Description of problem: I have seen two HP IA64 machines, hp-matterhorn1.rhts.bos.redhat.com hp-olympia1.rhts.bos.redhat.com failed to read /proc/acpi/dsdt: No such device. There is no such problem on RHEL5-U2 Kernel though. Version-Release number of selected component (if applicable): Both 4.7 and 4.6.z Kernel, kernel-2.6.9-78.EL kernel-2.6.9-67.0.20.EL How reproducible: Always Steps to Reproduce: cat /proc/acpi/dsdt Actual results: No such device Expected results: Display some content.
What appears to be happening is that the ACPI code attempts to reserve 128K+ memory for the dsdt table. a) I don't know if the calculation for the dsdt is correct in RHEL4, and b) I'm going to load RHEL5 (which works) to see if the same size reservation is made. I asked lwoodman (in private email) about this and he did say that the upper limit of kmalloc in RHEL4 is 128K. P.
>a) I don't know if the calculation for the dsdt is correct in RHEL4, and b) I'm >going to load RHEL5 (which works) to see if the same size reservation is made. dsdt size is 0x226a5 on both RHEL4 & RHEL5. The problem here is that RHEL4 has a 128K limit on kmalloc. The code in question should use alloc_pages() to avoid this bug. P.
After thinking about this, I'm curious as to why the dsdt size is 0x226a5. Maybe that's a BIOS/ACPI Table error? To make matters worse, it is alloc'd many times during a 'cat /proc/acpi/dsdt' ... P.
Doug -- comment #3? P.
Created attachment 312765 [details] ia64 workaround for this issue
Adding Alex and Tony -- HP engineers. Do you know if the dsdt is supposed to be that large on these systems? P.
Taking out of NEEDINFO. I downloaded acpidump, compiled, and ran it on matterhorn1. AFAICT, it looks like the table is valid. P.
Prarit, Can you find the DSDT in the ACPI dump? If so, bytes 7:4 give the length of the table. That is the amount of space that should be allocated.
Hi Tony, yup I found that using the acpidump utility. The amount there corresponds to what is being allocated -- and the table looks fine, IMO. So I posted the patch to RHKL. P.
Prarit, Since you've posted the patch, I assume that this problem has been solved and we can close this BZ. Correct?
Hi Tony, This BZ is in POST -- where it should be for now. dzickus will move it into MODIFIED, and eventually it will be CLOSED NEXTRELEASE. ie) leave it in POST. P.
Thanks, Prarit. Just doing a little housekeeping. :)
Updating PM score.
It might be interesting to check this bug, bug 464406 - [4.7.z] SGI Altix ACPI Problem . https://bugzilla.redhat.com/show_bug.cgi?id=464406
The machine hp-rx8640-03.rhts.bos.redhat.com is also affected.
Committed in 78.29.EL . RPMS are available at http://people.redhat.com/vgoyal/rhel4/
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on therefore solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHSA-2009-1024.html