Comment 38 Alok Kataria 2008-10-11 23:01:53 EDT (In reply to comment #36) > Created an attachment (id=320098) [details] > vmware tsc calibration, but more generic to accomodate other hypervisors > > Why is that? If upstream goes with that (eg. -mm or Ingo's x86 tree), the same should be acceptable in a RHEL update. Otherwise only other patch that will make sense upstream is the 3rd patch of this series.
Thanks, Alok P.S. : I am on vacation till January and will have intermittent access to mails, so please be aware that responses might be delayed. VM on VMware ESX Server 3.5.0 build 84374 - 32-bit or 64-bit? 64-bit - What processor? i wanted to know what release are these patches going to be merged in. Presuming that is the case, I've marked the "fix 64 bit soft lockup warnings on vmware" as obsolete here in the BZ.
First up, ABI or the Application Binary Interface is a low level (binary level) interface exposed by some application. I have no objections to that. If you still think that we should add some checks to thwart the therotical problem too, then may be we can do the DMI stuff. The --without kabichk command line option turns off the checking and will allow your kernel build to complete.I'm assuming you just want the base kernel?
Attached patch provides preset lpj support for KVM hypervisor. When running on a hypervisor, however, it is possible to hit this scenario even when there is no real bug. Thanks, Alok Comment 52 Glauber Costa 2008-10-20 08:54:50 EDT Rik, AFAICS, these patches do not solve the problem of "timekeeping" for KVM, only for vmware. Here's what I *meant* to say: Just to be clear, the latest patch that VMware attached (named "Disable softlockup with the TSC based algorithm under VMware") is supposed to replace patch
If you mean the functionality to detect whether we are running on VMware, yes i plan to push patches for this upstream but that is going to be a lenghty process What does uname -r return on your system? Instead, I went with the upstream solution (namely git hash > > 3da757daf86e498872855f0b5e101f763ba79499), and use the "lpj_fine" variable to > > basically do the same. > > Okay, but this patch Skipping the PIT routing check comes for free (which is also a problem under KVM).
This isn't setup in the upstream kernel, but is > > in this patchset. You signed out in another tab or window. So there is no pressing need for getting that patch too atleast for the 2.6.27 tree. Now, this one may be very different because > of the constant 1000HZ RHEL-5 kernel vs.
Thank you for the summary. https://www.ibm.com/developerworks/forums/thread.jspa?threadID=355717 Top AlanBartlett Forum Moderator Posts: 9299 Joined: 2007/10/22 11:30:09 Location: ~/Earth/UK/England/Suffolk Contact: Contact AlanBartlett Website Re: Problem REcompiling the kernel optimized for k8 Quote Postby AlanBartlett » 2008/06/10 00:53:47 D'oh. :oops: This is not odd because some variable called CONFIG_MODVERSIONS had not been set during the kernel build process thus disabling generation of checksums for the exported symbols. make: Entering directory `/home/GIT/RHEL7.2/rhel7/redhat' *** ERROR - ABI BREAKAGE WAS DETECTED *** The following symbols have been changed (this will cause an ABI breakage): dev_get_stats invalidate_bdev scsi_host_alloc hci_register_dev dev_addr_add blk_queue_merge_bvec vmalloc_to_page
TEST is "None". Again, thanks for taking the time with this; hopefully we won't have to go through too much more to get it done. Please let us know how things progress. Comment 17 Rik van Riel 2008-09-25 19:48:30 EDT I understand that you would like to see the patches in RHEL 5.3, however it is very late in the development cycle to
Thanks, Alok Comment 15 Rik van Riel 2008-09-25 18:03:58 EDT These patches look reasonable to me. When Alan comes online, he will provide more detailed explanations, I'm sure.which thread?Sorry, I forgot to paste the thread URL. fix lockup warnings for 64 bit (forward ported to latest RHEL 5) https://bugzilla.redhat.com/attachment.cgi?id=319719 7. [BUGFIX] Do not overcompensate for divider= option https://bugzilla.redhat.com/attachment.cgi?id=319817 8. Does > upstream need this?
Thanks, Alok Comment 43 Marcelo Tosatti 2008-10-15 16:46:03 EDT Created attachment 320485 [details] KVM support for x86_hyper_vendor and preset lpj Alok, My bad - sorry. Ah, OK, great. > > > > 3) Also in the second patch, during time initialization in both i386 and > > x86_64, in the local patch we set "preset_lpj" to This has been an entirely new and exciting experience for me since I became a Red Hat associate in May last year.
Home | New | Search | [?] | Reports | Requests | Help | NewAccount | Log In [x] | Forgot Password Login: [x] | Report Bugzilla Bug Legal Red Hat Hi Rik, What is the status on these patches ? That is, x86_64 already has the ability to use a TSC timesource now, yes? Hi Rik, While testing these patches we found a bug with the 32 bit kernels.
Comment 122 RHEL Product and Program Management 2009-02-16 10:07:47 EST Updating PM score. The solution is to change the kernel's timekeeping algorithm to keep time based on how much time has elapsed according to a time counter rather than by counting interrupts. CentOS The Community ENTerprise Operating System Skip to content Search Advanced search Quick links Unanswered posts Active topics Search The team FAQ Login Register Board index CentOS 5 CentOS 5 - Is that intented?
Patches to improve timekeeping for RHEL kernels running under VMware. If the messages are innocuous, they shouldn't be appearing on the console at all. OK, good. I merged the backdoor fix into the "vm tsc calibration" patch.
Please see my replied inline. (In reply to comment #3) > Poking a random IO port to detect whether or not we are running on VMware is > really not the Comment 97 Rik van Riel 2008-12-11 16:18:07 EST Thank you for that additional patch, Alok. check-kabi [ -k Module.kabi ] [ -s Module.symvers ] """ if __name__ == "__main__": symvers_file = "" kabi_file = "" opts, args = getopt.getopt(sys.argv[1:], 'hk:s:') for o, v in opts: if The Error - Abi Breakage Was Detected error may be caused by windows system files damage.
Comment 31 Alok Kataria 2008-10-08 12:44:55 EDT (In reply to comment #30) > *** ERROR - ABI BREAKAGE WAS DETECTED *** > > The following symbols have been changed (this will So that now its only called when running under VMware hypervisor. Anyway, we have a few options: 1) Disable the softlockup warning completely when running on hypervisors because the warning will more often than not be a false positive. 2) Try to This fix should be backported to the RHEL kernel for these patches to work correctly, else we cannot relaibly depend on the value of x86_hyper_vendor.