Created attachment 356100 [details] Patch which corrects in correct detection of vsmp under CONFIG_X86_VSMP Description of problem: Redhat provides kernel sources as a source rpm. The linux kernel sources can be extracted and optionally rebuilt with extra CONFIG options. One of the kernel patches applied has been partially pulled/ported causing code under CONFIG_X86_VSMP to be incorrect. While we understand that Redhat will not support recompiled kernels, the fact that upstream fixes have been partially pulled/pulled incorrectly results in in-correctness of code under CONFIG_X86_VSMP. The patch in question is: linux-2.6-x86_64-amd-8-socket-apicid-patches.patch 1. is_vsmp_box() will evaluate to return 0 even on a vSMPowered machine with CONFIG_X86_VSMP turned on 2. The is_vsmp_box() check is incorrect. This results in unexpected behavior when kernel is recompiled with CONFIG_X86_VSMP Version-Release number of selected component (if applicable): kernel-2.6.18-128.1.16.el5.src.rpm How reproducible: Always Steps to Reproduce: One would need a vSMPowered system to reproduce 1. Recompile sources extracted from the kernel source rpm by turning on CONFIG_X86_VSMP 2. Boot resulting kernel on vSMPowered system 3. You should see the following lines in the kernel dmesg: vSMP CTL: capabilities:0x00001fff control:0x000001b1 vSMP CTL: control set to:0x000001a1 If you don't, then the vsmp detection is incorrect. Actual results: The prints described above missing. Since this code path and prints actually set a 'fast' mode of operation on vSMPowered systems a kernel recompiled with CONFIG_X86_VSMP turned on works in slow mode rather than fast mode resulting in reduced performance. Expected results: Prints describe above to appear in dmesg Additional info: Here's a patch fixing the issue -- on top of RHEL sources. This points out bit rot on RHEL unsupported code path.
Created attachment 357179 [details] RHEL5 fix for this issue Updated to latest. P.
This request was evaluated by Red Hat Product Management for inclusion in a Red Hat Enterprise Linux maintenance release. Product Management has requested further review of this request by Red Hat Engineering, for potential inclusion in a Red Hat Enterprise Linux Update release for currently deployed products. This request is not yet committed for inclusion in an Update release.
@Kiran, We need to confirm that there is commitment to test for the resolution of this request during the RHEL 5.5 test phase, if it is accepted into the release. Please post a confirmation before Oct 16th, 2009, including the contact information for testing engineers.
(In reply to comment #5) > @Kiran, > > We need to confirm that there is commitment to test > for the resolution of this request during the RHEL 5.5 test > phase, if it is accepted into the release. > > Please post a confirmation before Oct 16th, 2009, > including the contact information for testing engineers. Yes, we are 100% committed to test and help Redhat in any possible way to resolve this issue. Josh Hunt <josh> is the designated engineer to test this fix/release. You can contact me as well. I can send you phone numbers in private if needed.
in kernel-2.6.18-170.el5 You can download this test kernel from http://people.redhat.com/dzickus/el5 Please do NOT transition this bugzilla state to VERIFIED until our QE team has sent specific instructions indicating when to do so. However feel free to provide a comment indicating that this fix has been verified.
I have recompiled your kernel source package with CONFIG_X86_VSMP and it resolves the issue. Here is a dmesg snippet: <dmesg> sizeof(buffer_head)=96 bytes sizeof(skbuff)=248 bytes migration_cost=97,10082,208756 checking if image is initramfs... it is Freeing initrd memory: 2478k freed vSMP CTL: capabilities:0x00001fff control:0x000001b1 vSMP CTL: control set to:0x000001a1 NET: Registered protocol family 16 ACPI: bus type pci registered PCI: Using configuration type 1 </dmesg> Thanks!
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-2010-0178.html