies:
If any newer SSP for the Cell Console is installed
on your system, you must re-apply the SSP after
installing this Cell Console patch.
PHSS_27603
Release Date: Available
Id = HSLco37704 SR:H555008887
PHSS_27603 (or later) is required for this defect fix.
Supersedes:
PHSS_27594 PHSS_26884
Equivalent Patches:
PHSS_28338:
s700: 10.20
s800: 10.20
Patch Package Size: 1320 KBytes
Installation Instructions:
Please review all instructions and the Hewlett-Packard
SupportLine User Guide or your Hewlett-Packard support terms
and conditions for precautions, scope of license,
restrictions, and, limitation of liability and warranties,
before installing this patch.
------------------------------------------------------------
1. Back up your system before installing a patch.
2. Login as root.
3. Copy the patch to the /tmp directory.
4. Move to the /tmp directory and unshar the patch:
cd /tmp
sh PHSS_28339
5. Run swinstall to install the patch:
swinstall -x autoreboot=true -x patch_match_target=true \
-s /tmp/PHSS_28339.depot
By default swinstall will archive the original software in
/var/adm/sw/save/PHSS_28339. If you do not wish to retain a
copy of the original software, include the patch_save_files
option in the swinstall command above:
-x patch_save_files=false
WARNING: If patch_save_files is false when a patch is installed,
the patch cannot be deinstalled. Please be careful
when using this feature.
For future reference, the contents of the PHSS_28339.text file is
available in the product readme:
swlist -l product -a readme -d /tmp/PHSS_28339.depot
To put this patch on a magnetic tape and install from the
tape drive, use the command:
dd if=/tmp/PHSS_28339.depot of=/dev/rmt/0m bs=2k
Special Installation Instructions:
PHSS_28339:
NOTE!
This patch has to be installed on the Installation
Server host.
The OB2 SSEA agent must be redistributed to the SSEA
client system(s) after the installation of the patch
in order to get the fixes into effect.
For distribution of the SSEA agent, refer to
the OmniBackII Installation guide.
Patch includes the following Site Specific Patches:
SSPUX410_010
SSPUX410_011
SSPUX410_032
SSPUX410_041
SSPUX410_044
SSPUX410_046
SSPUX410_055
-----End of Document ID: PHSS_28339------------------------------------------
Document ID: PHKL_28488
Date Loaded: 20030218
Title: s700_800 11.00 copyin EFAULT, LDCD access type
Patch Name: PHKL_28488
Patch Description: s700_800 11.00 copyin EFAULT, LDCD access type
Creation Date: 03/01/15
Post Date: 03/02/18
Hardware Platforms - OS Releases:
s700: 11.00
s800: 11.00
Products: N/A
Filesets:
OS-Core.CORE2-KRN,fr=B.11.00,fa=HP-UX_B.11.00_32,v=HP
OS-Core.CORE2-KRN,fr=B.11.00,fa=HP-UX_B.11.00_64,v=HP
Automatic Reboot?: Yes
Status: General Release
Critical:
Yes
PHKL_28488: HANG
PHKL_24005: OTHER
application sees EFAULT even while using a valid
address.
Category Tags:
defect_repair general_release critical halts_system
Path Name: /hp-ux_patches/s700_800/11.X/PHKL_28488
Symptoms:
PHKL_28488:
( SR:8606287496 CR:JAGae51435 )
A 64bit application calling fork() may experience an
interruptible hang.
PHKL_24005:
( SR: 8606157079 CR:JAGad26413)
System returns Bad address (EFAULT) even when the
application supplied a valid address.
Defect Description:
PHKL_28488:
( SR:8606287496 CR:JAGae51435 )
If the first access to a copy on write page after a fork()
is an LDCD (load and clear double) instruction, when
resolving the protection (access) fault, the access type
is incorrectly determined to be a READ instead of a WRITE,
thus resulting in a recursive protection fault and
interruptible process hang.
Resolution:
Updated the access type checking routines to correctly
return WRITE for LDCD instructions.
PHKL_24005:
( SR: 8606157079 CR:JAGad26413)
The system checks the access rights for the page twice
before copying a page in memory. The second check is
done if the page was not already in memory and was
brought in by a fault. This second check can fail
under a race condition.
Resolution:
The solution is not to check the access rights again if
the fault is resolved - resolving the fault ensures that
the page is in memory with proper access rights.
Enhancement:
No
SR:
8606157079 8606287496
Patch Files:
OS-Core.CORE2-KRN,fr=B.11.00,fa=HP-UX_B.11.00_32,v=HP:
/usr/conf/lib/libhp-ux.a(vm_mapcall.o)
OS-Core.CORE2-KRN,fr=B.11.00,fa=HP-UX_B.11.00_64,v=HP:
/usr/conf/lib/libhp-ux.a(vm_mapcall.o)
what(1) Output:
OS-Core.CORE2-KRN,fr=B.11.00,fa=HP-UX_B.11.00_32,v=HP:
/usr/conf/lib/libhp-ux.a(vm_mapcall.o):
vm_mapcall.c $Date: 2003/01/06 08:21:25 $Revision: r
11ros/4 PATCH_11.00 (PHKL_28488)
OS-Core.CORE2-KRN,fr=B.11.00,fa=HP-UX_B.11.00_64,v=HP:
/usr/conf/lib/libhp-ux.a(vm_mapcall.o):
vm_mapcall.c $Date: 2003/01/06 08:21:25 $Revision: r
11ros/4 PATCH_11.00 (PHKL_28488)
cksum(1) Output:
OS-Core.CORE2-KRN,fr=B.11.00,fa=HP-UX_B.11.00_32,v=HP:
1540202931 9380 /usr/conf/lib/libhp-ux.a(vm_mapcall.o)
OS-Core.CORE2-KRN,fr=B.11.00,fa=HP-UX_B.11.00_64,v=HP:
1973053988 28152 /usr/conf/lib/libhp-ux.a(vm_mapcall.o)
Patch Conflicts: None
Patch Dependencies:
s700: 11.00: PHKL_18543
s800: 11.00: PHKL_18543
Hardware Dependencies: None
Other Dependencies: None
Supersedes:
PHKL_24005
Equivalent Patches: None
Patch Package Size: 60 KBytes
Installation Instructions:
Please review all instructions and the Hewlett-Packard
SupportLine User Guide or your Hewlett-Packard support terms
and conditions for precautions, scope of license,
restrictions, and, limitation of liability and warranties,
before installing this patch.
------------------------------------------------------------
1. Back up your system before installing a patch.
2. Login as root.
3. Copy the patch to the /tmp directory.
4. Move to the /tmp directory and unshar the patch:
cd /tmp
sh PHKL_28488
5. Run swinstall to install the patch:
swinstall -x autoreboot=true -x patch_match_target=true \
-s /tmp/PHKL_28488.depot
By default swinstall will archive the original software in
/var/adm/sw/save/PHKL_28488. If you do not wish to retain a
copy of the original software, include the patch_save_files
option in the swinstall command above:
-x patch_save_files=false
WARNING: If patch_save_files is false when a patch is installed,
the patch cannot be deinstalled. Please be careful
when using this feature.
For future reference, the contents of the PHKL_28488.text file is
available in the product readme:
swlist -l product -a readme -d /tmp/PHKL_28488.depot
To put this patch on a magnetic tape and install from the
tape drive, use the command:
dd if=/tmp/PHKL_28488.depot of=/dev/rmt/0m bs=2k
Special Installation Instructions: None
-----End of Document ID: PHKL_28488------------------------------------------
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]