Hello, Here are some more recommendations for backports from unstable to 4.1, based upon the XenServer patch queue. Patches with a * do not apply cleanly, and I can provide patches if you wish to backport them. The rest should apply cleanly (unless I have missed a dependency on other patches we have in our queue, which case I apologize) The following are bugfixes: * 23101:dd386a4b6595 # x86: fix tboot after c/s 23013:65d26504e843 (ACPI cleanup) 23686:7c39a2c0d870 # pygrub: fix solaris kernel issue 23781:0849b0e59e24 # pm: don''t truncate processors'' ACPI IDs to 8 bits 23813:5535d7ce2673 # xen, vtd: Fix device check for devices behind PCIe-to-PCI bridges 25162:478bec603d3d # xen: Fix schedule()''s grabbing of the schedule lock 25487:baa85434d0ec # x86/mm: fix mod_l1_entry() return value when encountering r/o MMIO page 25489:cc46bd403bc4 # x86/PCI: fix guest_io_read() when pci_cfg_ok() denies access The following are not strictly bugfixes, but I feel they should be considered: * 23726:fd97ca086df6 # x86: update Intel CPUID masking code to latest spec 23731:48f72b389b04 # x86: add change missing in c/s 23726:fd97ca086df6 24261:64088ba60263 # x86/cpuidle: add Westmere-EX support to hw residencies reading logic 24970:33980e36597b # x86/cpuidle: deny access to the I/O port used for EM_SYSIO * 25482:34c725807d21 # SVM: fix performance decrease with asid assignment -- Andrew Cooper - Dom0 Kernel Engineer, Citrix XenServer T: +44 (0)1223 225 900, http://www.citrix.com
Andrew, I applied all the below, except 23101:dd386a4b6595 as the dependent patch (23013:65d26504e843) is not in OSS 4.1. -- Keir On 06/07/2012 16:36, "Andrew Cooper" <andrew.cooper3@citrix.com> wrote:> Hello, > > Here are some more recommendations for backports from unstable to 4.1, > based upon the XenServer patch queue. Patches with a * do not apply > cleanly, and I can provide patches if you wish to backport them. The > rest should apply cleanly (unless I have missed a dependency on other > patches we have in our queue, which case I apologize) > > The following are bugfixes: > > * 23101:dd386a4b6595 # x86: fix tboot after c/s 23013:65d26504e843 (ACPI > cleanup) > 23686:7c39a2c0d870 # pygrub: fix solaris kernel issue > 23781:0849b0e59e24 # pm: don''t truncate processors'' ACPI IDs to 8 bits > 23813:5535d7ce2673 # xen, vtd: Fix device check for devices behind > PCIe-to-PCI bridges > 25162:478bec603d3d # xen: Fix schedule()''s grabbing of the schedule lock > 25487:baa85434d0ec # x86/mm: fix mod_l1_entry() return value when > encountering r/o MMIO page > 25489:cc46bd403bc4 # x86/PCI: fix guest_io_read() when pci_cfg_ok() > denies access > > > The following are not strictly bugfixes, but I feel they should be > considered: > > * 23726:fd97ca086df6 # x86: update Intel CPUID masking code to latest spec > 23731:48f72b389b04 # x86: add change missing in c/s 23726:fd97ca086df6 > 24261:64088ba60263 # x86/cpuidle: add Westmere-EX support to hw > residencies reading logic > 24970:33980e36597b # x86/cpuidle: deny access to the I/O port used for > EM_SYSIO > * 25482:34c725807d21 # SVM: fix performance decrease with asid assignment
On 09/07/12 10:32, Keir Fraser wrote:> Andrew, > > I applied all the below, except 23101:dd386a4b6595 as the dependent patch > (23013:65d26504e843) is not in OSS 4.1. > > -- KeirThanks. Sorry for that patch - it appears it is actually a minimal mix of 23101 and 23013 to try and make TXTboot work, rather than a strict backport of 23101. Is there an eta for 4.1.3 ? ~Andrew> > On 06/07/2012 16:36, "Andrew Cooper" <andrew.cooper3@citrix.com> wrote: > >> Hello, >> >> Here are some more recommendations for backports from unstable to 4.1, >> based upon the XenServer patch queue. Patches with a * do not apply >> cleanly, and I can provide patches if you wish to backport them. The >> rest should apply cleanly (unless I have missed a dependency on other >> patches we have in our queue, which case I apologize) >> >> The following are bugfixes: >> >> * 23101:dd386a4b6595 # x86: fix tboot after c/s 23013:65d26504e843 (ACPI >> cleanup) >> 23686:7c39a2c0d870 # pygrub: fix solaris kernel issue >> 23781:0849b0e59e24 # pm: don''t truncate processors'' ACPI IDs to 8 bits >> 23813:5535d7ce2673 # xen, vtd: Fix device check for devices behind >> PCIe-to-PCI bridges >> 25162:478bec603d3d # xen: Fix schedule()''s grabbing of the schedule lock >> 25487:baa85434d0ec # x86/mm: fix mod_l1_entry() return value when >> encountering r/o MMIO page >> 25489:cc46bd403bc4 # x86/PCI: fix guest_io_read() when pci_cfg_ok() >> denies access >> >> >> The following are not strictly bugfixes, but I feel they should be >> considered: >> >> * 23726:fd97ca086df6 # x86: update Intel CPUID masking code to latest spec >> 23731:48f72b389b04 # x86: add change missing in c/s 23726:fd97ca086df6 >> 24261:64088ba60263 # x86/cpuidle: add Westmere-EX support to hw >> residencies reading logic >> 24970:33980e36597b # x86/cpuidle: deny access to the I/O port used for >> EM_SYSIO >> * 25482:34c725807d21 # SVM: fix performance decrease with asid assignment >-- Andrew Cooper - Dom0 Kernel Engineer, Citrix XenServer T: +44 (0)1223 225 900, http://www.citrix.com
>>> On 09.07.12 at 11:32, Keir Fraser <keir@xen.org> wrote: > I applied all the below, except 23101:dd386a4b6595 as the dependent patch > (23013:65d26504e843) is not in OSS 4.1.Please revert the backport of 25487:baa85434d0ec - it is wrong in 4.1-testing (where the function still returns a boolean-like value). Jan> On 06/07/2012 16:36, "Andrew Cooper" <andrew.cooper3@citrix.com> wrote: > >> Hello, >> >> Here are some more recommendations for backports from unstable to 4.1, >> based upon the XenServer patch queue. Patches with a * do not apply >> cleanly, and I can provide patches if you wish to backport them. The >> rest should apply cleanly (unless I have missed a dependency on other >> patches we have in our queue, which case I apologize) >> >> The following are bugfixes: >> >> * 23101:dd386a4b6595 # x86: fix tboot after c/s 23013:65d26504e843 (ACPI >> cleanup) >> 23686:7c39a2c0d870 # pygrub: fix solaris kernel issue >> 23781:0849b0e59e24 # pm: don''t truncate processors'' ACPI IDs to 8 bits >> 23813:5535d7ce2673 # xen, vtd: Fix device check for devices behind >> PCIe-to-PCI bridges >> 25162:478bec603d3d # xen: Fix schedule()''s grabbing of the schedule lock >> 25487:baa85434d0ec # x86/mm: fix mod_l1_entry() return value when >> encountering r/o MMIO page >> 25489:cc46bd403bc4 # x86/PCI: fix guest_io_read() when pci_cfg_ok() >> denies access >> >> >> The following are not strictly bugfixes, but I feel they should be >> considered: >> >> * 23726:fd97ca086df6 # x86: update Intel CPUID masking code to latest spec >> 23731:48f72b389b04 # x86: add change missing in c/s 23726:fd97ca086df6 >> 24261:64088ba60263 # x86/cpuidle: add Westmere-EX support to hw >> residencies reading logic >> 24970:33980e36597b # x86/cpuidle: deny access to the I/O port used for >> EM_SYSIO >> * 25482:34c725807d21 # SVM: fix performance decrease with asid assignment > > > > _______________________________________________ > Xen-devel mailing list > Xen-devel@lists.xen.org > http://lists.xen.org/xen-devel