Red Hat Bugzilla – Bug 33271
MP table in the EBDA can be UNSAFE
Last modified: 2005-10-31 17:00:50 EST
From Bugzilla Helper: User-Agent: Mozilla/4.0 (compatible; MSIE 5.5; Windows 98; Win 9x 4.90) On an IBM Netfinity 5100 (865821Y model : 2 PIII-733, IBM ServeRaid 4L, 512Mo ECC) we have this following message at boot (kernel 2.2.17-14): "WARNING: MP table in the EBDA can be UNSAFE, contact linux- smp@vger.kernel.org if you experience SMP problems!" What is the potential risk? And how can we fix it ? Reproducible: Always Steps to Reproduce: 1.Every time at boot. 2. 3. We've changed the 2 PIII-733. We've changes the motherboard. [root@netfinity /root]# dmesg | more Linux version 2.2.17-14 (root@localhost) (gcc version egcs-2.91.66 19990314/Linux (egcs-1.1.2 release)) #1 SMP Wed Mar 21 14:32:28 CET 2001 Intel MultiProcessor Specification v1.4 Virtual Wire compatibility mode. OEM ID: IBM GNK Product ID: Teton SMP APIC at: 0xFEE00000 Processor #1 Pentium(tm) Pro APIC version 17 Processor #0 Pentium(tm) Pro APIC version 17 I/O APIC #14 Version 17 at 0xFEC00000. I/O APIC #15 Version 17 at 0xFEC01000. Processors: 2 WARNING: MP table in the EBDA can be UNSAFE, contact linux- smp@vger.kernel.org if you experience SMP problems! mapped APIC to ffffe000 (fee00000) mapped IOAPIC to ffffd000 (fec00000) mapped IOAPIC to ffffc000 (fec01000) Detected 731059 kHz processor. Console: colour VGA+ 80x25 Calibrating delay loop... 1458.18 BogoMIPS Memory: 516996k/524224k available (1032k kernel code, 432k reserved, 5708k data, 56k init, 0k bigmem ) Dentry hash table entries: 262144 (order 9, 2048k) Buffer cache hash table entries: 524288 (order 9, 2048k) Page cache hash table entries: 131072 (order 7, 512k) Enabling fast FPU save and restore... done. Enabling unmasked SIMD FPU exception support... done. Checking 386/387 coupling... OK, FPU using exception 16 error reporting. Checking 'hlt' instruction... OK. POSIX conformance testing by UNIFIX mtrr: v1.35a (19990819) Richard Gooch (rgooch@atnf.csiro.au) Intel machine check architecture supported. Intel machine check reporting enabled on CPU#1. per-CPU timeslice cutoff: 49.93 usecs. CPU1: Intel Pentium III (Coppermine) stepping 03 calibrating APIC timer ... ..... CPU clock speed is 730.9681 MHz. ..... system bus clock speed is 132.9032 MHz. Booting processor 0 eip 2000 Calibrating delay loop... 1461.45 BogoMIPS Intel machine check reporting enabled on CPU#0. OK. CPU0: Intel Pentium III (Coppermine) stepping 03 Total of 2 processors activated (2919.63 BogoMIPS). enabling symmetric IO mode... ...done. ENABLING IO-APIC IRQs init IO_APIC IRQs IO-APIC (apicid-pin) 14-0, 14-5, 15-0, 15-1, 15-2, 15-3, 15-5, 15-8, 15- 9, 15-12, 15-13, 15-14, 15- 15 not connected. ..MP-BIOS bug: 8254 timer not connected to IO-APIC ...trying to set up timer as ExtINT... .. (found pin 0) ... works. number of MP IRQ sources: 20. number of IO-APIC #14 registers: 16. number of IO-APIC #15 registers: 16. testing the IO APIC....................... IO APIC #14...... .... register #00:0E000000 ....... : physical APIC id: 0E .... register #01: 000F0011 ....... : max redirection entries: 000F ....... : IO APIC version: 0011 .... register #02: 00000000 ....... : arbitration: 00 .... IRQ redirection table: --cut
It is only a debugging message and may be safely ignored.
In fact, we have 3 netfinity 5100 with Redhat 6.2 and on only one of them we have the message "MP table in the EBDA can be UNSAFE". The most disturbing is that we have not this message on both other machines, so can we really ignore this message? What is its meaning? Thanks a lot.
It's the kernel detecting that the bios has put a critical piece of information in a memory-location that can be trashed by (older versions of) LILO. This is a bios bug (eg it violates a spec), however recent version of LILO are believed to have a workaround for this. If this workaround would fail to work, you would see a lot of crashes in SMP mode.