Patch Name: PHKL_13768 Patch Description: s700 10.20 Panic,Hang in memory management cumulative patch Creation Date: 98/01/19 Post Date: 98/02/03 Hardware Platforms - OS Releases: s700: 10.20 Products: N/A Filesets: OS-Core.CORE-KRN Automatic Reboot?: Yes Status: General Superseded Critical: Yes PHKL_13768: OTHER Prevents MC/Service Guard TOC on PA8000 systems. PHKL_11164: PANIC PHKL_11055: HANG PHKL_8683: PANIC Path Name: /hp-ux_patches/s700/10.X/PHKL_13768 Symptoms: PHKL_13768: Temporary system hang on PA8000 systems, possibly resulting in TOC to preserve system integrity (if running MC/Service Guard). When analysing the system crash dump, a processor would be executing in the kernel routine alloc_large_page(). PHKL_11164: The system message buffer would show many "sysmap: rmap ovflo, lost [...]". Eventually, the system would panic with "kalloc: out of kernel virtual space". This problem would only be seen on PA8000 systems. PHKL_11055: Large memory systems could hang while trying to allocate kernel memory. A TOC crash dump would show a processor executing in alloc_large_page() while other processors would be spinning waiting for the pfdat_lock spinlock. This problem would only be seen on PA8000 systems. PHKL_8683: System pacin with data page fault on ICS. Defect Description: PHKL_13768: On PA8000 systems, when free physical memory becomes heavily fragmented, the time needed to find free large pages (superpages) increases drastically. During this time (possibly several seconds) the kernel would preempt any user or system processes. This could result in MC/Service Guard TOC'ing the system. The fix was to yield the cpu when spending too much time in alloc_large_page(). PHKL_11164: The variable page size memory allocator was preallocating kernel virtual space aligned on the largest possible superpage size instead of the largest available. The excess was then freed, causing gaps (fragmentation) in the sysmap. The fix was to allocate kernel virtual space only after the determination of the largest available superpage size. This problem would only be seen on PA8000 systems. PHKL_11055: Large memory MP systems could hang if the processors were mallocing kernel dynamic data at the same time. The contention on the pfdat_lock spinlock would caused excessive cpu time being burned spinning. This problem would only be seen on PA8000 (PA2.0) systems. PHKL_8683: While on the ICS, sysmemunreserve() was deferencing a no longer valid uarea pointer although its caller kalloc() took care of specifying " uarea/proc checking". The fix was to honor the flag passed by the callers. SR: 5000716225 5003366971 5003367979 5003399188 Patch Files: /usr/conf/lib/libhp-ux.a(vm_kern.o) /usr/conf/lib/libhp-ux.a(vm_page.o) what(1) Output: /usr/conf/lib/libhp-ux.a(vm_kern.o): vm_kern.c $Date: 97/05/20 15:19:14 $ $Revision: 1. 9.98.5 $ PATCH_10.20 (PHKL_11164) /usr/conf/lib/libhp-ux.a(vm_page.o): vm_page.c $Date: 98/01/19 05:27:04 $ $Revision: 1. 91.98.13 $ PATCH_10.20 (PHKL_13768) cksum(1) Output: 1473182150 10480 /usr/conf/lib/libhp-ux.a(vm_kern.o) 3813709095 21236 /usr/conf/lib/libhp-ux.a(vm_page.o) Patch Conflicts: None Patch Dependencies: None Hardware Dependencies: None Other Dependencies: None Supersedes: PHKL_8683 PHKL_11055 PHKL_11164 Equivalent Patches: PHKL_13769: s800: 10.20 Patch Package Size: 90 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_13768 5a. For a standalone system, run swinstall to install the patch: swinstall -x autoreboot=true -x match_target=true \ -s /tmp/PHKL_13768.depot 5b. For a homogeneous NFS Diskless cluster run swcluster on the server to install the patch on the server and the clients: swcluster -i -b This will invoke swcluster in the interactive mode and force all clients to be shut down. WARNING: All cluster clients must be shut down prior to the patch installation. Installing the patch while the clients are booted is unsupported and can lead to serious problems. The swcluster command will invoke an swinstall session in which you must specify: alternate root path - default is /export/shared_root/OS_700 source depot path - /tmp/PHKL_13768.depot To complete the installation, select the patch by choosing "Actions -> Match What Target Has" and then "Actions -> Install" from the Menubar. 5c. For a heterogeneous NFS Diskless cluster: - run swinstall on the server as in step 5a to install the patch on the cluster server. - run swcluster on the server as in step 5b to install the patch on the cluster clients. By default swinstall will archive the original software in /var/adm/sw/patch/PHKL_13768. If you do not wish to retain a copy of the original software, you can create an empty file named /var/adm/sw/patch/PATCH_NOSAVE. Warning: If this file exists when a patch is installed, the patch cannot be deinstalled. Please be careful when using this feature. It is recommended that you move the PHKL_13768.text file to /var/adm/sw/patch for future reference. To put this patch on a magnetic tape and install from the tape drive, use the command: dd if=/tmp/PHKL_13768.depot of=/dev/rmt/0m bs=2k Special Installation Instructions: None