* preparations for 4.11.2 @ 2019-03-18 16:13 Jan Beulich 2019-04-26 11:59 ` [Xen-devel] " Andrew Cooper 0 siblings, 1 reply; 31+ messages in thread From: Jan Beulich @ 2019-03-18 16:13 UTC (permalink / raw) To: xen-devel Cc: Anthony Perard, Ian Jackson, Stefano Stabellini, Wei Liu, Lars Kurth All, the release is due by the end of the month, but will likely don't make it before early April. Please point out backports you find missing from the respective staging branch, but which you consider relevant. The one commit I've queued already on top of what was just pushed is 22e2f8dddf x86/e820: fix build with gcc9 Jan _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: preparations for 4.11.2 @ 2019-04-26 11:59 ` Andrew Cooper 0 siblings, 0 replies; 31+ messages in thread From: Andrew Cooper @ 2019-04-26 11:59 UTC (permalink / raw) To: xen-devel; +Cc: Ian Jackson On 18/03/2019 16:13, Jan Beulich wrote: > All, > > the release is due by the end of the month, but will likely don't make > it before early April. Please point out backports you find missing from > the respective staging branch, but which you consider relevant. The > one commit I've queued already on top of what was just pushed is > > 22e2f8dddf x86/e820: fix build with gcc9 ffb60a58df48419c1f2607cd3cc919fa2bfc9c2d "tools/misc/xenpm: fix getting info when some CPUs are offline" for 4.11 and earlier. ~Andrew _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [Xen-devel] preparations for 4.11.2 @ 2019-04-26 11:59 ` Andrew Cooper 0 siblings, 0 replies; 31+ messages in thread From: Andrew Cooper @ 2019-04-26 11:59 UTC (permalink / raw) To: xen-devel; +Cc: Ian Jackson On 18/03/2019 16:13, Jan Beulich wrote: > All, > > the release is due by the end of the month, but will likely don't make > it before early April. Please point out backports you find missing from > the respective staging branch, but which you consider relevant. The > one commit I've queued already on top of what was just pushed is > > 22e2f8dddf x86/e820: fix build with gcc9 ffb60a58df48419c1f2607cd3cc919fa2bfc9c2d "tools/misc/xenpm: fix getting info when some CPUs are offline" for 4.11 and earlier. ~Andrew _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: preparations for 4.11.2 @ 2019-04-26 12:02 ` Andrew Cooper 0 siblings, 0 replies; 31+ messages in thread From: Andrew Cooper @ 2019-04-26 12:02 UTC (permalink / raw) To: xen-devel, Ian Jackson On 26/04/2019 12:59, Andrew Cooper wrote: > On 18/03/2019 16:13, Jan Beulich wrote: >> All, >> >> the release is due by the end of the month, but will likely don't make >> it before early April. Please point out backports you find missing from >> the respective staging branch, but which you consider relevant. The >> one commit I've queued already on top of what was just pushed is >> >> 22e2f8dddf x86/e820: fix build with gcc9 > ffb60a58df48419c1f2607cd3cc919fa2bfc9c2d "tools/misc/xenpm: fix getting > info when some CPUs are offline" for 4.11 and earlier. Oh, and 677e64dbe315343620c3b266e9eb16623b118038 "tools/ocaml: Dup2 /dev/null to stdin in daemonize()" again for 4.12 and earlier. ~Andrew _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [Xen-devel] preparations for 4.11.2 @ 2019-04-26 12:02 ` Andrew Cooper 0 siblings, 0 replies; 31+ messages in thread From: Andrew Cooper @ 2019-04-26 12:02 UTC (permalink / raw) To: xen-devel, Ian Jackson On 26/04/2019 12:59, Andrew Cooper wrote: > On 18/03/2019 16:13, Jan Beulich wrote: >> All, >> >> the release is due by the end of the month, but will likely don't make >> it before early April. Please point out backports you find missing from >> the respective staging branch, but which you consider relevant. The >> one commit I've queued already on top of what was just pushed is >> >> 22e2f8dddf x86/e820: fix build with gcc9 > ffb60a58df48419c1f2607cd3cc919fa2bfc9c2d "tools/misc/xenpm: fix getting > info when some CPUs are offline" for 4.11 and earlier. Oh, and 677e64dbe315343620c3b266e9eb16623b118038 "tools/ocaml: Dup2 /dev/null to stdin in daemonize()" again for 4.12 and earlier. ~Andrew _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: preparations for 4.11.2 @ 2019-05-16 15:11 ` Andrew Cooper 0 siblings, 0 replies; 31+ messages in thread From: Andrew Cooper @ 2019-05-16 15:11 UTC (permalink / raw) To: xen-devel, Ian Jackson On 26/04/2019 13:02, Andrew Cooper wrote: > On 26/04/2019 12:59, Andrew Cooper wrote: >> On 18/03/2019 16:13, Jan Beulich wrote: >>> All, >>> >>> the release is due by the end of the month, but will likely don't make >>> it before early April. Please point out backports you find missing from >>> the respective staging branch, but which you consider relevant. The >>> one commit I've queued already on top of what was just pushed is >>> >>> 22e2f8dddf x86/e820: fix build with gcc9 >> ffb60a58df48419c1f2607cd3cc919fa2bfc9c2d "tools/misc/xenpm: fix getting >> info when some CPUs are offline" for 4.11 and earlier. > Oh, and 677e64dbe315343620c3b266e9eb16623b118038 "tools/ocaml: Dup2 > /dev/null to stdin in daemonize()" again for 4.12 and earlier. In addition, 2ec5339ec921 "tools/libxl: correct vcpu affinity output with sparse physical cpu map" 129025fe3093 "oxenstored: Don't re-open a xenctrl handle for every domain introduction" 7b20a865bc10 "tools/ocaml: Release the global lock before invoking block syscalls" c393b64dcee6 "tools/libxc: Fix issues with libxc and Xen having different featureset lengths" 82855aba5bf9 "tools/libxc: Fix error handling in get_cpuid_domain_info()" 48dab9767d2e "tools/xl: use libxl_domain_info to get domain type for vcpu-pin" 365aabb6e502 "tools/libxendevicemodel: add xendevicemodel_modified_memory_bulk to map" is possibly a candidate, but is also complicated by the stable SONAME. It is perhaps easiest to ignore, seeing as the issue has already gone unnoticed for 2 years. ~Andrew _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [Xen-devel] preparations for 4.11.2 @ 2019-05-16 15:11 ` Andrew Cooper 0 siblings, 0 replies; 31+ messages in thread From: Andrew Cooper @ 2019-05-16 15:11 UTC (permalink / raw) To: xen-devel, Ian Jackson On 26/04/2019 13:02, Andrew Cooper wrote: > On 26/04/2019 12:59, Andrew Cooper wrote: >> On 18/03/2019 16:13, Jan Beulich wrote: >>> All, >>> >>> the release is due by the end of the month, but will likely don't make >>> it before early April. Please point out backports you find missing from >>> the respective staging branch, but which you consider relevant. The >>> one commit I've queued already on top of what was just pushed is >>> >>> 22e2f8dddf x86/e820: fix build with gcc9 >> ffb60a58df48419c1f2607cd3cc919fa2bfc9c2d "tools/misc/xenpm: fix getting >> info when some CPUs are offline" for 4.11 and earlier. > Oh, and 677e64dbe315343620c3b266e9eb16623b118038 "tools/ocaml: Dup2 > /dev/null to stdin in daemonize()" again for 4.12 and earlier. In addition, 2ec5339ec921 "tools/libxl: correct vcpu affinity output with sparse physical cpu map" 129025fe3093 "oxenstored: Don't re-open a xenctrl handle for every domain introduction" 7b20a865bc10 "tools/ocaml: Release the global lock before invoking block syscalls" c393b64dcee6 "tools/libxc: Fix issues with libxc and Xen having different featureset lengths" 82855aba5bf9 "tools/libxc: Fix error handling in get_cpuid_domain_info()" 48dab9767d2e "tools/xl: use libxl_domain_info to get domain type for vcpu-pin" 365aabb6e502 "tools/libxendevicemodel: add xendevicemodel_modified_memory_bulk to map" is possibly a candidate, but is also complicated by the stable SONAME. It is perhaps easiest to ignore, seeing as the issue has already gone unnoticed for 2 years. ~Andrew _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: preparations for 4.11.2 @ 2019-05-16 15:20 ` Jan Beulich 0 siblings, 0 replies; 31+ messages in thread From: Jan Beulich @ 2019-05-16 15:20 UTC (permalink / raw) To: Andrew Cooper, Ian Jackson; +Cc: xen-devel >>> On 16.05.19 at 17:11, <andrew.cooper3@citrix.com> wrote: > On 26/04/2019 13:02, Andrew Cooper wrote: >> On 26/04/2019 12:59, Andrew Cooper wrote: >>> On 18/03/2019 16:13, Jan Beulich wrote: >>>> All, >>>> >>>> the release is due by the end of the month, but will likely don't make >>>> it before early April. Please point out backports you find missing from >>>> the respective staging branch, but which you consider relevant. The >>>> one commit I've queued already on top of what was just pushed is >>>> >>>> 22e2f8dddf x86/e820: fix build with gcc9 >>> ffb60a58df48419c1f2607cd3cc919fa2bfc9c2d "tools/misc/xenpm: fix getting >>> info when some CPUs are offline" for 4.11 and earlier. >> Oh, and 677e64dbe315343620c3b266e9eb16623b118038 "tools/ocaml: Dup2 >> /dev/null to stdin in daemonize()" again for 4.12 and earlier. > > In addition, > > 2ec5339ec921 "tools/libxl: correct vcpu affinity output with sparse > physical cpu map" > 129025fe3093 "oxenstored: Don't re-open a xenctrl handle for every > domain introduction" > 7b20a865bc10 "tools/ocaml: Release the global lock before invoking block > syscalls" > c393b64dcee6 "tools/libxc: Fix issues with libxc and Xen having > different featureset lengths" > 82855aba5bf9 "tools/libxc: Fix error handling in get_cpuid_domain_info()" > 48dab9767d2e "tools/xl: use libxl_domain_info to get domain type for > vcpu-pin" > > 365aabb6e502 "tools/libxendevicemodel: add > xendevicemodel_modified_memory_bulk to map" is possibly a candidate, but > is also complicated by the stable SONAME. It is perhaps easiest to > ignore, seeing as the issue has already gone unnoticed for 2 years. Unless these are really urgent to put in, I'd like them to be deferred to 4.11.3. With XSA-297 out we've already started the (leaf tree) tagging process, so I was really hoping to get this much delayed release out rather sooner than later. Jan _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [Xen-devel] preparations for 4.11.2 @ 2019-05-16 15:20 ` Jan Beulich 0 siblings, 0 replies; 31+ messages in thread From: Jan Beulich @ 2019-05-16 15:20 UTC (permalink / raw) To: Andrew Cooper, Ian Jackson; +Cc: xen-devel >>> On 16.05.19 at 17:11, <andrew.cooper3@citrix.com> wrote: > On 26/04/2019 13:02, Andrew Cooper wrote: >> On 26/04/2019 12:59, Andrew Cooper wrote: >>> On 18/03/2019 16:13, Jan Beulich wrote: >>>> All, >>>> >>>> the release is due by the end of the month, but will likely don't make >>>> it before early April. Please point out backports you find missing from >>>> the respective staging branch, but which you consider relevant. The >>>> one commit I've queued already on top of what was just pushed is >>>> >>>> 22e2f8dddf x86/e820: fix build with gcc9 >>> ffb60a58df48419c1f2607cd3cc919fa2bfc9c2d "tools/misc/xenpm: fix getting >>> info when some CPUs are offline" for 4.11 and earlier. >> Oh, and 677e64dbe315343620c3b266e9eb16623b118038 "tools/ocaml: Dup2 >> /dev/null to stdin in daemonize()" again for 4.12 and earlier. > > In addition, > > 2ec5339ec921 "tools/libxl: correct vcpu affinity output with sparse > physical cpu map" > 129025fe3093 "oxenstored: Don't re-open a xenctrl handle for every > domain introduction" > 7b20a865bc10 "tools/ocaml: Release the global lock before invoking block > syscalls" > c393b64dcee6 "tools/libxc: Fix issues with libxc and Xen having > different featureset lengths" > 82855aba5bf9 "tools/libxc: Fix error handling in get_cpuid_domain_info()" > 48dab9767d2e "tools/xl: use libxl_domain_info to get domain type for > vcpu-pin" > > 365aabb6e502 "tools/libxendevicemodel: add > xendevicemodel_modified_memory_bulk to map" is possibly a candidate, but > is also complicated by the stable SONAME. It is perhaps easiest to > ignore, seeing as the issue has already gone unnoticed for 2 years. Unless these are really urgent to put in, I'd like them to be deferred to 4.11.3. With XSA-297 out we've already started the (leaf tree) tagging process, so I was really hoping to get this much delayed release out rather sooner than later. Jan _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: preparations for 4.11.2 @ 2019-05-16 15:23 ` Andrew Cooper 0 siblings, 0 replies; 31+ messages in thread From: Andrew Cooper @ 2019-05-16 15:23 UTC (permalink / raw) To: Jan Beulich, Ian Jackson; +Cc: xen-devel On 16/05/2019 16:20, Jan Beulich wrote: >>>> On 16.05.19 at 17:11, <andrew.cooper3@citrix.com> wrote: >> On 26/04/2019 13:02, Andrew Cooper wrote: >>> On 26/04/2019 12:59, Andrew Cooper wrote: >>>> On 18/03/2019 16:13, Jan Beulich wrote: >>>>> All, >>>>> >>>>> the release is due by the end of the month, but will likely don't make >>>>> it before early April. Please point out backports you find missing from >>>>> the respective staging branch, but which you consider relevant. The >>>>> one commit I've queued already on top of what was just pushed is >>>>> >>>>> 22e2f8dddf x86/e820: fix build with gcc9 >>>> ffb60a58df48419c1f2607cd3cc919fa2bfc9c2d "tools/misc/xenpm: fix getting >>>> info when some CPUs are offline" for 4.11 and earlier. >>> Oh, and 677e64dbe315343620c3b266e9eb16623b118038 "tools/ocaml: Dup2 >>> /dev/null to stdin in daemonize()" again for 4.12 and earlier. >> In addition, >> >> 2ec5339ec921 "tools/libxl: correct vcpu affinity output with sparse >> physical cpu map" >> 129025fe3093 "oxenstored: Don't re-open a xenctrl handle for every >> domain introduction" >> 7b20a865bc10 "tools/ocaml: Release the global lock before invoking block >> syscalls" >> c393b64dcee6 "tools/libxc: Fix issues with libxc and Xen having >> different featureset lengths" >> 82855aba5bf9 "tools/libxc: Fix error handling in get_cpuid_domain_info()" >> 48dab9767d2e "tools/xl: use libxl_domain_info to get domain type for >> vcpu-pin" >> >> 365aabb6e502 "tools/libxendevicemodel: add >> xendevicemodel_modified_memory_bulk to map" is possibly a candidate, but >> is also complicated by the stable SONAME. It is perhaps easiest to >> ignore, seeing as the issue has already gone unnoticed for 2 years. > Unless these are really urgent to put in, I'd like them to be deferred > to 4.11.3. With XSA-297 out we've already started the (leaf tree) > tagging process, so I was really hoping to get this much delayed > release out rather sooner than later. At a minimum, ffb60a58df4 and ffb60a58df4 need backporting, because they are breakage of tools caused by our recommendation to turn off SMT in recent XSAs. Everything else can be deferred if necessary. ~Andrew _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [Xen-devel] preparations for 4.11.2 @ 2019-05-16 15:23 ` Andrew Cooper 0 siblings, 0 replies; 31+ messages in thread From: Andrew Cooper @ 2019-05-16 15:23 UTC (permalink / raw) To: Jan Beulich, Ian Jackson; +Cc: xen-devel On 16/05/2019 16:20, Jan Beulich wrote: >>>> On 16.05.19 at 17:11, <andrew.cooper3@citrix.com> wrote: >> On 26/04/2019 13:02, Andrew Cooper wrote: >>> On 26/04/2019 12:59, Andrew Cooper wrote: >>>> On 18/03/2019 16:13, Jan Beulich wrote: >>>>> All, >>>>> >>>>> the release is due by the end of the month, but will likely don't make >>>>> it before early April. Please point out backports you find missing from >>>>> the respective staging branch, but which you consider relevant. The >>>>> one commit I've queued already on top of what was just pushed is >>>>> >>>>> 22e2f8dddf x86/e820: fix build with gcc9 >>>> ffb60a58df48419c1f2607cd3cc919fa2bfc9c2d "tools/misc/xenpm: fix getting >>>> info when some CPUs are offline" for 4.11 and earlier. >>> Oh, and 677e64dbe315343620c3b266e9eb16623b118038 "tools/ocaml: Dup2 >>> /dev/null to stdin in daemonize()" again for 4.12 and earlier. >> In addition, >> >> 2ec5339ec921 "tools/libxl: correct vcpu affinity output with sparse >> physical cpu map" >> 129025fe3093 "oxenstored: Don't re-open a xenctrl handle for every >> domain introduction" >> 7b20a865bc10 "tools/ocaml: Release the global lock before invoking block >> syscalls" >> c393b64dcee6 "tools/libxc: Fix issues with libxc and Xen having >> different featureset lengths" >> 82855aba5bf9 "tools/libxc: Fix error handling in get_cpuid_domain_info()" >> 48dab9767d2e "tools/xl: use libxl_domain_info to get domain type for >> vcpu-pin" >> >> 365aabb6e502 "tools/libxendevicemodel: add >> xendevicemodel_modified_memory_bulk to map" is possibly a candidate, but >> is also complicated by the stable SONAME. It is perhaps easiest to >> ignore, seeing as the issue has already gone unnoticed for 2 years. > Unless these are really urgent to put in, I'd like them to be deferred > to 4.11.3. With XSA-297 out we've already started the (leaf tree) > tagging process, so I was really hoping to get this much delayed > release out rather sooner than later. At a minimum, ffb60a58df4 and ffb60a58df4 need backporting, because they are breakage of tools caused by our recommendation to turn off SMT in recent XSAs. Everything else can be deferred if necessary. ~Andrew _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: preparations for 4.11.2 @ 2019-05-16 15:29 ` Jan Beulich 0 siblings, 0 replies; 31+ messages in thread From: Jan Beulich @ 2019-05-16 15:29 UTC (permalink / raw) To: Andrew Cooper, Ian Jackson; +Cc: xen-devel >>> On 16.05.19 at 17:23, <andrew.cooper3@citrix.com> wrote: > On 16/05/2019 16:20, Jan Beulich wrote: >>>>> On 16.05.19 at 17:11, <andrew.cooper3@citrix.com> wrote: >>> On 26/04/2019 13:02, Andrew Cooper wrote: >>>> On 26/04/2019 12:59, Andrew Cooper wrote: >>>>> On 18/03/2019 16:13, Jan Beulich wrote: >>>>>> All, >>>>>> >>>>>> the release is due by the end of the month, but will likely don't make >>>>>> it before early April. Please point out backports you find missing from >>>>>> the respective staging branch, but which you consider relevant. The >>>>>> one commit I've queued already on top of what was just pushed is >>>>>> >>>>>> 22e2f8dddf x86/e820: fix build with gcc9 >>>>> ffb60a58df48419c1f2607cd3cc919fa2bfc9c2d "tools/misc/xenpm: fix getting >>>>> info when some CPUs are offline" for 4.11 and earlier. >>>> Oh, and 677e64dbe315343620c3b266e9eb16623b118038 "tools/ocaml: Dup2 >>>> /dev/null to stdin in daemonize()" again for 4.12 and earlier. >>> In addition, >>> >>> 2ec5339ec921 "tools/libxl: correct vcpu affinity output with sparse >>> physical cpu map" >>> 129025fe3093 "oxenstored: Don't re-open a xenctrl handle for every >>> domain introduction" >>> 7b20a865bc10 "tools/ocaml: Release the global lock before invoking block >>> syscalls" >>> c393b64dcee6 "tools/libxc: Fix issues with libxc and Xen having >>> different featureset lengths" >>> 82855aba5bf9 "tools/libxc: Fix error handling in get_cpuid_domain_info()" >>> 48dab9767d2e "tools/xl: use libxl_domain_info to get domain type for >>> vcpu-pin" >>> >>> 365aabb6e502 "tools/libxendevicemodel: add >>> xendevicemodel_modified_memory_bulk to map" is possibly a candidate, but >>> is also complicated by the stable SONAME. It is perhaps easiest to >>> ignore, seeing as the issue has already gone unnoticed for 2 years. >> Unless these are really urgent to put in, I'd like them to be deferred >> to 4.11.3. With XSA-297 out we've already started the (leaf tree) >> tagging process, so I was really hoping to get this much delayed >> release out rather sooner than later. > > At a minimum, ffb60a58df4 and ffb60a58df4 need backporting, because they > are breakage of tools caused by our recommendation to turn off SMT in > recent XSAs. > > Everything else can be deferred if necessary. Well, it's mostly an all or nothing thing: If another osstest flight is going to be needed anyway, then perhaps the full set could still be put in. But we're already late by 1.5 months... Jan _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [Xen-devel] preparations for 4.11.2 @ 2019-05-16 15:29 ` Jan Beulich 0 siblings, 0 replies; 31+ messages in thread From: Jan Beulich @ 2019-05-16 15:29 UTC (permalink / raw) To: Andrew Cooper, Ian Jackson; +Cc: xen-devel >>> On 16.05.19 at 17:23, <andrew.cooper3@citrix.com> wrote: > On 16/05/2019 16:20, Jan Beulich wrote: >>>>> On 16.05.19 at 17:11, <andrew.cooper3@citrix.com> wrote: >>> On 26/04/2019 13:02, Andrew Cooper wrote: >>>> On 26/04/2019 12:59, Andrew Cooper wrote: >>>>> On 18/03/2019 16:13, Jan Beulich wrote: >>>>>> All, >>>>>> >>>>>> the release is due by the end of the month, but will likely don't make >>>>>> it before early April. Please point out backports you find missing from >>>>>> the respective staging branch, but which you consider relevant. The >>>>>> one commit I've queued already on top of what was just pushed is >>>>>> >>>>>> 22e2f8dddf x86/e820: fix build with gcc9 >>>>> ffb60a58df48419c1f2607cd3cc919fa2bfc9c2d "tools/misc/xenpm: fix getting >>>>> info when some CPUs are offline" for 4.11 and earlier. >>>> Oh, and 677e64dbe315343620c3b266e9eb16623b118038 "tools/ocaml: Dup2 >>>> /dev/null to stdin in daemonize()" again for 4.12 and earlier. >>> In addition, >>> >>> 2ec5339ec921 "tools/libxl: correct vcpu affinity output with sparse >>> physical cpu map" >>> 129025fe3093 "oxenstored: Don't re-open a xenctrl handle for every >>> domain introduction" >>> 7b20a865bc10 "tools/ocaml: Release the global lock before invoking block >>> syscalls" >>> c393b64dcee6 "tools/libxc: Fix issues with libxc and Xen having >>> different featureset lengths" >>> 82855aba5bf9 "tools/libxc: Fix error handling in get_cpuid_domain_info()" >>> 48dab9767d2e "tools/xl: use libxl_domain_info to get domain type for >>> vcpu-pin" >>> >>> 365aabb6e502 "tools/libxendevicemodel: add >>> xendevicemodel_modified_memory_bulk to map" is possibly a candidate, but >>> is also complicated by the stable SONAME. It is perhaps easiest to >>> ignore, seeing as the issue has already gone unnoticed for 2 years. >> Unless these are really urgent to put in, I'd like them to be deferred >> to 4.11.3. With XSA-297 out we've already started the (leaf tree) >> tagging process, so I was really hoping to get this much delayed >> release out rather sooner than later. > > At a minimum, ffb60a58df4 and ffb60a58df4 need backporting, because they > are breakage of tools caused by our recommendation to turn off SMT in > recent XSAs. > > Everything else can be deferred if necessary. Well, it's mostly an all or nothing thing: If another osstest flight is going to be needed anyway, then perhaps the full set could still be put in. But we're already late by 1.5 months... Jan _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: preparations for 4.11.2 @ 2019-05-16 16:21 ` Ian Jackson 0 siblings, 0 replies; 31+ messages in thread From: Ian Jackson @ 2019-05-16 16:21 UTC (permalink / raw) To: Jan Beulich; +Cc: Andrew Cooper, xen-devel Jan Beulich writes ("Re: [Xen-devel] preparations for 4.11.2"): > Unless these are really urgent to put in, I'd like them to be deferred > to 4.11.3. With XSA-297 out we've already started the (leaf tree) > tagging process, so I was really hoping to get this much delayed > release out rather sooner than later. I'm sorry to throw a spanner in the works but at the very least tools/ocaml: Dup2 /dev/null to stdin in daemonize() fixes quite an alarming bug. And Andy is making a case for the SMT fixes. Currently I have these (not yet pushed): 989a2ec4f3ba tools/xl: use libxl_domain_info to get domain type for vcpu-pin b55ff4c879ac tools/libxl: correct vcpu affinity output with sparse physical cpu map 4b72470175a5 tools/ocaml: Dup2 /dev/null to stdin in daemonize() 5c6be595b1bc tools/misc/xenpm: fix getting info when some CPUs are offline Andy also mentioned this: > c393b64dcee6 "tools/libxc: Fix issues with libxc and Xen having > different featureset lengths" which I think may be important but I asked a question about it. Sorry, but I think it may be best to wait. I'm open to being persuaded... Regards, Ian. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [Xen-devel] preparations for 4.11.2 @ 2019-05-16 16:21 ` Ian Jackson 0 siblings, 0 replies; 31+ messages in thread From: Ian Jackson @ 2019-05-16 16:21 UTC (permalink / raw) To: Jan Beulich; +Cc: Andrew Cooper, xen-devel Jan Beulich writes ("Re: [Xen-devel] preparations for 4.11.2"): > Unless these are really urgent to put in, I'd like them to be deferred > to 4.11.3. With XSA-297 out we've already started the (leaf tree) > tagging process, so I was really hoping to get this much delayed > release out rather sooner than later. I'm sorry to throw a spanner in the works but at the very least tools/ocaml: Dup2 /dev/null to stdin in daemonize() fixes quite an alarming bug. And Andy is making a case for the SMT fixes. Currently I have these (not yet pushed): 989a2ec4f3ba tools/xl: use libxl_domain_info to get domain type for vcpu-pin b55ff4c879ac tools/libxl: correct vcpu affinity output with sparse physical cpu map 4b72470175a5 tools/ocaml: Dup2 /dev/null to stdin in daemonize() 5c6be595b1bc tools/misc/xenpm: fix getting info when some CPUs are offline Andy also mentioned this: > c393b64dcee6 "tools/libxc: Fix issues with libxc and Xen having > different featureset lengths" which I think may be important but I asked a question about it. Sorry, but I think it may be best to wait. I'm open to being persuaded... Regards, Ian. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: preparations for 4.11.2 @ 2019-05-17 6:20 ` Jan Beulich 0 siblings, 0 replies; 31+ messages in thread From: Jan Beulich @ 2019-05-17 6:20 UTC (permalink / raw) To: Andrew Cooper, Ian Jackson; +Cc: xen-devel >>> On 16.05.19 at 18:21, <ian.jackson@citrix.com> wrote: > Sorry, but I think it may be best to wait. I'm open to being > persuaded... Okay - as also indicated on irc, with the weekend in between and with the most recent flight having failed anyway it shouldn't be too much of an extra delay. Yet to be honest - most or all of these should have been requested and committed several weeks ago, soon after the mail at the root of this thread was sent. I wonder if I need to add a hard cut-off date to these initiator mails that I send. I'd prefer not to, not the least because of unforeseen issues like the recent month-long (or more?) osstest stall. Jan _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [Xen-devel] preparations for 4.11.2 @ 2019-05-17 6:20 ` Jan Beulich 0 siblings, 0 replies; 31+ messages in thread From: Jan Beulich @ 2019-05-17 6:20 UTC (permalink / raw) To: Andrew Cooper, Ian Jackson; +Cc: xen-devel >>> On 16.05.19 at 18:21, <ian.jackson@citrix.com> wrote: > Sorry, but I think it may be best to wait. I'm open to being > persuaded... Okay - as also indicated on irc, with the weekend in between and with the most recent flight having failed anyway it shouldn't be too much of an extra delay. Yet to be honest - most or all of these should have been requested and committed several weeks ago, soon after the mail at the root of this thread was sent. I wonder if I need to add a hard cut-off date to these initiator mails that I send. I'd prefer not to, not the least because of unforeseen issues like the recent month-long (or more?) osstest stall. Jan _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: preparations for 4.11.2 @ 2019-05-17 16:07 ` Ian Jackson 0 siblings, 0 replies; 31+ messages in thread From: Ian Jackson @ 2019-05-17 16:07 UTC (permalink / raw) To: Jan Beulich; +Cc: Andrew Cooper, xen-devel Jan Beulich writes ("Re: preparations for 4.11.2"): > Okay - as also indicated on irc, with the weekend in between and > with the most recent flight having failed anyway it shouldn't be > too much of an extra delay. Right. OK, I have pushed my queue now. > Yet to be honest - most or all of these > should have been requested and committed several weeks ago, > soon after the mail at the root of this thread was sent. I wonder > if I need to add a hard cut-off date to these initiator mails that I > send. I'd prefer not to, not the least because of unforeseen > issues like the recent month-long (or more?) osstest stall. I think stating a cut-off date would be a very sensible idea. If circumstances change you can always say "because of X, this deadline is being waived/extended". Ian. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [Xen-devel] preparations for 4.11.2 @ 2019-05-17 16:07 ` Ian Jackson 0 siblings, 0 replies; 31+ messages in thread From: Ian Jackson @ 2019-05-17 16:07 UTC (permalink / raw) To: Jan Beulich; +Cc: Andrew Cooper, xen-devel Jan Beulich writes ("Re: preparations for 4.11.2"): > Okay - as also indicated on irc, with the weekend in between and > with the most recent flight having failed anyway it shouldn't be > too much of an extra delay. Right. OK, I have pushed my queue now. > Yet to be honest - most or all of these > should have been requested and committed several weeks ago, > soon after the mail at the root of this thread was sent. I wonder > if I need to add a hard cut-off date to these initiator mails that I > send. I'd prefer not to, not the least because of unforeseen > issues like the recent month-long (or more?) osstest stall. I think stating a cut-off date would be a very sensible idea. If circumstances change you can always say "because of X, this deadline is being waived/extended". Ian. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: preparations for 4.11.2 @ 2019-05-16 16:17 ` Ian Jackson 0 siblings, 0 replies; 31+ messages in thread From: Ian Jackson @ 2019-05-16 16:17 UTC (permalink / raw) To: Andrew Cooper; +Cc: xen-devel Andrew Cooper writes ("Re: [Xen-devel] preparations for 4.11.2"): > In addition, Thanks. ==== wanting discussion: ==== > 365aabb6e502 "tools/libxendevicemodel: add > xendevicemodel_modified_memory_bulk to map" is possibly a candidate, but > is also complicated by the stable SONAME. It is perhaps easiest to > ignore, seeing as the issue has already gone unnoticed for 2 years. We would be bumping the minor version. I think it is ABI compatible. So I am inclined to backport this one but I haven't done so yet. > 129025fe3093 "oxenstored: Don't re-open a xenctrl handle for every > domain introduction" Can you justify how this is a bugfix ? It doesn't seem like backport material to me. > 7b20a865bc10 "tools/ocaml: Release the global lock before invoking block > syscalls" This *really* doesn't look like a bugfix, let alone a backport candidate ! Removing a lock for performance reasons ! > c393b64dcee6 "tools/libxc: Fix issues with libxc and Xen having > different featureset lengths" The compatibility implications here are not clearly spelled out in the commit message. AFAICT, after this commit, the effect is: - new tools will work with old hypervisor - old tools will not necessariloy work with old hypervisor I assume that we are talking here about old and new code with the same Xen version, eg as a result of a security fix. The previous behaviour, ie, what happens without this patch, is not entirely clear to me. > 82855aba5bf9 "tools/libxc: Fix error handling in get_cpuid_domain_info()" This might break some callers, mightn't it ? What callers ? Or is there an argument that there aren't callers which will be broken ? ==== done: ==== > 2ec5339ec921 "tools/libxl: correct vcpu affinity output with sparse > physical cpu map" Backported to 4.11 and 4.10. > 48dab9767d2e "tools/xl: use libxl_domain_info to get domain type for > vcpu-pin" Backported to 4.12, 4.11, 4.10. Ian. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [Xen-devel] preparations for 4.11.2 @ 2019-05-16 16:17 ` Ian Jackson 0 siblings, 0 replies; 31+ messages in thread From: Ian Jackson @ 2019-05-16 16:17 UTC (permalink / raw) To: Andrew Cooper; +Cc: xen-devel Andrew Cooper writes ("Re: [Xen-devel] preparations for 4.11.2"): > In addition, Thanks. ==== wanting discussion: ==== > 365aabb6e502 "tools/libxendevicemodel: add > xendevicemodel_modified_memory_bulk to map" is possibly a candidate, but > is also complicated by the stable SONAME. It is perhaps easiest to > ignore, seeing as the issue has already gone unnoticed for 2 years. We would be bumping the minor version. I think it is ABI compatible. So I am inclined to backport this one but I haven't done so yet. > 129025fe3093 "oxenstored: Don't re-open a xenctrl handle for every > domain introduction" Can you justify how this is a bugfix ? It doesn't seem like backport material to me. > 7b20a865bc10 "tools/ocaml: Release the global lock before invoking block > syscalls" This *really* doesn't look like a bugfix, let alone a backport candidate ! Removing a lock for performance reasons ! > c393b64dcee6 "tools/libxc: Fix issues with libxc and Xen having > different featureset lengths" The compatibility implications here are not clearly spelled out in the commit message. AFAICT, after this commit, the effect is: - new tools will work with old hypervisor - old tools will not necessariloy work with old hypervisor I assume that we are talking here about old and new code with the same Xen version, eg as a result of a security fix. The previous behaviour, ie, what happens without this patch, is not entirely clear to me. > 82855aba5bf9 "tools/libxc: Fix error handling in get_cpuid_domain_info()" This might break some callers, mightn't it ? What callers ? Or is there an argument that there aren't callers which will be broken ? ==== done: ==== > 2ec5339ec921 "tools/libxl: correct vcpu affinity output with sparse > physical cpu map" Backported to 4.11 and 4.10. > 48dab9767d2e "tools/xl: use libxl_domain_info to get domain type for > vcpu-pin" Backported to 4.12, 4.11, 4.10. Ian. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: preparations for 4.11.2 @ 2019-05-16 17:09 ` Andrew Cooper 0 siblings, 0 replies; 31+ messages in thread From: Andrew Cooper @ 2019-05-16 17:09 UTC (permalink / raw) To: Ian Jackson; +Cc: xen-devel On 16/05/2019 17:17, Ian Jackson wrote: > Andrew Cooper writes ("Re: [Xen-devel] preparations for 4.11.2"): >> In addition, > Thanks. > > ==== wanting discussion: ==== > >> 365aabb6e502 "tools/libxendevicemodel: add >> xendevicemodel_modified_memory_bulk to map" is possibly a candidate, but >> is also complicated by the stable SONAME. It is perhaps easiest to >> ignore, seeing as the issue has already gone unnoticed for 2 years. > We would be bumping the minor version. I think it is ABI compatible. > So I am inclined to backport this one but I haven't done so yet. > >> 129025fe3093 "oxenstored: Don't re-open a xenctrl handle for every >> domain introduction" > Can you justify how this is a bugfix ? It doesn't seem like backport > material to me. It was found from strace (while investigating an unrelated issue), but given how many issues we've had in the past with {o,}xenstored exceeding its FD limit, I'd still put it in the category of bugfix. It balloons the worst-case FD requirements by as many concurrent domain starts as the rest of dom0 can manage. > >> 7b20a865bc10 "tools/ocaml: Release the global lock before invoking block >> syscalls" > This *really* doesn't look like a bugfix, let alone a backport > candidate ! Removing a lock for performance reasons ! Of course its a backport candidate, and it is a bugfix even if most of the time it is only observed as a perf improvement. The Ocaml FFI says "thou shalt not make a syscall holding this lock", because while that lock is held, everything is single threaded. IIRC, this particular issue lead to a partial outage of one of our HTTP API endpoints. > >> c393b64dcee6 "tools/libxc: Fix issues with libxc and Xen having >> different featureset lengths" > The compatibility implications here are not clearly spelled out in the > commit message. AFAICT, after this commit, the effect is: > - new tools will work with old hypervisor > - old tools will not necessariloy work with old hypervisor > I assume that we are talking here about old and new code with the same > Xen version, eg as a result of a security fix. > > The previous behaviour, ie, what happens without this patch, is not > entirely clear to me. This was an unintended consequence of XSA-253 (Spectre/Meltdown) where the length of the featureset did increase in a security fix. In the period of time between installing updated dom0 userspace packages, and rebooting into the new hypervisor, attempting to start a guest results in libc heap corruption and an abort(). Because libxl doesn't used the partially-improved CPUID functionality yet, it doesn't hit the second bug of incoming migrates getting intermittently rejected due to 4/8 bytes of heap metadata being included in the CPUID safety check. >> 82855aba5bf9 "tools/libxc: Fix error handling in get_cpuid_domain_info()" > This might break some callers, mightn't it ? What callers ? Or is > there an argument that there aren't callers which will be broken ? This was from the same bit of debugging as above, and ISTR caused some error messages in higher callers to print junk instead of the real error. ~Andrew _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [Xen-devel] preparations for 4.11.2 @ 2019-05-16 17:09 ` Andrew Cooper 0 siblings, 0 replies; 31+ messages in thread From: Andrew Cooper @ 2019-05-16 17:09 UTC (permalink / raw) To: Ian Jackson; +Cc: xen-devel On 16/05/2019 17:17, Ian Jackson wrote: > Andrew Cooper writes ("Re: [Xen-devel] preparations for 4.11.2"): >> In addition, > Thanks. > > ==== wanting discussion: ==== > >> 365aabb6e502 "tools/libxendevicemodel: add >> xendevicemodel_modified_memory_bulk to map" is possibly a candidate, but >> is also complicated by the stable SONAME. It is perhaps easiest to >> ignore, seeing as the issue has already gone unnoticed for 2 years. > We would be bumping the minor version. I think it is ABI compatible. > So I am inclined to backport this one but I haven't done so yet. > >> 129025fe3093 "oxenstored: Don't re-open a xenctrl handle for every >> domain introduction" > Can you justify how this is a bugfix ? It doesn't seem like backport > material to me. It was found from strace (while investigating an unrelated issue), but given how many issues we've had in the past with {o,}xenstored exceeding its FD limit, I'd still put it in the category of bugfix. It balloons the worst-case FD requirements by as many concurrent domain starts as the rest of dom0 can manage. > >> 7b20a865bc10 "tools/ocaml: Release the global lock before invoking block >> syscalls" > This *really* doesn't look like a bugfix, let alone a backport > candidate ! Removing a lock for performance reasons ! Of course its a backport candidate, and it is a bugfix even if most of the time it is only observed as a perf improvement. The Ocaml FFI says "thou shalt not make a syscall holding this lock", because while that lock is held, everything is single threaded. IIRC, this particular issue lead to a partial outage of one of our HTTP API endpoints. > >> c393b64dcee6 "tools/libxc: Fix issues with libxc and Xen having >> different featureset lengths" > The compatibility implications here are not clearly spelled out in the > commit message. AFAICT, after this commit, the effect is: > - new tools will work with old hypervisor > - old tools will not necessariloy work with old hypervisor > I assume that we are talking here about old and new code with the same > Xen version, eg as a result of a security fix. > > The previous behaviour, ie, what happens without this patch, is not > entirely clear to me. This was an unintended consequence of XSA-253 (Spectre/Meltdown) where the length of the featureset did increase in a security fix. In the period of time between installing updated dom0 userspace packages, and rebooting into the new hypervisor, attempting to start a guest results in libc heap corruption and an abort(). Because libxl doesn't used the partially-improved CPUID functionality yet, it doesn't hit the second bug of incoming migrates getting intermittently rejected due to 4/8 bytes of heap metadata being included in the CPUID safety check. >> 82855aba5bf9 "tools/libxc: Fix error handling in get_cpuid_domain_info()" > This might break some callers, mightn't it ? What callers ? Or is > there an argument that there aren't callers which will be broken ? This was from the same bit of debugging as above, and ISTR caused some error messages in higher callers to print junk instead of the real error. ~Andrew _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: preparations for 4.11.2 @ 2019-05-17 9:50 ` Ian Jackson 0 siblings, 0 replies; 31+ messages in thread From: Ian Jackson @ 2019-05-17 9:50 UTC (permalink / raw) To: Andrew Cooper; +Cc: xen-devel Andrew Cooper writes ("Re: [Xen-devel] preparations for 4.11.2"): > On 16/05/2019 17:17, Ian Jackson wrote: > > Andrew Cooper writes ("Re: [Xen-devel] preparations for 4.11.2"): > >> 129025fe3093 "oxenstored: Don't re-open a xenctrl handle for every > >> domain introduction" > > Can you justify how this is a bugfix ? It doesn't seem like backport > > material to me. > > It was found from strace (while investigating an unrelated issue), but > given how many issues we've had in the past with {o,}xenstored exceeding > its FD limit, I'd still put it in the category of bugfix. > > It balloons the worst-case FD requirements by as many concurrent domain > starts as the rest of dom0 can manage. Thanks for the clarification. > >> 7b20a865bc10 "tools/ocaml: Release the global lock before invoking block > >> syscalls" > > This *really* doesn't look like a bugfix, let alone a backport > > candidate ! Removing a lock for performance reasons ! > > Of course its a backport candidate, and it is a bugfix even if most of > the time it is only observed as a perf improvement. > > The Ocaml FFI says "thou shalt not make a syscall holding this lock", > because while that lock is held, everything is single threaded. That wasn't mentioned in the commit message, so I assumed that the reason was the usual one for removing locks, namely simply to increase cpu concurrency. So these are bugfixes, but they're not particularly low risk based just on the code. How long has XS been running these patches ? The answer to that may give me some confidence that for users of Xen stable branches, the possible reward of fixing a mysterious bad behaviour is better to take the risk of these patches having bugs. > >> c393b64dcee6 "tools/libxc: Fix issues with libxc and Xen having > >> different featureset lengths" > > The compatibility implications here are not clearly spelled out in the > > commit message. AFAICT, after this commit, the effect is: > > - new tools will work with old hypervisor > > - old tools will not necessariloy work with old hypervisor > > I assume that we are talking here about old and new code with the same > > Xen version, eg as a result of a security fix. > > > > The previous behaviour, ie, what happens without this patch, is not > > entirely clear to me. > > This was an unintended consequence of XSA-253 (Spectre/Meltdown) where > the length of the featureset did increase in a security fix. Cripes. OK, so that is definitely wanted. It applies cleanly back to 4.8, so I have done that. I guess it needs to be applied to 4.7 too ? I get conflicts. Would you care to try to fix that up or shall I ? > >> 82855aba5bf9 "tools/libxc: Fix error handling in get_cpuid_domain_info()" > > This might break some callers, mightn't it ? What callers ? Or is > > there an argument that there aren't callers which will be broken ? > > This was from the same bit of debugging as above, and ISTR caused some > error messages in higher callers to print junk instead of the real error. Right. I can see that that is annoying. But given that the existing behaviour is inconsistent, it is possible that *other* callers somewhere are relying on the other behaviour and they may even check errno values or something. IOW you are changing the API of this function from "omg what happens depends on what error it even was" to something sane and uniform. That is still an API change though, and maybe not sensible to backport, if the only benefit is improved error messages. It seems to me this would depend on how likely it is that there are any callers that would be broken by the change and in what situations that would break. Thanks for the info. Ian. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [Xen-devel] preparations for 4.11.2 @ 2019-05-17 9:50 ` Ian Jackson 0 siblings, 0 replies; 31+ messages in thread From: Ian Jackson @ 2019-05-17 9:50 UTC (permalink / raw) To: Andrew Cooper; +Cc: xen-devel Andrew Cooper writes ("Re: [Xen-devel] preparations for 4.11.2"): > On 16/05/2019 17:17, Ian Jackson wrote: > > Andrew Cooper writes ("Re: [Xen-devel] preparations for 4.11.2"): > >> 129025fe3093 "oxenstored: Don't re-open a xenctrl handle for every > >> domain introduction" > > Can you justify how this is a bugfix ? It doesn't seem like backport > > material to me. > > It was found from strace (while investigating an unrelated issue), but > given how many issues we've had in the past with {o,}xenstored exceeding > its FD limit, I'd still put it in the category of bugfix. > > It balloons the worst-case FD requirements by as many concurrent domain > starts as the rest of dom0 can manage. Thanks for the clarification. > >> 7b20a865bc10 "tools/ocaml: Release the global lock before invoking block > >> syscalls" > > This *really* doesn't look like a bugfix, let alone a backport > > candidate ! Removing a lock for performance reasons ! > > Of course its a backport candidate, and it is a bugfix even if most of > the time it is only observed as a perf improvement. > > The Ocaml FFI says "thou shalt not make a syscall holding this lock", > because while that lock is held, everything is single threaded. That wasn't mentioned in the commit message, so I assumed that the reason was the usual one for removing locks, namely simply to increase cpu concurrency. So these are bugfixes, but they're not particularly low risk based just on the code. How long has XS been running these patches ? The answer to that may give me some confidence that for users of Xen stable branches, the possible reward of fixing a mysterious bad behaviour is better to take the risk of these patches having bugs. > >> c393b64dcee6 "tools/libxc: Fix issues with libxc and Xen having > >> different featureset lengths" > > The compatibility implications here are not clearly spelled out in the > > commit message. AFAICT, after this commit, the effect is: > > - new tools will work with old hypervisor > > - old tools will not necessariloy work with old hypervisor > > I assume that we are talking here about old and new code with the same > > Xen version, eg as a result of a security fix. > > > > The previous behaviour, ie, what happens without this patch, is not > > entirely clear to me. > > This was an unintended consequence of XSA-253 (Spectre/Meltdown) where > the length of the featureset did increase in a security fix. Cripes. OK, so that is definitely wanted. It applies cleanly back to 4.8, so I have done that. I guess it needs to be applied to 4.7 too ? I get conflicts. Would you care to try to fix that up or shall I ? > >> 82855aba5bf9 "tools/libxc: Fix error handling in get_cpuid_domain_info()" > > This might break some callers, mightn't it ? What callers ? Or is > > there an argument that there aren't callers which will be broken ? > > This was from the same bit of debugging as above, and ISTR caused some > error messages in higher callers to print junk instead of the real error. Right. I can see that that is annoying. But given that the existing behaviour is inconsistent, it is possible that *other* callers somewhere are relying on the other behaviour and they may even check errno values or something. IOW you are changing the API of this function from "omg what happens depends on what error it even was" to something sane and uniform. That is still an API change though, and maybe not sensible to backport, if the only benefit is improved error messages. It seems to me this would depend on how likely it is that there are any callers that would be broken by the change and in what situations that would break. Thanks for the info. Ian. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: preparations for 4.11.2 @ 2019-05-17 16:04 ` Ian Jackson 0 siblings, 0 replies; 31+ messages in thread From: Ian Jackson @ 2019-05-17 16:04 UTC (permalink / raw) To: Andrew Cooper, xen-devel Ian Jackson writes ("Re: [Xen-devel] preparations for 4.11.2"): > Andrew Cooper writes ("Re: [Xen-devel] preparations for 4.11.2"): > > On 16/05/2019 17:17, Ian Jackson wrote: > > > Andrew Cooper writes ("Re: [Xen-devel] preparations for 4.11.2"): > > >> 129025fe3093 "oxenstored: Don't re-open a xenctrl handle for every > > >> domain introduction" ... > > >> 7b20a865bc10 "tools/ocaml: Release the global lock before invoking block > > >> syscalls" ... > So these are bugfixes, but they're not particularly low risk based > just on the code. How long has XS been running these patches ? The > answer to that may give me some confidence that for users of Xen > stable branches, the possible reward of fixing a mysterious bad > behaviour is better to take the risk of these patches having bugs. Based on this: 12:17 <andyhhp> XS has been using those ocaml changes for longer than they've been upstream 12:19 <andyhhp> although if you're still hesitant, it really isn't the end of the world. Your decision here doesn't affect XS - we've already got them backported in the patchqueue I have taken 129025fe3093 "oxenstored: Don't re-open a xenctrl handle..." to 4.11 and 4.10. 7b20a865bc10 "tools/ocaml: Release the global lock..." does not apply cleanly. Do you happen to have a version for 4.11 and/or 4.10 ? I am not convinced I ought to try to fix the backport myself particularly if Citrix XS have been running a textually different patch for a long time... The rest of this I think is still in question and IMO not a blocker for 4.11.2. Ian. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [Xen-devel] preparations for 4.11.2 @ 2019-05-17 16:04 ` Ian Jackson 0 siblings, 0 replies; 31+ messages in thread From: Ian Jackson @ 2019-05-17 16:04 UTC (permalink / raw) To: Andrew Cooper, xen-devel Ian Jackson writes ("Re: [Xen-devel] preparations for 4.11.2"): > Andrew Cooper writes ("Re: [Xen-devel] preparations for 4.11.2"): > > On 16/05/2019 17:17, Ian Jackson wrote: > > > Andrew Cooper writes ("Re: [Xen-devel] preparations for 4.11.2"): > > >> 129025fe3093 "oxenstored: Don't re-open a xenctrl handle for every > > >> domain introduction" ... > > >> 7b20a865bc10 "tools/ocaml: Release the global lock before invoking block > > >> syscalls" ... > So these are bugfixes, but they're not particularly low risk based > just on the code. How long has XS been running these patches ? The > answer to that may give me some confidence that for users of Xen > stable branches, the possible reward of fixing a mysterious bad > behaviour is better to take the risk of these patches having bugs. Based on this: 12:17 <andyhhp> XS has been using those ocaml changes for longer than they've been upstream 12:19 <andyhhp> although if you're still hesitant, it really isn't the end of the world. Your decision here doesn't affect XS - we've already got them backported in the patchqueue I have taken 129025fe3093 "oxenstored: Don't re-open a xenctrl handle..." to 4.11 and 4.10. 7b20a865bc10 "tools/ocaml: Release the global lock..." does not apply cleanly. Do you happen to have a version for 4.11 and/or 4.10 ? I am not convinced I ought to try to fix the backport myself particularly if Citrix XS have been running a textually different patch for a long time... The rest of this I think is still in question and IMO not a blocker for 4.11.2. Ian. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: preparations for 4.11.2 @ 2019-05-16 16:02 ` Ian Jackson 0 siblings, 0 replies; 31+ messages in thread From: Ian Jackson @ 2019-05-16 16:02 UTC (permalink / raw) To: Andrew Cooper; +Cc: xen-devel, Christian Lindig Andrew Cooper writes ("Re: [Xen-devel] preparations for 4.11.2"): > Oh, and 677e64dbe315343620c3b266e9eb16623b118038 "tools/ocaml: Dup2 > /dev/null to stdin in daemonize()" again for 4.12 and earlier. This is already in 4.12. It's quite alarming so I decided to backport it all the way to 4.7 (last release in security support). Ian. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [Xen-devel] preparations for 4.11.2 @ 2019-05-16 16:02 ` Ian Jackson 0 siblings, 0 replies; 31+ messages in thread From: Ian Jackson @ 2019-05-16 16:02 UTC (permalink / raw) To: Andrew Cooper; +Cc: xen-devel, Christian Lindig Andrew Cooper writes ("Re: [Xen-devel] preparations for 4.11.2"): > Oh, and 677e64dbe315343620c3b266e9eb16623b118038 "tools/ocaml: Dup2 > /dev/null to stdin in daemonize()" again for 4.12 and earlier. This is already in 4.12. It's quite alarming so I decided to backport it all the way to 4.7 (last release in security support). Ian. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: preparations for 4.11.2 @ 2019-05-16 15:54 ` Ian Jackson 0 siblings, 0 replies; 31+ messages in thread From: Ian Jackson @ 2019-05-16 15:54 UTC (permalink / raw) To: Andrew Cooper; +Cc: xen-devel Andrew Cooper writes ("Re: [Xen-devel] preparations for 4.11.2"): > ffb60a58df48419c1f2607cd3cc919fa2bfc9c2d "tools/misc/xenpm: fix getting > info when some CPUs are offline" for 4.11 and earlier. Thanks, queued for 4.11 and 4.10. Ian. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
* Re: [Xen-devel] preparations for 4.11.2 @ 2019-05-16 15:54 ` Ian Jackson 0 siblings, 0 replies; 31+ messages in thread From: Ian Jackson @ 2019-05-16 15:54 UTC (permalink / raw) To: Andrew Cooper; +Cc: xen-devel Andrew Cooper writes ("Re: [Xen-devel] preparations for 4.11.2"): > ffb60a58df48419c1f2607cd3cc919fa2bfc9c2d "tools/misc/xenpm: fix getting > info when some CPUs are offline" for 4.11 and earlier. Thanks, queued for 4.11 and 4.10. Ian. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel ^ permalink raw reply [flat|nested] 31+ messages in thread
end of thread, other threads:[~2019-05-17 16:07 UTC | newest] Thread overview: 31+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2019-03-18 16:13 preparations for 4.11.2 Jan Beulich 2019-04-26 11:59 ` Andrew Cooper 2019-04-26 11:59 ` [Xen-devel] " Andrew Cooper 2019-04-26 12:02 ` Andrew Cooper 2019-04-26 12:02 ` [Xen-devel] " Andrew Cooper 2019-05-16 15:11 ` Andrew Cooper 2019-05-16 15:11 ` [Xen-devel] " Andrew Cooper 2019-05-16 15:20 ` Jan Beulich 2019-05-16 15:20 ` [Xen-devel] " Jan Beulich 2019-05-16 15:23 ` Andrew Cooper 2019-05-16 15:23 ` [Xen-devel] " Andrew Cooper 2019-05-16 15:29 ` Jan Beulich 2019-05-16 15:29 ` [Xen-devel] " Jan Beulich 2019-05-16 16:21 ` Ian Jackson 2019-05-16 16:21 ` [Xen-devel] " Ian Jackson 2019-05-17 6:20 ` Jan Beulich 2019-05-17 6:20 ` [Xen-devel] " Jan Beulich 2019-05-17 16:07 ` Ian Jackson 2019-05-17 16:07 ` [Xen-devel] " Ian Jackson 2019-05-16 16:17 ` Ian Jackson 2019-05-16 16:17 ` [Xen-devel] " Ian Jackson 2019-05-16 17:09 ` Andrew Cooper 2019-05-16 17:09 ` [Xen-devel] " Andrew Cooper 2019-05-17 9:50 ` Ian Jackson 2019-05-17 9:50 ` [Xen-devel] " Ian Jackson 2019-05-17 16:04 ` Ian Jackson 2019-05-17 16:04 ` [Xen-devel] " Ian Jackson 2019-05-16 16:02 ` Ian Jackson 2019-05-16 16:02 ` [Xen-devel] " Ian Jackson 2019-05-16 15:54 ` Ian Jackson 2019-05-16 15:54 ` [Xen-devel] " Ian Jackson
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.