)]}' { "log": [ { "commit": "6f7da290413ba713f0cdd9ff1a2a9bb129ef4f6c", "tree": "be557c02369355dcae1e58fdc8f0f8c7ca25b0f4", "parents": [ "401e000ab90d7b81d8ea0735e3ff909548754876" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 02 16:07:02 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 02 16:07:02 2017 -0700" }, "message": "Linux 4.12\n" }, { "commit": "401e000ab90d7b81d8ea0735e3ff909548754876", "tree": "bc7eae657403c34bd4b13c84d769cc7e650745c7", "parents": [ "79c496816963aa0561868b43c2c950dfeb282639" ], "author": { "name": "Sylvain \u0027ythier\u0027 Hitier", "email": "sylvain.hitier@gmail.com", "time": "Sun Jul 02 15:21:56 2017 +0200" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 02 15:37:23 2017 -0700" }, "message": "moduleparam: fix doc: hwparam_irq configures an IRQ\n\nSigned-off-by: Sylvain \u0027ythier\u0027 Hitier \u003csylvain.hitier@gmail.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "79c496816963aa0561868b43c2c950dfeb282639", "tree": "e127b465e24c9ba27d996fcb7b70aa67f60ffb3e", "parents": [ "3a61a54cd72c93afa3b7246e3ed06f26ed37fde7", "854236363370995a609a10b03e35fd3dc5e9e4a1" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 02 11:53:44 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 02 11:53:44 2017 -0700" }, "message": "Merge branch \u0027upstream\u0027 of git://git.linux-mips.org/pub/scm/ralf/upstream-linus\n\nPull MIPS fixes from Ralf Baechle:\n \"Here\u0027s a final round of fixes for 4.12:\n\n - Fix misordered instructions in assembly code making kenel startup\n via UHB unreliable.\n\n - Fix special case of MADDF and MADDF emulation.\n\n - Fix alignment issue in address calculation in pm-cps on 64 bit.\n\n - Fix IRQ tracing \u0026 lockdep when rescheduling\n\n - Systems with MAARs require post-DMA cache flushes.\n\n The reordering fix and the MADDF/MSUBF fix have sat in linux-next for\n a number of days. The others haven\u0027t propagated from my pull tree to\n linux-next yet but all have survived manual testing and Imagination\u0027s\n automated test system and there are no pending bug reports\"\n\n* \u0027upstream\u0027 of git://git.linux-mips.org/pub/scm/ralf/upstream-linus:\n MIPS: Avoid accidental raw backtrace\n MIPS: Perform post-DMA cache flushes on systems with MAARs\n MIPS: Fix IRQ tracing \u0026 lockdep when rescheduling\n MIPS: pm-cps: Drop manual cache-line alignment of ready_count\n MIPS: math-emu: Handle zero accumulator case in MADDF and MSUBF separately\n MIPS: head: Reorder instructions missing a delay slot\n" }, { "commit": "3a61a54cd72c93afa3b7246e3ed06f26ed37fde7", "tree": "545290a2ba291bf70959bea1ad6b18650b8c7ec6", "parents": [ "e18aca0236a2dac4a134ace4685e97ad09d3605b", "9e25ebfe56ece7541cd10a20d715cbdd148a2e06" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 02 10:09:40 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jul 02 10:09:40 2017 -0700" }, "message": "Merge branch \u0027fixes\u0027 of git://git.armlinux.org.uk/~rmk/linux-arm\n\nPull ARM fix from Russell King:\n \"One final fix for 4.12 - Doug found a boot failure case triggered by\n requesting a non-even MB vmalloc size\"\n\n* \u0027fixes\u0027 of git://git.armlinux.org.uk/~rmk/linux-arm:\n ARM: 8685/1: ensure memblock-limit is pmd-aligned\n" }, { "commit": "e18aca0236a2dac4a134ace4685e97ad09d3605b", "tree": "ed071ab72fa4d27cf5397f3bcd09cdf84206d813", "parents": [ "a527bf61404cd36fedd81dc165a03b6f5529092e", "79298acc4ba097e9ab78644e3e38902d73547c92" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Jul 01 09:10:17 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Jul 01 09:10:17 2017 -0700" }, "message": "Merge branch \u0027x86-urgent-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull x86 fixes from Thomas Gleixner:\n \"Fixlets for x86:\n\n - Prevent kexec crash when KASLR is enabled, which was caused by an\n address calculation bug\n\n - Restore the freeing of PUDs on memory hot remove\n\n - Correct a negated pointer check in the intel uncore performance\n monitoring driver\n\n - Plug a memory leak in an error exit path in the RDT code\"\n\n* \u0027x86-urgent-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n x86/intel_rdt: Fix memory leak on mount failure\n x86/boot/KASLR: Fix kexec crash due to \u0027virt_addr\u0027 calculation bug\n x86/boot/KASLR: Add checking for the offset of kernel virtual address randomization\n perf/x86/intel/uncore: Fix wrong box pointer check\n x86/mm/hotplug: Fix BUG_ON() after hot-remove by not freeing PUD\n" }, { "commit": "a527bf61404cd36fedd81dc165a03b6f5529092e", "tree": "49efc79a4bc1a4c8dc662e4105863fd127e86ef9", "parents": [ "46589d7ab79a4b5f4d3183fafe1c8fb011f82210", "e3c2c4fb524b548e58108a8820c0ac7787667584" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Jul 01 08:46:52 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Jul 01 08:46:52 2017 -0700" }, "message": "Merge branch \u0027perf-urgent-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull perf fix from Thomas Gleixner:\n \"The last fix for perf for this cycles:\n\n - Prevent a segfault when kernel.kptr_restrict\u003d2 is set by avoiding a\n null pointer dereference\"\n\n* \u0027perf-urgent-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n perf machine: Fix segfault for kernel.kptr_restrict\u003d2\n" }, { "commit": "46589d7ab79a4b5f4d3183fafe1c8fb011f82210", "tree": "04f5e86f1313a46e198127cba94dd01e43dcc65c", "parents": [ "fc93274ab5a455b1e31aea5d2c9e8ba1f699fded", "1d80df93d9d170eba8c3654719e5857307efd077" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Jul 01 08:39:13 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Jul 01 08:39:13 2017 -0700" }, "message": "Merge tag \u0027pinctrl-v4.12-4\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl\n\nPull pinctrl fix from Linus Walleij:\n \"Brian noticed that this regression has not got a proper fix for the\n entire merge window and consequently we need to revert the offending\n commit.\n\n It\u0027s part of the RT-mainstream work, the dance goes like this, two\n steps forward, one step back.\n\n Summary:\n\n - A last fix for v4.12, an IRQ problem reported early in the merge\n window appears not to have been properly fixed, so the offending\n commit will be reverted and we will find the proper fix for v4.13.\n Hopefully\"\n\n* tag \u0027pinctrl-v4.12-4\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl:\n Revert \"pinctrl: rockchip: avoid hardirq-unsafe functions in irq_chip\"\n" }, { "commit": "fc93274ab5a455b1e31aea5d2c9e8ba1f699fded", "tree": "e9cc70e49331791d16663e2ffc899e11d5f8bce7", "parents": [ "c0a0c7a4e1200bfea439b9444e6d6b4bede9db23", "c06632ea054c49510efacb42c52aab693c45b7ba" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Jul 01 08:24:54 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Jul 01 08:24:54 2017 -0700" }, "message": "Merge tag \u0027gpio-v4.12-4\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio\n\nPull last minute fixes for GPIO from Linus Walleij:\n\n - Fix another ACPI problem with broken BIOSes.\n\n - Filter out the right GPIO events, making a very user-visible bug go\n away.\n\n* tag \u0027gpio-v4.12-4\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio:\n gpio: acpi: Skip _AEI entries without a handler rather then aborting the scan\n gpiolib: fix filtering out unwanted events\n" }, { "commit": "c0a0c7a4e1200bfea439b9444e6d6b4bede9db23", "tree": "93de4bfcfc0892c8ed37c029254aaa1adec2256f", "parents": [ "fbd576295d6f98b20356d1e6efbc8f976b17c8af", "9e52b32567126fe146f198971364f68d3bc5233f" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 30 17:18:57 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 30 17:18:57 2017 -0700" }, "message": "Merge tag \u0027trace-v4.12-rc5\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace\n\nPull last-minute tracing fixes from Steven Rostedt:\n \"Two fixes:\n\n One is for a crash when using the :mod: trace probe command into\n stack_trace_filter. This bug was introduced during the last merge\n window.\n\n The other was there forever. It\u0027s a small bug that makes it impossible\n to name a module function for kprobes when the module starts with a\n digit\"\n\n* tag \u0027trace-v4.12-rc5\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace:\n tracing/kprobes: Allow to create probe with a module name starting with a digit\n ftrace: Fix regression with module command in stack_trace_filter\n" }, { "commit": "fbd576295d6f98b20356d1e6efbc8f976b17c8af", "tree": "8cb8ca1ba24c6355b52793048fa10079e979d3f2", "parents": [ "dbd1877754338dc98818d200cd25cf398d8b5d9e" ], "author": { "name": "Zack Weinberg", "email": "zackw@panix.com", "time": "Wed Jun 14 08:14:28 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 30 13:53:07 2017 -0700" }, "message": "uapi/linux/a.out.h: don\u0027t use deprecated system-specific predefines.\n\nuapi/linux/a.out.h uses a number of predefined macros that are\ndeprecated because they\u0027re in the application namespace\n(e.g. \u0027#ifdef linux\u0027 instead of \u0027#ifdef __linux__\u0027).\nThis patch either corrects or just removes them if they are not\napplicable to Linux.\n\nThe primary reason this is worth bothering to fix, considering how\nobsolete a.out binary support is, is that the GCC build process\nconsiders this such a severe error that it will copy the header into a\nprivate directory and change the macro names, which causes future\nupdates to the header to be masked. This header probably doesn\u0027t get\nupdated very often anymore, but it is the _only_ uapi header that gets\nthis treatment, so IMHO it is worth patching just to drive that number\nall the way to zero.\n\nSigned-off-by: Zack Weinberg \u003czackw@panix.com\u003e\n[hch: removed dead conditionals]\nSigned-off-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "dbd1877754338dc98818d200cd25cf398d8b5d9e", "tree": "4d9f75bcae6fb266a5082c6af7abb1cec2573752", "parents": [ "b4df2e3537bef7d867338063da56b557c50f68f2" ], "author": { "name": "Jakub Kicinski", "email": "jakub.kicinski@netronome.com", "time": "Wed Jun 28 21:25:48 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 30 13:49:53 2017 -0700" }, "message": "hashtable: remove repeated phrase from a comment\n\n\"in a rcu enabled hashtable\" is repeated twice in a comment.\n\nSigned-off-by: Jakub Kicinski \u003cjakub.kicinski@netronome.com\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "79298acc4ba097e9ab78644e3e38902d73547c92", "tree": "a8e0f66c760d12d29e5ed4e920c1e738bf96bc9c", "parents": [ "8eabf42ae5237e6b699aeac687b5b629e3537c8d" ], "author": { "name": "Vikas Shivappa", "email": "vikas.shivappa@linux.intel.com", "time": "Mon Jun 26 11:55:49 2017 -0700" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Fri Jun 30 21:20:00 2017 +0200" }, "message": "x86/intel_rdt: Fix memory leak on mount failure\n\nIf mount fails, the kn_info directory is not freed causing memory leak.\n\nAdd the missing error handling path.\n\nFixes: 4e978d06dedb (\"x86/intel_rdt: Add \"info\" files to resctrl file system\")\nSigned-off-by: Vikas Shivappa \u003cvikas.shivappa@linux.intel.com\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: ravi.v.shankar@intel.com\nCc: tony.luck@intel.com\nCc: fenghua.yu@intel.com\nCc: peterz@infradead.org\nCc: vikas.shivappa@intel.com\nCc: andi.kleen@intel.com\nCc: stable@vger.kernel.org\nLink: http://lkml.kernel.org/r/1498503368-20173-3-git-send-email-vikas.shivappa@linux.intel.com\n\n" }, { "commit": "b4df2e3537bef7d867338063da56b557c50f68f2", "tree": "267992a93dfdd25c1fed985da62c14603cf6abd1", "parents": [ "27ab862a3afdd7a0285b69a7475f8af7bd2434c4", "d6bd8194e2867e85ac2de63486d3b83ccfae4e62" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 30 10:55:34 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 30 10:55:34 2017 -0700" }, "message": "Merge tag \u0027powerpc-4.12-8\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux\n\nPull powerpc fixes from Michael Ellerman:\n \"Hopefully the last two powerpc fixes for 4.12.\n\n The CXL one is larger than I\u0027d usually send at rc7, but it fixes new\n code this cycle, so better to have it working for the release. It was\n actually sent a few weeks back but got blocked in testing behind\n another fix that was causing issues.\n\n We are still tracking one crash in v4.12-rc7, but only one person has\n reproduced it and the commit identified by bisect doesn\u0027t touch any of\n the relevant code, so I think it\u0027s 50/50 whether that commit is\n actually the problem or it\u0027s some code layout / toolchain issue.\n\n Two fixes for code we merged this cycle:\n\n - cxl: Fixes for Coherent Accelerator Interface Architecture 2.0\n\n - Avoid miscompilation w/GCC 4.6.3 on 32-bit - don\u0027t inline\n copy_to/from_user()\n\n Thanks to Al Viro, Larry Finger, Christophe Lombard\"\n\n* tag \u0027powerpc-4.12-8\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:\n powerpc/32: Avoid miscompilation w/GCC 4.6.3 - don\u0027t inline copy_to/from_user()\n cxl: Fixes for Coherent Accelerator Interface Architecture 2.0\n" }, { "commit": "27ab862a3afdd7a0285b69a7475f8af7bd2434c4", "tree": "ba8b6be81c3019b1e59c00ec526113c779f65747", "parents": [ "4adc6b9382d5a2c1f39fa44b1e6ff29467e844e5", "84a21dbdef0b96d773599c33c2afbb002198d303" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 30 10:37:48 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 30 10:37:48 2017 -0700" }, "message": "Merge tag \u0027iommu-fixes-v4.12-rc7\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu\n\nPull IOMMU fixes from Joerg Roedel:\n \"Two fixes:\n\n - A fix for AMD IOMMU interrupt remapping code when IRQs are\n forwarded directly to KVM guests\n\n - Fixed check in the recently merged code to allow tboot with\n Intel VT-d disabled\"\n\n* tag \u0027iommu-fixes-v4.12-rc7\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu:\n iommu/amd: Fix interrupt remapping when disable guest_mode\n iommu/vt-d: Correctly disable Intel IOMMU force on\n" }, { "commit": "4adc6b9382d5a2c1f39fa44b1e6ff29467e844e5", "tree": "0d06384a369bee0e16fdd3e56557fc75084dd0d3", "parents": [ "86c3e00afd64a507aef954cd581547f28807e51a", "d94815f917da770d42c377786dc428f542e38f71" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 30 10:30:26 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 30 10:30:26 2017 -0700" }, "message": "Merge tag \u0027sound-4.12\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound\n\nPull sound fixes from Takashi Iwai:\n \"Two last-minute HD-audio fixes\"\n\n* tag \u0027sound-4.12\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:\n ALSA: hda - Fix endless loop of codec configure\n ALSA: hda - set input_path bitmap to zero after moving it to new place\n" }, { "commit": "86c3e00afd64a507aef954cd581547f28807e51a", "tree": "df0ad2fcc7749a282683acc5aac473c398b2bbc5", "parents": [ "4d8a991d460d4fa4829beaffdcba45a217ca0fa7", "fbaf94ee3cd5cfa85fe2d4d04be968820cb00556" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 30 10:22:59 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 30 10:22:59 2017 -0700" }, "message": "Merge branch \u0027overlayfs-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/vfs\n\nPull overlayfs fixes from Miklos Szeredi:\n \"Fix two bugs in copy-up code. One introduced in 4.11 and one in\n 4.12-rc\"\n\n* \u0027overlayfs-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/vfs:\n ovl: don\u0027t set origin on broken lower hardlink\n ovl: copy-up: don\u0027t unlock between lookup and link\n" }, { "commit": "8eabf42ae5237e6b699aeac687b5b629e3537c8d", "tree": "8c44cf4bae96bf80da05770b0212dff252c27e9e", "parents": [ "b892cb873ced2af57dc5a018557d128c53ed6ae0" ], "author": { "name": "Baoquan He", "email": "bhe@redhat.com", "time": "Tue Jun 27 20:39:06 2017 +0800" }, "committer": { "name": "Ingo Molnar", "email": "mingo@kernel.org", "time": "Fri Jun 30 08:53:14 2017 +0200" }, "message": "x86/boot/KASLR: Fix kexec crash due to \u0027virt_addr\u0027 calculation bug\n\nKernel text KASLR is separated into physical address and virtual\naddress randomization. And for virtual address randomization, we\nonly randomiza to get an offset between 16M and KERNEL_IMAGE_SIZE.\nSo the initial value of \u0027virt_addr\u0027 should be LOAD_PHYSICAL_ADDR,\nbut not the original kernel loading address \u0027output\u0027.\n\nThe bug will cause kernel boot failure if kernel is loaded at a different\nposition than the address, 16M, which is decided at compiled time.\nKexec/kdump is such practical case.\n\nTo fix it, just assign LOAD_PHYSICAL_ADDR to virt_addr as initial\nvalue.\n\nTested-by: Dave Young \u003cdyoung@redhat.com\u003e\nSigned-off-by: Baoquan He \u003cbhe@redhat.com\u003e\nCc: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nCc: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nFixes: 8391c73 (\"x86/KASLR: Randomize virtual address separately\")\nLink: http://lkml.kernel.org/r/1498567146-11990-3-git-send-email-bhe@redhat.com\nSigned-off-by: Ingo Molnar \u003cmingo@kernel.org\u003e\n" }, { "commit": "b892cb873ced2af57dc5a018557d128c53ed6ae0", "tree": "31b01dbb7af587f6901862e071336f43a6a46c10", "parents": [ "80c65fdb4c6920e332a9781a3de5877594b07522" ], "author": { "name": "Baoquan He", "email": "bhe@redhat.com", "time": "Tue Jun 27 20:39:05 2017 +0800" }, "committer": { "name": "Ingo Molnar", "email": "mingo@kernel.org", "time": "Fri Jun 30 08:53:14 2017 +0200" }, "message": "x86/boot/KASLR: Add checking for the offset of kernel virtual address randomization\n\nFor kernel text KASLR, the virtual address is confined to area of 1G,\n[0xffffffff80000000, 0xffffffffc0000000). For the implemenataion of\nvirtual address randomization, we only randomize to get an offset\nbetween 16M and 1G, then add this offset to the starting address,\n0xffffffff80000000. Here 16M is the offset which is decided at linking\nstage. So the amount of the local variable \u0027virt_addr\u0027 which respresents\nthe offset plus the kernel output size can not exceed KERNEL_IMAGE_SIZE.\n\nAdd a debug check for the offset. If out of bounds, print error\nmessage and hang there.\n\nSuggested-by: Ingo Molnar \u003cmingo@kernel.org\u003e\nSigned-off-by: Baoquan He \u003cbhe@redhat.com\u003e\nCc: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nCc: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nLink: http://lkml.kernel.org/r/1498567146-11990-2-git-send-email-bhe@redhat.com\nSigned-off-by: Ingo Molnar \u003cmingo@kernel.org\u003e\n" }, { "commit": "9e52b32567126fe146f198971364f68d3bc5233f", "tree": "a13bfbc453fe0005ff13f13997aefa299636b359", "parents": [ "0f17976568b3f72e676450af0c0db6f8752253d6" ], "author": { "name": "Sabrina Dubroca", "email": "sd@queasysnail.net", "time": "Thu Jun 22 11:24:42 2017 +0200" }, "committer": { "name": "Steven Rostedt (VMware)", "email": "rostedt@goodmis.org", "time": "Thu Jun 29 23:13:23 2017 -0400" }, "message": "tracing/kprobes: Allow to create probe with a module name starting with a digit\n\nAlways try to parse an address, since kstrtoul() will safely fail when\ngiven a symbol as input. If that fails (which will be the case for a\nsymbol), try to parse a symbol instead.\n\nThis allows creating a probe such as:\n\n p:probe/vlan_gro_receive 8021q:vlan_gro_receive+0\n\nWhich is necessary for this command to work:\n\n perf probe -m 8021q -a vlan_gro_receive\n\nLink: http://lkml.kernel.org/r/fd72d666f45b114e2c5b9cf7e27b91de1ec966f1.1498122881.git.sd@queasysnail.net\n\nCc: stable@vger.kernel.org\nFixes: 413d37d1e (\"tracing: Add kprobe-based event tracer\")\nAcked-by: Masami Hiramatsu \u003cmhiramat@kernel.org\u003e\nSigned-off-by: Sabrina Dubroca \u003csd@queasysnail.net\u003e\nSigned-off-by: Steven Rostedt (VMware) \u003crostedt@goodmis.org\u003e\n" }, { "commit": "854236363370995a609a10b03e35fd3dc5e9e4a1", "tree": "018ed1df38ce86f1576e664a3104de867ec95028", "parents": [ "cad482c1b1844136fc6d9e33f10d3386344fb045" ], "author": { "name": "James Hogan", "email": "james.hogan@imgtec.com", "time": "Thu Jun 29 15:05:04 2017 +0100" }, "committer": { "name": "Ralf Baechle", "email": "ralf@linux-mips.org", "time": "Fri Jun 30 04:42:15 2017 +0200" }, "message": "MIPS: Avoid accidental raw backtrace\n\nSince commit 81a76d7119f6 (\"MIPS: Avoid using unwind_stack() with\nusermode\") show_backtrace() invokes the raw backtracer when\ncp0_status \u0026 ST0_KSU indicates user mode to fix issues on EVA kernels\nwhere user and kernel address spaces overlap.\n\nHowever this is used by show_stack() which creates its own pt_regs on\nthe stack and leaves cp0_status uninitialised in most of the code paths.\nThis results in the non deterministic use of the raw back tracer\ndepending on the previous stack content.\n\nshow_stack() deals exclusively with kernel mode stacks anyway, so\nexplicitly initialise regs.cp0_status to KSU_KERNEL (i.e. 0) to ensure\nwe get a useful backtrace.\n\nFixes: 81a76d7119f6 (\"MIPS: Avoid using unwind_stack() with usermode\")\nSigned-off-by: James Hogan \u003cjames.hogan@imgtec.com\u003e\nCc: linux-mips@linux-mips.org\nCc: \u003cstable@vger.kernel.org\u003e # 3.15+\nPatchwork: https://patchwork.linux-mips.org/patch/16656/\nSigned-off-by: Ralf Baechle \u003cralf@linux-mips.org\u003e\n" }, { "commit": "cad482c1b1844136fc6d9e33f10d3386344fb045", "tree": "a3f0f718e15f5e03bb9fc19c860719462ec9bc3f", "parents": [ "d8550860d910c6b7b70f830f59003b33daaa52c9" ], "author": { "name": "Paul Burton", "email": "paul.burton@imgtec.com", "time": "Tue Jun 13 10:01:08 2017 -0700" }, "committer": { "name": "Ralf Baechle", "email": "ralf@linux-mips.org", "time": "Fri Jun 30 04:40:46 2017 +0200" }, "message": "MIPS: Perform post-DMA cache flushes on systems with MAARs\n\nRecent CPUs from Imagination Technologies such as the I6400 or P6600 are\nable to speculatively fetch data from memory into caches. This means\nthat if used in a system with non-coherent DMA they require that caches\nbe invalidated after a device performs DMA, and before the CPU reads the\nDMA\u0027d data, in order to ensure that stale values weren\u0027t speculatively\nprefetched.\n\nSuch CPUs also introduced Memory Accessibility Attribute Registers\n(MAARs) in order to control the regions in which they are allowed to\nspeculate. Thus we can use the presence of MAARs as a good indication\nthat the CPU requires the above cache maintenance. Use the presence of\nMAARs to determine the result of cpu_needs_post_dma_flush() in the\ndefault case, in order to handle these recent CPUs correctly.\n\nNote that the return type of cpu_needs_post_dma_flush() is changed to\nbool, such that it\u0027s clearer what\u0027s happening when cpu_has_maar is cast\nto bool for the return value. If this patch were backported to a\npre-v4.7 kernel then MIPS_CPU_MAAR was 1ull\u003c\u003c34, so when cast to an int\nwe would incorrectly return 0. It so happens that MIPS_CPU_MAAR is\ncurrently 1ull\u003c\u003c30, so when truncated to an int gives a non-zero value\nanyway, but even so the implicit conversion from long long int to bool\nmakes it clearer to understand what will happen than the implicit\nconversion from long long int to int would. The bool return type also\nfits this usage better semantically, so seems like an all-round win.\n\nThanks to Ed for spotting the issue for pre-v4.7 kernels \u0026 suggesting\nthe return type change.\n\nSigned-off-by: Paul Burton \u003cpaul.burton@imgtec.com\u003e\nReviewed-by: Bryan O\u0027Donoghue \u003cpure.logic@nexus-software.ie\u003e\nTested-by: Bryan O\u0027Donoghue \u003cpure.logic@nexus-software.ie\u003e\nCc: Ed Blake \u003ced.blake@imgtec.com\u003e\nCc: linux-mips@linux-mips.org\nPatchwork: https://patchwork.linux-mips.org/patch/16363/\nSigned-off-by: Ralf Baechle \u003cralf@linux-mips.org\u003e\n" }, { "commit": "d8550860d910c6b7b70f830f59003b33daaa52c9", "tree": "1bb4c74af466c3e5787d2f9c5b571736a3ca8082", "parents": [ "161c51ccb7a6faf45ffe09aa5cf1ad85ccdad503" ], "author": { "name": "Paul Burton", "email": "paul.burton@imgtec.com", "time": "Fri Mar 03 15:26:05 2017 -0800" }, "committer": { "name": "Ralf Baechle", "email": "ralf@linux-mips.org", "time": "Fri Jun 30 04:40:18 2017 +0200" }, "message": "MIPS: Fix IRQ tracing \u0026 lockdep when rescheduling\n\nWhen the scheduler sets TIF_NEED_RESCHED \u0026 we call into the scheduler\nfrom arch/mips/kernel/entry.S we disable interrupts. This is true\nregardless of whether we reach work_resched from syscall_exit_work,\nresume_userspace or by looping after calling schedule(). Although we\ndisable interrupts in these paths we don\u0027t call trace_hardirqs_off()\nbefore calling into C code which may acquire locks, and we therefore\nleave lockdep with an inconsistent view of whether interrupts are\ndisabled or not when CONFIG_PROVE_LOCKING \u0026 CONFIG_DEBUG_LOCKDEP are\nboth enabled.\n\nWithout tracing this interrupt state lockdep will print warnings such\nas the following once a task returns from a syscall via\nsyscall_exit_partial with TIF_NEED_RESCHED set:\n\n[ 49.927678] ------------[ cut here ]------------\n[ 49.934445] WARNING: CPU: 0 PID: 1 at kernel/locking/lockdep.c:3687 check_flags.part.41+0x1dc/0x1e8\n[ 49.946031] DEBUG_LOCKS_WARN_ON(current-\u003ehardirqs_enabled)\n[ 49.946355] CPU: 0 PID: 1 Comm: init Not tainted 4.10.0-00439-gc9fd5d362289-dirty #197\n[ 49.963505] Stack : 0000000000000000 ffffffff81bb5d6a 0000000000000006 ffffffff801ce9c4\n[ 49.974431] 0000000000000000 0000000000000000 0000000000000000 000000000000004a\n[ 49.985300] ffffffff80b7e487 ffffffff80a24498 a8000000ff160000 ffffffff80ede8b8\n[ 49.996194] 0000000000000001 0000000000000000 0000000000000000 0000000077c8030c\n[ 50.007063] 000000007fd8a510 ffffffff801cd45c 0000000000000000 a8000000ff127c88\n[ 50.017945] 0000000000000000 ffffffff801cf928 0000000000000001 ffffffff80a24498\n[ 50.028827] 0000000000000000 0000000000000001 0000000000000000 0000000000000000\n[ 50.039688] 0000000000000000 a8000000ff127bd0 0000000000000000 ffffffff805509bc\n[ 50.050575] 00000000140084e0 0000000000000000 0000000000000000 0000000000040a00\n[ 50.061448] 0000000000000000 ffffffff8010e1b0 0000000000000000 ffffffff805509bc\n[ 50.072327] ...\n[ 50.076087] Call Trace:\n[ 50.079869] [\u003cffffffff8010e1b0\u003e] show_stack+0x80/0xa8\n[ 50.086577] [\u003cffffffff805509bc\u003e] dump_stack+0x10c/0x190\n[ 50.093498] [\u003cffffffff8015dde0\u003e] __warn+0xf0/0x108\n[ 50.099889] [\u003cffffffff8015de34\u003e] warn_slowpath_fmt+0x3c/0x48\n[ 50.107241] [\u003cffffffff801c15b4\u003e] check_flags.part.41+0x1dc/0x1e8\n[ 50.114961] [\u003cffffffff801c239c\u003e] lock_is_held_type+0x8c/0xb0\n[ 50.122291] [\u003cffffffff809461b8\u003e] __schedule+0x8c0/0x10f8\n[ 50.129221] [\u003cffffffff80946a60\u003e] schedule+0x30/0x98\n[ 50.135659] [\u003cffffffff80106278\u003e] work_resched+0x8/0x34\n[ 50.142397] ---[ end trace 0cb4f6ef5b99fe21 ]---\n[ 50.148405] possible reason: unannotated irqs-off.\n[ 50.154600] irq event stamp: 400463\n[ 50.159566] hardirqs last enabled at (400463): [\u003cffffffff8094edc8\u003e] _raw_spin_unlock_irqrestore+0x40/0xa8\n[ 50.171981] hardirqs last disabled at (400462): [\u003cffffffff8094eb98\u003e] _raw_spin_lock_irqsave+0x30/0xb0\n[ 50.183897] softirqs last enabled at (400450): [\u003cffffffff8016580c\u003e] __do_softirq+0x4ac/0x6a8\n[ 50.195015] softirqs last disabled at (400425): [\u003cffffffff80165e78\u003e] irq_exit+0x110/0x128\n\nFix this by using the TRACE_IRQS_OFF macro to call trace_hardirqs_off()\nwhen CONFIG_TRACE_IRQFLAGS is enabled. This is done before invoking\nschedule() following the work_resched label because:\n\n 1) Interrupts are disabled regardless of the path we take to reach\n work_resched() \u0026 schedule().\n\n 2) Performing the tracing here avoids the need to do it in paths which\n disable interrupts but don\u0027t call out to C code before hitting a\n path which uses the RESTORE_SOME macro that will call\n trace_hardirqs_on() or trace_hardirqs_off() as appropriate.\n\nWe call trace_hardirqs_on() using the TRACE_IRQS_ON macro before calling\nsyscall_trace_leave() for similar reasons, ensuring that lockdep has a\nconsistent view of state after we re-enable interrupts.\n\nSigned-off-by: Paul Burton \u003cpaul.burton@imgtec.com\u003e\nFixes: 1da177e4c3f4 (\"Linux-2.6.12-rc2\")\nCc: linux-mips@linux-mips.org\nCc: stable \u003cstable@vger.kernel.org\u003e\nPatchwork: https://patchwork.linux-mips.org/patch/15385/\nSigned-off-by: Ralf Baechle \u003cralf@linux-mips.org\u003e\n" }, { "commit": "161c51ccb7a6faf45ffe09aa5cf1ad85ccdad503", "tree": "2fdcdcb84cdab3befe0a7a8d18d795135dc15b5f", "parents": [ "ddbfff7429a75d954bf5bdff9f2222bceb4c236a" ], "author": { "name": "Paul Burton", "email": "paul.burton@imgtec.com", "time": "Thu Mar 02 14:02:40 2017 -0800" }, "committer": { "name": "Ralf Baechle", "email": "ralf@linux-mips.org", "time": "Fri Jun 30 04:38:55 2017 +0200" }, "message": "MIPS: pm-cps: Drop manual cache-line alignment of ready_count\n\nWe allocate memory for a ready_count variable per-CPU, which is accessed\nvia a cached non-coherent TLB mapping to perform synchronisation between\nthreads within the core using LL/SC instructions. In order to ensure\nthat the variable is contained within its own data cache line we\nallocate 2 lines worth of memory \u0026 align the resulting pointer to a line\nboundary. This is however unnecessary, since kmalloc is guaranteed to\nreturn memory which is at least cache-line aligned (see\nARCH_DMA_MINALIGN). Stop the redundant manual alignment.\n\nBesides cleaning up the code \u0026 avoiding needless work, this has the side\neffect of avoiding an arithmetic error found by Bryan on 64 bit systems\ndue to the 32 bit size of the former dlinesz. This led the ready_count\nvariable to have its upper 32b cleared erroneously for MIPS64 kernels,\ncausing problems when ready_count was later used on MIPS64 via cpuidle.\n\nSigned-off-by: Paul Burton \u003cpaul.burton@imgtec.com\u003e\nFixes: 3179d37ee1ed (\"MIPS: pm-cps: add PM state entry code for CPS systems\")\nReported-by: Bryan O\u0027Donoghue \u003cbryan.odonoghue@imgtec.com\u003e\nReviewed-by: Bryan O\u0027Donoghue \u003cbryan.odonoghue@imgtec.com\u003e\nTested-by: Bryan O\u0027Donoghue \u003cbryan.odonoghue@imgtec.com\u003e\nCc: linux-mips@linux-mips.org\nCc: stable \u003cstable@vger.kernel.org\u003e # v3.16+\nPatchwork: https://patchwork.linux-mips.org/patch/15383/\nSigned-off-by: Ralf Baechle \u003cralf@linux-mips.org\u003e\n" }, { "commit": "9e25ebfe56ece7541cd10a20d715cbdd148a2e06", "tree": "c16253ba6097189912ebe1924d58367b9e4a77c2", "parents": [ "d360a687d99577110c181e67ebfb9a1b6fed63a2" ], "author": { "name": "Doug Berger", "email": "opendmb@gmail.com", "time": "Thu Jun 29 18:41:36 2017 +0100" }, "committer": { "name": "Russell King", "email": "rmk+kernel@armlinux.org.uk", "time": "Thu Jun 29 23:10:12 2017 +0100" }, "message": "ARM: 8685/1: ensure memblock-limit is pmd-aligned\n\nThe pmd containing memblock_limit is cleared by prepare_page_table()\nwhich creates the opportunity for early_alloc() to allocate unmapped\nmemory if memblock_limit is not pmd aligned causing a boot-time hang.\n\nCommit 965278dcb8ab (\"ARM: 8356/1: mm: handle non-pmd-aligned end of RAM\")\nattempted to resolve this problem, but there is a path through the\nadjust_lowmem_bounds() routine where if all memory regions start and\nend on pmd-aligned addresses the memblock_limit will be set to\narm_lowmem_limit.\n\nSince arm_lowmem_limit can be affected by the vmalloc early parameter,\nthe value of arm_lowmem_limit may not be pmd-aligned. This commit\ncorrects this oversight such that memblock_limit is always rounded\ndown to pmd-alignment.\n\nFixes: 965278dcb8ab (\"ARM: 8356/1: mm: handle non-pmd-aligned end of RAM\")\nSigned-off-by: Doug Berger \u003copendmb@gmail.com\u003e\nSuggested-by: Mark Rutland \u003cmark.rutland@arm.com\u003e\nSigned-off-by: Russell King \u003crmk+kernel@armlinux.org.uk\u003e\n" }, { "commit": "4d8a991d460d4fa4829beaffdcba45a217ca0fa7", "tree": "51edc7b944c963780f75a0475683a6fb188c5199", "parents": [ "27bc344014289a298a9fad4b828fcf61daa60aff", "d58299a478c416c0b48e4b31c6332fe7beb63000" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jun 29 14:30:07 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jun 29 14:30:07 2017 -0700" }, "message": "Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net\n\nPull networking fixes from David Miller:\n\n 1) Need to access netdev-\u003enum_rx_queues behind an accessor in netvsc\n driver otherwise the build breaks with some configs, from Arnd\n Bergmann.\n\n 2) Add dummy xfrm_dev_event() so that build doesn\u0027t fail when\n CONFIG_XFRM_OFFLOAD is not set. From Hangbin Liu.\n\n 3) Don\u0027t OOPS when pfkey_msg2xfrm_state() signals an erros, from Dan\n Carpenter.\n\n 4) Fix MCDI command size for filter operations in sfc driver, from\n Martin Habets.\n\n 5) Fix UFO segmenting so that we don\u0027t calculate incorrect checksums,\n from Michal Kubecek.\n\n 6) When ipv6 datagram connects fail, reset destination address and\n port. From Wei Wang.\n\n 7) TCP disconnect must reset the cached receive DST, from WANG Cong.\n\n 8) Fix sign extension bug on 32-bit in dev_get_stats(), from Eric\n Dumazet.\n\n 9) fman driver has to depend on HAS_DMA, from Madalin Bucur.\n\n10) Fix bpf pointer leak with xadd in verifier, from Daniel Borkmann.\n\n11) Fix negative page counts with GFO, from Michal Kubecek.\n\n* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (41 commits)\n sfc: fix attempt to translate invalid filter ID\n net: handle NAPI_GRO_FREE_STOLEN_HEAD case also in napi_frags_finish()\n bpf: prevent leaking pointer via xadd on unpriviledged\n arcnet: com20020-pci: add missing pdev setup in netdev structure\n arcnet: com20020-pci: fix dev_id calculation\n arcnet: com20020: remove needless base_addr assignment\n Trivial fix to spelling mistake in arc_printk message\n arcnet: change irq handler to lock irqsave\n rocker: move dereference before free\n mlxsw: spectrum_router: Fix NULL pointer dereference\n net: sched: Fix one possible panic when no destroy callback\n virtio-net: serialize tx routine during reset\n net: usb: asix88179_178a: Add support for the Belkin B2B128\n fsl/fman: add dependency on HAS_DMA\n net: prevent sign extension in dev_get_stats()\n tcp: reset sk_rx_dst in tcp_disconnect()\n net: ipv6: reset daddr and dport in sk if connect() fails\n bnx2x: Don\u0027t log mc removal needlessly\n bnxt_en: Fix netpoll handling.\n bnxt_en: Add missing logic to handle TPA end error conditions.\n ...\n" }, { "commit": "27bc344014289a298a9fad4b828fcf61daa60aff", "tree": "37b04a28d5afd9e95522889a425b64fae95cd8f2", "parents": [ "374bf8831aa99a445d6d26c56a65fb35db747d91", "00a0ea33b495ee6149bf5a77ac5807ce87323abb" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jun 29 14:23:02 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jun 29 14:23:02 2017 -0700" }, "message": "Merge tag \u0027for-4.12/dm-fixes-5\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm\n\nPull device mapper fixes from Mike Snitzer:\n\n - dm thinp fix for crash that will occur when metadata device failure\n races with discard passdown to the underlying data device.\n\n - dm raid fix to not access the superblock\u0027s \u003e\u003d 1.9.0 \u0027sectors\u0027 member\n unconditionally.\n\n* tag \u0027for-4.12/dm-fixes-5\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm:\n dm thin: do not queue freed thin mapping for next stage processing\n dm raid: fix oops on upgrading to extended superblock format\n" }, { "commit": "374bf8831aa99a445d6d26c56a65fb35db747d91", "tree": "fa7f85706041be3969a2d0397f08f4f933b54a81", "parents": [ "6474924e2b5ddb0030c355558966adcbe3b49022", "9ae3b3f52c62ddd5eb12c57f195f4f38121faa01" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jun 29 14:10:37 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Thu Jun 29 14:10:37 2017 -0700" }, "message": "Merge branch \u0027for-linus\u0027 of git://git.kernel.dk/linux-block\n\nPull block fixes from Jens Axboe:\n \"Two fixes that should go into this release.\n\n One is an nvme regression fix from Keith, fixing a missing queue\n freeze if the controller is being reset. This causes the reset to\n hang.\n\n The other is a fix for a leak of the bio protection info, if smaller\n sized O_DIRECT is used. This fix should be more involved as we have\n other problematic paths in the kernel, but given as this isn\u0027t a\n regression in this series, we\u0027ll tackle those for 4.13\"\n\n* \u0027for-linus\u0027 of git://git.kernel.dk/linux-block:\n block: provide bio_uninit() free freeing integrity/task associations\n nvme/pci: Fix stuck nvme reset\n" }, { "commit": "d58299a478c416c0b48e4b31c6332fe7beb63000", "tree": "7e1c5c96244bdb703fae5368db5f9c146e05e427", "parents": [ "e44699d2c28067f69698ccb68dd3ddeacfebc434" ], "author": { "name": "Edward Cree", "email": "ecree@solarflare.com", "time": "Thu Jun 29 16:50:06 2017 +0100" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Thu Jun 29 15:59:38 2017 -0400" }, "message": "sfc: fix attempt to translate invalid filter ID\n\nWhen filter insertion fails with no rollback, we were trying to convert\n EFX_EF10_FILTER_ID_INVALID to an id to store in \u0027ids\u0027 (which is either\n vlan-\u003euc or vlan-\u003emc). This would WARN_ON_ONCE and then record a bogus\n filter ID of 0x1fff, neither of which is a good thing.\n\nFixes: 0ccb998bf46d (\"sfc: fix filter_id misinterpretation in edge case\")\nSigned-off-by: Edward Cree \u003cecree@solarflare.com\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "e44699d2c28067f69698ccb68dd3ddeacfebc434", "tree": "100559d942c9c465ed1362127051c89b16bcdfa6", "parents": [ "6bdf6abc56b53103324dfd270a86580306e1a232" ], "author": { "name": "Michal Kubeček", "email": "mkubecek@suse.cz", "time": "Thu Jun 29 11:13:36 2017 +0200" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Thu Jun 29 15:54:13 2017 -0400" }, "message": "net: handle NAPI_GRO_FREE_STOLEN_HEAD case also in napi_frags_finish()\n\nRecently I started seeing warnings about pages with refcount -1. The\nproblem was traced to packets being reused after their head was merged into\na GRO packet by skb_gro_receive(). While bisecting the issue pointed to\ncommit c21b48cc1bbf (\"net: adjust skb-\u003etruesize in ___pskb_trim()\") and\nI have never seen it on a kernel with it reverted, I believe the real\nproblem appeared earlier when the option to merge head frag in GRO was\nimplemented.\n\nHandling NAPI_GRO_FREE_STOLEN_HEAD state was only added to GRO_MERGED_FREE\nbranch of napi_skb_finish() so that if the driver uses napi_gro_frags()\nand head is merged (which in my case happens after the skb_condense()\ncall added by the commit mentioned above), the skb is reused including the\nhead that has been merged. As a result, we release the page reference\ntwice and eventually end up with negative page refcount.\n\nTo fix the problem, handle NAPI_GRO_FREE_STOLEN_HEAD in napi_frags_finish()\nthe same way it\u0027s done in napi_skb_finish().\n\nFixes: d7e8883cfcf4 (\"net: make GRO aware of skb-\u003ehead_frag\")\nSigned-off-by: Michal Kubecek \u003cmkubecek@suse.cz\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "6bdf6abc56b53103324dfd270a86580306e1a232", "tree": "2ab300507010ec4dc0895936ae04718984823086", "parents": [ "00778f7cadd393630a23e83ffaf60b70fe460879" ], "author": { "name": "Daniel Borkmann", "email": "daniel@iogearbox.net", "time": "Thu Jun 29 03:04:59 2017 +0200" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Thu Jun 29 15:44:34 2017 -0400" }, "message": "bpf: prevent leaking pointer via xadd on unpriviledged\n\nLeaking kernel addresses on unpriviledged is generally disallowed,\nfor example, verifier rejects the following:\n\n 0: (b7) r0 \u003d 0\n 1: (18) r2 \u003d 0xffff897e82304400\n 3: (7b) *(u64 *)(r1 +48) \u003d r2\n R2 leaks addr into ctx\n\nDoing pointer arithmetic on them is also forbidden, so that they\ndon\u0027t turn into unknown value and then get leaked out. However,\nthere\u0027s xadd as a special case, where we don\u0027t check the src reg\nfor being a pointer register, e.g. the following will pass:\n\n 0: (b7) r0 \u003d 0\n 1: (7b) *(u64 *)(r1 +48) \u003d r0\n 2: (18) r2 \u003d 0xffff897e82304400 ; map\n 4: (db) lock *(u64 *)(r1 +48) +\u003d r2\n 5: (95) exit\n\nWe could store the pointer into skb-\u003ecb, loose the type context,\nand then read it out from there again to leak it eventually out\nof a map value. Or more easily in a different variant, too:\n\n 0: (bf) r6 \u003d r1\n 1: (7a) *(u64 *)(r10 -8) \u003d 0\n 2: (bf) r2 \u003d r10\n 3: (07) r2 +\u003d -8\n 4: (18) r1 \u003d 0x0\n 6: (85) call bpf_map_lookup_elem#1\n 7: (15) if r0 \u003d\u003d 0x0 goto pc+3\n R0\u003dmap_value(ks\u003d8,vs\u003d8,id\u003d0),min_value\u003d0,max_value\u003d0 R6\u003dctx R10\u003dfp\n 8: (b7) r3 \u003d 0\n 9: (7b) *(u64 *)(r0 +0) \u003d r3\n 10: (db) lock *(u64 *)(r0 +0) +\u003d r6\n 11: (b7) r0 \u003d 0\n 12: (95) exit\n\n from 7 to 11: R0\u003dinv,min_value\u003d0,max_value\u003d0 R6\u003dctx R10\u003dfp\n 11: (b7) r0 \u003d 0\n 12: (95) exit\n\nPrevent this by checking xadd src reg for pointer types. Also\nadd a couple of test cases related to this.\n\nFixes: 1be7f75d1668 (\"bpf: enable non-root eBPF programs\")\nFixes: 17a5267067f3 (\"bpf: verifier (add verifier core)\")\nSigned-off-by: Daniel Borkmann \u003cdaniel@iogearbox.net\u003e\nAcked-by: Alexei Starovoitov \u003cast@kernel.org\u003e\nAcked-by: Martin KaFai Lau \u003ckafai@fb.com\u003e\nAcked-by: Edward Cree \u003cecree@solarflare.com\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "80c65fdb4c6920e332a9781a3de5877594b07522", "tree": "7a4d448c21dc2ba5c6e81a2828f449f60dad31f3", "parents": [ "98fe3633c5a44e5ee3d642907739eb0210407886" ], "author": { "name": "Kan Liang", "email": "kan.liang@intel.com", "time": "Thu Jun 29 12:09:26 2017 -0700" }, "committer": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Thu Jun 29 21:28:13 2017 +0200" }, "message": "perf/x86/intel/uncore: Fix wrong box pointer check\n\nShould not init a NULL box. It will cause system crash.\nThe issue looks like caused by a typo.\n\nThis was not noticed because there is no NULL box. Also, for most\nboxes, they are enabled by default. The init code is not critical.\n\nFixes: fff4b87e594a (\"perf/x86/intel/uncore: Make package handling more robust\")\nSigned-off-by: Kan Liang \u003ckan.liang@intel.com\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: stable@vger.kernel.org\nLink: http://lkml.kernel.org/r/20170629190926.2456-1-kan.liang@intel.com\n\n" }, { "commit": "00778f7cadd393630a23e83ffaf60b70fe460879", "tree": "71d298b46dea34756a8273467111f39de854d373", "parents": [ "acb4b7df48b539cb391287921de57e4e5fae3460", "2a0ea04c83ab82c3852c9171d2fa5cd9a1432c9b" ], "author": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Thu Jun 29 15:18:38 2017 -0400" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Thu Jun 29 15:18:38 2017 -0400" }, "message": "Merge branch \u0027arcnet-fixes\u0027\n\nMichael Grzeschik says:\n\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\narcnet: Collection of latest fixes\n\nHere we sum up the recent fixes I collected on the way to use and\nstabilise the framework. Part of it is an possible deadlock that we\nprevent as well to fix the calculation of the dev_id that can be setup\nby an rotary encoder. Beside that we added an trivial spelling patch and\nfix some wrong and missing assignments that improves the code footprint.\n\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "2a0ea04c83ab82c3852c9171d2fa5cd9a1432c9b", "tree": "71d298b46dea34756a8273467111f39de854d373", "parents": [ "cb108619f2fc77846bf7a7543517f3487f455b24" ], "author": { "name": "Michael Grzeschik", "email": "m.grzeschik@pengutronix.de", "time": "Wed Jun 28 18:28:37 2017 +0200" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Thu Jun 29 15:18:37 2017 -0400" }, "message": "arcnet: com20020-pci: add missing pdev setup in netdev structure\n\nWe add the pdev data to the pci devices netdev structure. This way\nthe interface get consistent device names in the userspace (udev).\n\nSigned-off-by: Michael Grzeschik \u003cm.grzeschik@pengutronix.de\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "cb108619f2fc77846bf7a7543517f3487f455b24", "tree": "97efed19db0fb87fcdba34fd04f32afdb32b1878", "parents": [ "0d494fcf867f95040b5b67e4bc5af739bcda37da" ], "author": { "name": "Michael Grzeschik", "email": "m.grzeschik@pengutronix.de", "time": "Wed Jun 28 18:28:36 2017 +0200" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Thu Jun 29 15:18:36 2017 -0400" }, "message": "arcnet: com20020-pci: fix dev_id calculation\n\nThe dev_id was miscalculated. Only the two bits 4-5 are relevant for the\nMA1 card. PCIARC1 and PCIFB2 use the four bits 4-7 for id selection.\n\nSigned-off-by: Michael Grzeschik \u003cm.grzeschik@pengutronix.de\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "0d494fcf867f95040b5b67e4bc5af739bcda37da", "tree": "ddf3bbdea36a2257271e55d241cdd5edd48b1a16", "parents": [ "06908d7aee8d62a80cabfd134d0354dc4d2794bc" ], "author": { "name": "Michael Grzeschik", "email": "m.grzeschik@pengutronix.de", "time": "Wed Jun 28 18:28:35 2017 +0200" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Thu Jun 29 15:18:36 2017 -0400" }, "message": "arcnet: com20020: remove needless base_addr assignment\n\nThe assignment is superfluous.\n\nSigned-off-by: Michael Grzeschik \u003cm.grzeschik@pengutronix.de\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "06908d7aee8d62a80cabfd134d0354dc4d2794bc", "tree": "506184da24c9aeb15f81ed024db533f4e0458fea", "parents": [ "5b85840320151f61e04d83a23ef2567a07094503" ], "author": { "name": "Colin Ian King", "email": "colin.king@canonical.com", "time": "Wed Jun 28 18:28:34 2017 +0200" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Thu Jun 29 15:18:36 2017 -0400" }, "message": "Trivial fix to spelling mistake in arc_printk message\n\nSigned-off-by: Colin Ian King \u003ccolin.king@canonical.com\u003e\nSigned-off-by: Michael Grzeschik \u003cm.grzeschik@pengutronix.de\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "5b85840320151f61e04d83a23ef2567a07094503", "tree": "d0bd1b5ef5b1cd2b0a87c4664f197762f8896c78", "parents": [ "acb4b7df48b539cb391287921de57e4e5fae3460" ], "author": { "name": "Michael Grzeschik", "email": "m.grzeschik@pengutronix.de", "time": "Wed Jun 28 18:28:33 2017 +0200" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Thu Jun 29 15:18:35 2017 -0400" }, "message": "arcnet: change irq handler to lock irqsave\n\nThis patch prevents the arcnet driver from the following deadlock.\n\n[ 41.273910] \u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\n[ 41.280397] [ INFO: SOFTIRQ-safe -\u003e SOFTIRQ-unsafe lock order detected ]\n[ 41.287433] 4.4.0-00034-gc0ae784 #536 Not tainted\n[ 41.292366] ------------------------------------------------------\n[ 41.298863] arcecho/233 [HC0[0]:SC0[2]:HE0:SE0] is trying to acquire:\n[ 41.305628] (\u0026(\u0026lp-\u003elock)-\u003erlock){+.+...}, at: [\u003cbf083bc8\u003e] arcnet_send_packet+0x60/0x1c0 [arcnet]\n[ 41.315199]\n[ 41.315199] and this task is already holding:\n[ 41.321324] (_xmit_ARCNET#2){+.-...}, at: [\u003cc06b934c\u003e] packet_direct_xmit+0xfc/0x1c8\n[ 41.329593] which would create a new lock dependency:\n[ 41.334893] (_xmit_ARCNET#2){+.-...} -\u003e (\u0026(\u0026lp-\u003elock)-\u003erlock){+.+...}\n[ 41.341801]\n[ 41.341801] but this new dependency connects a SOFTIRQ-irq-safe lock:\n[ 41.350108] (_xmit_ARCNET#2){+.-...}\n... which became SOFTIRQ-irq-safe at:\n[ 41.357539] [\u003cc06f8fc8\u003e] _raw_spin_lock+0x30/0x40\n[ 41.362677] [\u003cc063ab8c\u003e] dev_watchdog+0x5c/0x264\n[ 41.367723] [\u003cc0094edc\u003e] call_timer_fn+0x6c/0xf4\n[ 41.372759] [\u003cc00950b8\u003e] run_timer_softirq+0x154/0x210\n[ 41.378340] [\u003cc0036b30\u003e] __do_softirq+0x144/0x298\n[ 41.383469] [\u003cc0036fb4\u003e] irq_exit+0xcc/0x130\n[ 41.388138] [\u003cc0085c50\u003e] __handle_domain_irq+0x60/0xb4\n[ 41.393728] [\u003cc0014578\u003e] __irq_svc+0x58/0x78\n[ 41.398402] [\u003cc0010274\u003e] arch_cpu_idle+0x24/0x3c\n[ 41.403443] [\u003cc007127c\u003e] cpu_startup_entry+0x1f8/0x25c\n[ 41.409029] [\u003cc09adc90\u003e] start_kernel+0x3c0/0x3cc\n[ 41.414170]\n[ 41.414170] to a SOFTIRQ-irq-unsafe lock:\n[ 41.419931] (\u0026(\u0026lp-\u003elock)-\u003erlock){+.+...}\n... which became SOFTIRQ-irq-unsafe at:\n[ 41.427996] ... [\u003cc06f8fc8\u003e] _raw_spin_lock+0x30/0x40\n[ 41.433409] [\u003cbf083d54\u003e] arcnet_interrupt+0x2c/0x800 [arcnet]\n[ 41.439646] [\u003cc0089120\u003e] handle_nested_irq+0x8c/0xec\n[ 41.445063] [\u003cc03c1170\u003e] regmap_irq_thread+0x190/0x314\n[ 41.450661] [\u003cc0087244\u003e] irq_thread_fn+0x1c/0x34\n[ 41.455700] [\u003cc0087548\u003e] irq_thread+0x13c/0x1dc\n[ 41.460649] [\u003cc0050f10\u003e] kthread+0xe4/0xf8\n[ 41.465158] [\u003cc000f810\u003e] ret_from_fork+0x14/0x24\n[ 41.470207]\n[ 41.470207] other info that might help us debug this:\n[ 41.470207]\n[ 41.478627] Possible interrupt unsafe locking scenario:\n[ 41.478627]\n[ 41.485763] CPU0 CPU1\n[ 41.490521] ---- ----\n[ 41.495279] lock(\u0026(\u0026lp-\u003elock)-\u003erlock);\n[ 41.499414] local_irq_disable();\n[ 41.505636] lock(_xmit_ARCNET#2);\n[ 41.511967] lock(\u0026(\u0026lp-\u003elock)-\u003erlock);\n[ 41.518741] \u003cInterrupt\u003e\n[ 41.521490] lock(_xmit_ARCNET#2);\n[ 41.525356]\n[ 41.525356] *** DEADLOCK ***\n[ 41.525356]\n[ 41.531587] 1 lock held by arcecho/233:\n[ 41.535617] #0: (_xmit_ARCNET#2){+.-...}, at: [\u003cc06b934c\u003e] packet_direct_xmit+0xfc/0x1c8\n[ 41.544355]\nthe dependencies between SOFTIRQ-irq-safe lock and the holding lock:\n[ 41.552362] -\u003e (_xmit_ARCNET#2){+.-...} ops: 27 {\n[ 41.557357] HARDIRQ-ON-W at:\n[ 41.560664] [\u003cc06f8fc8\u003e] _raw_spin_lock+0x30/0x40\n[ 41.567445] [\u003cc063ba28\u003e] dev_deactivate_many+0x114/0x304\n[ 41.574866] [\u003cc063bc3c\u003e] dev_deactivate+0x24/0x38\n[ 41.581646] [\u003cc0630374\u003e] linkwatch_do_dev+0x40/0x74\n[ 41.588613] [\u003cc06305d8\u003e] __linkwatch_run_queue+0xec/0x140\n[ 41.596120] [\u003cc0630658\u003e] linkwatch_event+0x2c/0x34\n[ 41.602991] [\u003cc004af30\u003e] process_one_work+0x188/0x40c\n[ 41.610131] [\u003cc004b200\u003e] worker_thread+0x4c/0x480\n[ 41.616912] [\u003cc0050f10\u003e] kthread+0xe4/0xf8\n[ 41.623048] [\u003cc000f810\u003e] ret_from_fork+0x14/0x24\n[ 41.629735] IN-SOFTIRQ-W at:\n[ 41.633039] [\u003cc06f8fc8\u003e] _raw_spin_lock+0x30/0x40\n[ 41.639820] [\u003cc063ab8c\u003e] dev_watchdog+0x5c/0x264\n[ 41.646508] [\u003cc0094edc\u003e] call_timer_fn+0x6c/0xf4\n[ 41.653190] [\u003cc00950b8\u003e] run_timer_softirq+0x154/0x210\n[ 41.660425] [\u003cc0036b30\u003e] __do_softirq+0x144/0x298\n[ 41.667201] [\u003cc0036fb4\u003e] irq_exit+0xcc/0x130\n[ 41.673518] [\u003cc0085c50\u003e] __handle_domain_irq+0x60/0xb4\n[ 41.680754] [\u003cc0014578\u003e] __irq_svc+0x58/0x78\n[ 41.687077] [\u003cc0010274\u003e] arch_cpu_idle+0x24/0x3c\n[ 41.693769] [\u003cc007127c\u003e] cpu_startup_entry+0x1f8/0x25c\n[ 41.701006] [\u003cc09adc90\u003e] start_kernel+0x3c0/0x3cc\n[ 41.707791] INITIAL USE at:\n[ 41.711003] [\u003cc06f8fc8\u003e] _raw_spin_lock+0x30/0x40\n[ 41.717696] [\u003cc063ba28\u003e] dev_deactivate_many+0x114/0x304\n[ 41.725026] [\u003cc063bc3c\u003e] dev_deactivate+0x24/0x38\n[ 41.731718] [\u003cc0630374\u003e] linkwatch_do_dev+0x40/0x74\n[ 41.738593] [\u003cc06305d8\u003e] __linkwatch_run_queue+0xec/0x140\n[ 41.746011] [\u003cc0630658\u003e] linkwatch_event+0x2c/0x34\n[ 41.752789] [\u003cc004af30\u003e] process_one_work+0x188/0x40c\n[ 41.759847] [\u003cc004b200\u003e] worker_thread+0x4c/0x480\n[ 41.766541] [\u003cc0050f10\u003e] kthread+0xe4/0xf8\n[ 41.772596] [\u003cc000f810\u003e] ret_from_fork+0x14/0x24\n[ 41.779198] }\n[ 41.780945] ... key at: [\u003cc124d620\u003e] netdev_xmit_lock_key+0x38/0x1c8\n[ 41.788192] ... acquired at:\n[ 41.791309] [\u003cc007bed8\u003e] lock_acquire+0x70/0x90\n[ 41.796361] [\u003cc06f9140\u003e] _raw_spin_lock_irqsave+0x40/0x54\n[ 41.802324] [\u003cbf083bc8\u003e] arcnet_send_packet+0x60/0x1c0 [arcnet]\n[ 41.808844] [\u003cc06b9380\u003e] packet_direct_xmit+0x130/0x1c8\n[ 41.814622] [\u003cc06bc7e4\u003e] packet_sendmsg+0x3b8/0x680\n[ 41.820034] [\u003cc05fe8b0\u003e] sock_sendmsg+0x14/0x24\n[ 41.825091] [\u003cc05ffd68\u003e] SyS_sendto+0xb8/0xe0\n[ 41.829956] [\u003cc05ffda8\u003e] SyS_send+0x18/0x20\n[ 41.834638] [\u003cc000f780\u003e] ret_fast_syscall+0x0/0x1c\n[ 41.839954]\n[ 41.841514]\nthe dependencies between the lock to be acquired and SOFTIRQ-irq-unsafe lock:\n[ 41.850302] -\u003e (\u0026(\u0026lp-\u003elock)-\u003erlock){+.+...} ops: 5 {\n[ 41.855644] HARDIRQ-ON-W at:\n[ 41.858945] [\u003cc06f8fc8\u003e] _raw_spin_lock+0x30/0x40\n[ 41.865726] [\u003cbf083d54\u003e] arcnet_interrupt+0x2c/0x800 [arcnet]\n[ 41.873607] [\u003cc0089120\u003e] handle_nested_irq+0x8c/0xec\n[ 41.880666] [\u003cc03c1170\u003e] regmap_irq_thread+0x190/0x314\n[ 41.887901] [\u003cc0087244\u003e] irq_thread_fn+0x1c/0x34\n[ 41.894593] [\u003cc0087548\u003e] irq_thread+0x13c/0x1dc\n[ 41.901195] [\u003cc0050f10\u003e] kthread+0xe4/0xf8\n[ 41.907338] [\u003cc000f810\u003e] ret_from_fork+0x14/0x24\n[ 41.914025] SOFTIRQ-ON-W at:\n[ 41.917328] [\u003cc06f8fc8\u003e] _raw_spin_lock+0x30/0x40\n[ 41.924106] [\u003cbf083d54\u003e] arcnet_interrupt+0x2c/0x800 [arcnet]\n[ 41.931981] [\u003cc0089120\u003e] handle_nested_irq+0x8c/0xec\n[ 41.939028] [\u003cc03c1170\u003e] regmap_irq_thread+0x190/0x314\n[ 41.946264] [\u003cc0087244\u003e] irq_thread_fn+0x1c/0x34\n[ 41.952954] [\u003cc0087548\u003e] irq_thread+0x13c/0x1dc\n[ 41.959548] [\u003cc0050f10\u003e] kthread+0xe4/0xf8\n[ 41.965689] [\u003cc000f810\u003e] ret_from_fork+0x14/0x24\n[ 41.972379] INITIAL USE at:\n[ 41.975595] [\u003cc06f8fc8\u003e] _raw_spin_lock+0x30/0x40\n[ 41.982283] [\u003cbf083d54\u003e] arcnet_interrupt+0x2c/0x800 [arcnet]\n[ 41.990063] [\u003cc0089120\u003e] handle_nested_irq+0x8c/0xec\n[ 41.997027] [\u003cc03c1170\u003e] regmap_irq_thread+0x190/0x314\n[ 42.004172] [\u003cc0087244\u003e] irq_thread_fn+0x1c/0x34\n[ 42.010766] [\u003cc0087548\u003e] irq_thread+0x13c/0x1dc\n[ 42.017267] [\u003cc0050f10\u003e] kthread+0xe4/0xf8\n[ 42.023314] [\u003cc000f810\u003e] ret_from_fork+0x14/0x24\n[ 42.029903] }\n[ 42.031648] ... key at: [\u003cbf0854cc\u003e] __key.42091+0x0/0xfffff0f8 [arcnet]\n[ 42.039255] ... acquired at:\n[ 42.042372] [\u003cc007bed8\u003e] lock_acquire+0x70/0x90\n[ 42.047413] [\u003cc06f9140\u003e] _raw_spin_lock_irqsave+0x40/0x54\n[ 42.053364] [\u003cbf083bc8\u003e] arcnet_send_packet+0x60/0x1c0 [arcnet]\n[ 42.059872] [\u003cc06b9380\u003e] packet_direct_xmit+0x130/0x1c8\n[ 42.065634] [\u003cc06bc7e4\u003e] packet_sendmsg+0x3b8/0x680\n[ 42.071030] [\u003cc05fe8b0\u003e] sock_sendmsg+0x14/0x24\n[ 42.076069] [\u003cc05ffd68\u003e] SyS_sendto+0xb8/0xe0\n[ 42.080926] [\u003cc05ffda8\u003e] SyS_send+0x18/0x20\n[ 42.085601] [\u003cc000f780\u003e] ret_fast_syscall+0x0/0x1c\n[ 42.090918]\n[ 42.092481]\n[ 42.092481] stack backtrace:\n[ 42.097065] CPU: 0 PID: 233 Comm: arcecho Not tainted 4.4.0-00034-gc0ae784 #536\n[ 42.104751] Hardware name: Generic AM33XX (Flattened Device Tree)\n[ 42.111183] [\u003cc0017ec8\u003e] (unwind_backtrace) from [\u003cc00139d0\u003e] (show_stack+0x10/0x14)\n[ 42.119337] [\u003cc00139d0\u003e] (show_stack) from [\u003cc02a82c4\u003e] (dump_stack+0x8c/0x9c)\n[ 42.126937] [\u003cc02a82c4\u003e] (dump_stack) from [\u003cc0078260\u003e] (check_usage+0x4bc/0x63c)\n[ 42.134815] [\u003cc0078260\u003e] (check_usage) from [\u003cc0078438\u003e] (check_irq_usage+0x58/0xb0)\n[ 42.142964] [\u003cc0078438\u003e] (check_irq_usage) from [\u003cc007aaa0\u003e] (__lock_acquire+0x1524/0x20b0)\n[ 42.151740] [\u003cc007aaa0\u003e] (__lock_acquire) from [\u003cc007bed8\u003e] (lock_acquire+0x70/0x90)\n[ 42.159886] [\u003cc007bed8\u003e] (lock_acquire) from [\u003cc06f9140\u003e] (_raw_spin_lock_irqsave+0x40/0x54)\n[ 42.168768] [\u003cc06f9140\u003e] (_raw_spin_lock_irqsave) from [\u003cbf083bc8\u003e] (arcnet_send_packet+0x60/0x1c0 [arcnet])\n[ 42.179115] [\u003cbf083bc8\u003e] (arcnet_send_packet [arcnet]) from [\u003cc06b9380\u003e] (packet_direct_xmit+0x130/0x1c8)\n[ 42.189182] [\u003cc06b9380\u003e] (packet_direct_xmit) from [\u003cc06bc7e4\u003e] (packet_sendmsg+0x3b8/0x680)\n[ 42.198059] [\u003cc06bc7e4\u003e] (packet_sendmsg) from [\u003cc05fe8b0\u003e] (sock_sendmsg+0x14/0x24)\n[ 42.206199] [\u003cc05fe8b0\u003e] (sock_sendmsg) from [\u003cc05ffd68\u003e] (SyS_sendto+0xb8/0xe0)\n[ 42.213978] [\u003cc05ffd68\u003e] (SyS_sendto) from [\u003cc05ffda8\u003e] (SyS_send+0x18/0x20)\n[ 42.221388] [\u003cc05ffda8\u003e] (SyS_send) from [\u003cc000f780\u003e] (ret_fast_syscall+0x0/0x1c)\n\nSigned-off-by: Michael Grzeschik \u003cm.grzeschik@pengutronix.de\u003e\n\n ---\n v1 -\u003e v2: removed unneeded zero assignment of flags\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "acb4b7df48b539cb391287921de57e4e5fae3460", "tree": "597b43a9f7267cf3afad24dd89182f369e925752", "parents": [ "6b27c8adf27edf1dabe2cdcfaa101ef7e2712415" ], "author": { "name": "Dan Carpenter", "email": "dan.carpenter@oracle.com", "time": "Wed Jun 28 14:44:21 2017 +0300" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Thu Jun 29 14:19:24 2017 -0400" }, "message": "rocker: move dereference before free\n\nMy static checker complains that ofdpa_neigh_del() can sometimes free\n\"found\". It just makes sense to use it first before deleting it.\n\nFixes: ecf244f753e0 (\"rocker: fix maybe-uninitialized warning\")\nSigned-off-by: Dan Carpenter \u003cdan.carpenter@oracle.com\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "6b27c8adf27edf1dabe2cdcfaa101ef7e2712415", "tree": "26b90658eccc357a4fbf4e734f5a1e745b3e5722", "parents": [ "c1a4872ebfb83b1af7144f7b29ac8c4b344a12a8" ], "author": { "name": "Ido Schimmel", "email": "idosch@mellanox.com", "time": "Wed Jun 28 09:03:12 2017 +0300" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Thu Jun 29 12:59:48 2017 -0400" }, "message": "mlxsw: spectrum_router: Fix NULL pointer dereference\n\nIn case a VLAN device is enslaved to a bridge we shouldn\u0027t create a\nrouter interface (RIF) for it when it\u0027s configured with an IP address.\nThis is already handled by the driver for other types of netdevs, such\nas physical ports and LAG devices.\n\nIf this IP address is then removed and the interface is subsequently\nunlinked from the bridge, a NULL pointer dereference can happen, as the\noriginal 802.1d FID was replaced with an rFID which was then deleted.\n\nTo reproduce:\n$ ip link set dev enp3s0np9 up\n$ ip link add name enp3s0np9.111 link enp3s0np9 type vlan id 111\n$ ip link set dev enp3s0np9.111 up\n$ ip link add name br0 type bridge\n$ ip link set dev br0 up\n$ ip link set enp3s0np9.111 master br0\n$ ip address add dev enp3s0np9.111 192.168.0.1/24\n$ ip address del dev enp3s0np9.111 192.168.0.1/24\n$ ip link set dev enp3s0np9.111 nomaster\n\nFixes: 99724c18fc66 (\"mlxsw: spectrum: Introduce support for router interfaces\")\nSigned-off-by: Ido Schimmel \u003cidosch@mellanox.com\u003e\nReported-by: Petr Machata \u003cpetrm@mellanox.com\u003e\nTested-by: Petr Machata \u003cpetrm@mellanox.com\u003e\nReviewed-by: Petr Machata \u003cpetrm@mellanox.com\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "c1a4872ebfb83b1af7144f7b29ac8c4b344a12a8", "tree": "bfd7474b87652f2ddbe4f2d6583fe081ebdb0c35", "parents": [ "713a98d90c5ea072c1bb00ef40617aee2cef2232" ], "author": { "name": "Gao Feng", "email": "gfree.wind@vip.163.com", "time": "Wed Jun 28 12:53:54 2017 +0800" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Thu Jun 29 12:55:12 2017 -0400" }, "message": "net: sched: Fix one possible panic when no destroy callback\n\nWhen qdisc fail to init, qdisc_create would invoke the destroy callback\nto cleanup. But there is no check if the callback exists really. So it\nwould cause the panic if there is no real destroy callback like the qdisc\ncodel, fq, and so on.\n\nTake codel as an example following:\nWhen a malicious user constructs one invalid netlink msg, it would cause\ncodel_init-\u003ecodel_change-\u003enla_parse_nested failed.\nThen kernel would invoke the destroy callback directly but qdisc codel\ndoesn\u0027t define one. It causes one panic as a result.\n\nNow add one the check for destroy to avoid the possible panic.\n\nFixes: 87b60cfacf9f (\"net_sched: fix error recovery at qdisc creation\")\nSigned-off-by: Gao Feng \u003cgfree.wind@vip.163.com\u003e\nAcked-by: Eric Dumazet \u003cedumazet@google.com\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "713a98d90c5ea072c1bb00ef40617aee2cef2232", "tree": "db93479ad5f5a93653fd59862d45744d7472ed12", "parents": [ "e20bd60bf62a2448be873653c7febca1d4d73afc" ], "author": { "name": "Jason Wang", "email": "jasowang@redhat.com", "time": "Wed Jun 28 09:51:03 2017 +0800" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Thu Jun 29 12:51:59 2017 -0400" }, "message": "virtio-net: serialize tx routine during reset\n\nWe don\u0027t hold any tx lock when trying to disable TX during reset, this\nwould lead a use after free since ndo_start_xmit() tries to access\nthe virtqueue which has already been freed. Fix this by using\nnetif_tx_disable() before freeing the vqs, this could make sure no tx\nafter vq freeing.\n\nReported-by: Jean-Philippe Menil \u003cjpmenil@gmail.com\u003e\nTested-by: Jean-Philippe Menil \u003cjpmenil@gmail.com\u003e\nFixes commit f600b6905015 (\"virtio_net: Add XDP support\")\nCc: John Fastabend \u003cjohn.fastabend@gmail.com\u003e\nSigned-off-by: Jason Wang \u003cjasowang@redhat.com\u003e\nAcked-by: Michael S. Tsirkin \u003cmst@redhat.com\u003e\nAcked-by: Robert McCabe \u003crobert.mccabe@rockwellcollins.com\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "0f17976568b3f72e676450af0c0db6f8752253d6", "tree": "6bd68d4113163f5205b7ee89a36b162e92b13d20", "parents": [ "32c1431eea4881a6b17bd7c639315010aeefa452" ], "author": { "name": "Steven Rostedt (VMware)", "email": "rostedt@goodmis.org", "time": "Thu Jun 29 10:05:45 2017 -0400" }, "committer": { "name": "Steven Rostedt (VMware)", "email": "rostedt@goodmis.org", "time": "Thu Jun 29 10:05:45 2017 -0400" }, "message": "ftrace: Fix regression with module command in stack_trace_filter\n\nWhen doing the following command:\n\n # echo \":mod:kvm_intel\" \u003e /sys/kernel/tracing/stack_trace_filter\n\nit triggered a crash.\n\nThis happened with the clean up of probes. It required all callers to the\nregex function (doing ftrace filtering) to have ops-\u003eprivate be a pointer to\na trace_array. But for the stack tracer, that is not the case.\n\nAllow for the ops-\u003eprivate to be NULL, and change the function command\ncallbacks to handle the trace_array pointer being NULL as well.\n\nFixes: d2afd57a4b96 (\"tracing/ftrace: Allow instances to have their own function probes\")\nSigned-off-by: Steven Rostedt (VMware) \u003crostedt@goodmis.org\u003e\n" }, { "commit": "1d80df93d9d170eba8c3654719e5857307efd077", "tree": "8dd1fa59a79bb26ad0371e944680ade98a252aff", "parents": [ "c0bc126f97fb929b3ae02c1c62322645d70eb408" ], "author": { "name": "Brian Norris", "email": "briannorris@chromium.org", "time": "Fri Jun 23 13:59:11 2017 -0700" }, "committer": { "name": "Linus Walleij", "email": "linus.walleij@linaro.org", "time": "Thu Jun 29 15:03:24 2017 +0200" }, "message": "Revert \"pinctrl: rockchip: avoid hardirq-unsafe functions in irq_chip\"\n\nThis reverts commit 88bb94216f59e10802aaf78c858a4146085faf18.\n\nIt introduced a new CONFIG_DEBUG_ATOMIC_SLEEP warning in v4.12-rc1:\n\n[ 7226.716713] BUG: sleeping function called from invalid context at kernel/locking/mutex.c:238\n[ 7226.716716] in_atomic(): 0, irqs_disabled(): 0, pid: 1708, name: bash\n[ 7226.716722] CPU: 1 PID: 1708 Comm: bash Not tainted 4.12.0-rc6+ #1213\n[ 7226.716724] Hardware name: Google Kevin (DT)\n[ 7226.716726] Call trace:\n[ 7226.716738] [\u003cffffff8008089928\u003e] dump_backtrace+0x0/0x24c\n[ 7226.716743] [\u003cffffff8008089b94\u003e] show_stack+0x20/0x28\n[ 7226.716749] [\u003cffffff8008371370\u003e] dump_stack+0x90/0xb0\n[ 7226.716755] [\u003cffffff80080cd2a0\u003e] ___might_sleep+0x10c/0x124\n[ 7226.716760] [\u003cffffff80080cd330\u003e] __might_sleep+0x78/0x88\n[ 7226.716765] [\u003cffffff800879e210\u003e] mutex_lock+0x2c/0x64\n[ 7226.716771] [\u003cffffff80083ad678\u003e] rockchip_irq_bus_lock+0x30/0x3c\n[ 7226.716777] [\u003cffffff80080f6d40\u003e] __irq_get_desc_lock+0x78/0x98\n[ 7226.716782] [\u003cffffff80080f7e6c\u003e] irq_set_irq_wake+0x44/0x12c\n[ 7226.716787] [\u003cffffff8008486e18\u003e] dev_pm_arm_wake_irq+0x4c/0x58\n[ 7226.716792] [\u003cffffff800848b80c\u003e] device_wakeup_arm_wake_irqs+0x3c/0x58\n[ 7226.716796] [\u003cffffff80084896fc\u003e] dpm_suspend_noirq+0xf8/0x3a0\n[ 7226.716800] [\u003cffffff80080f1384\u003e] suspend_devices_and_enter+0x1a4/0x9a8\n[ 7226.716803] [\u003cffffff80080f21ec\u003e] pm_suspend+0x664/0x6a4\n[ 7226.716807] [\u003cffffff80080f04d8\u003e] state_store+0xd4/0xf8\n...\n\nIt was reported on -rc1, and it\u0027s still not fixed in -rc6, so it should\njust be reverted.\n\nCc: John Keeping \u003cjohn@metanate.com\u003e\nSigned-off-by: Brian Norris \u003cbriannorris@chromium.org\u003e\nReviewed-by: Heiko Stuebner \u003cheiko@sntech.de\u003e\nSigned-off-by: Linus Walleij \u003clinus.walleij@linaro.org\u003e\n" }, { "commit": "c06632ea054c49510efacb42c52aab693c45b7ba", "tree": "2812bbdf3279ca1062fd6d06985d5bf77964f7fa", "parents": [ "ad537b822577fcc143325786cd6ad50d7b9df31c" ], "author": { "name": "Hans de Goede", "email": "hdegoede@redhat.com", "time": "Fri Jun 23 09:26:13 2017 +0200" }, "committer": { "name": "Linus Walleij", "email": "linus.walleij@linaro.org", "time": "Thu Jun 29 14:55:08 2017 +0200" }, "message": "gpio: acpi: Skip _AEI entries without a handler rather then aborting the scan\n\nacpi_walk_resources will stop as soon as the callback passed in returns\nan error status. On a x86 tablet I have the first GpioInt in the _AEI\nresource list has no handler defined in the DSDT, causing\nacpi_walk_resources to abort scanning the rest of the resource list,\nwhich does define valid ACPI GPIO events.\n\nThis commit changes the return for not finding a handler from\nAE_BAD_PARAMETER to AE_OK so that the rest of the resource list will\nget scanned normally in case of missing event handlers.\n\nSigned-off-by: Hans de Goede \u003chdegoede@redhat.com\u003e\nAcked-by: Mika Westerberg \u003cmika.westerberg@linux.intel.com\u003e\nAcked-by: Andy Shevchenko \u003candriy.shevchenko@linux.intel.com\u003e\nSigned-off-by: Linus Walleij \u003clinus.walleij@linaro.org\u003e\n" }, { "commit": "ad537b822577fcc143325786cd6ad50d7b9df31c", "tree": "0894a0efdc66032a12b792e2e40d0506732cde83", "parents": [ "c0bc126f97fb929b3ae02c1c62322645d70eb408" ], "author": { "name": "Bartosz Golaszewski", "email": "brgl@bgdev.pl", "time": "Fri Jun 23 13:45:16 2017 +0200" }, "committer": { "name": "Linus Walleij", "email": "linus.walleij@linaro.org", "time": "Thu Jun 29 11:33:46 2017 +0200" }, "message": "gpiolib: fix filtering out unwanted events\n\nGPIOEVENT_REQUEST_BOTH_EDGES is not a single flag, but a binary OR of\nGPIOEVENT_REQUEST_RISING_EDGE and GPIOEVENT_REQUEST_FALLING_EDGE.\n\nThe expression \u0027le-\u003eeflags \u0026 GPIOEVENT_REQUEST_BOTH_EDGES\u0027 we\u0027ll get\nevaluated to true even if only one event type was requested.\n\nFix it by checking both RISING \u0026 FALLING flags explicitly.\n\nCc: stable@vger.kernel.org\nFixes: 61f922db7221 (\"gpio: userspace ABI for reading GPIO line events\")\nSigned-off-by: Bartosz Golaszewski \u003cbrgl@bgdev.pl\u003e\nSigned-off-by: Linus Walleij \u003clinus.walleij@linaro.org\u003e\n" }, { "commit": "6474924e2b5ddb0030c355558966adcbe3b49022", "tree": "c217bdbd7999bd6a9d45e220a5f6efdd828f8df3", "parents": [ "e547204f1f4f314a7352f7b93332fdc1a2e4b24c" ], "author": { "name": "Tobias Klauser", "email": "tklauser@distanz.ch", "time": "Wed Jun 28 15:30:02 2017 +0200" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Jun 28 16:13:57 2017 -0700" }, "message": "arch: remove unused macro/function thread_saved_pc()\n\nThe only user of thread_saved_pc() in non-arch-specific code was removed\nin commit 8243d5597793 (\"sched/core: Remove pointless printout in\nsched_show_task()\"). Remove the implementations as well.\n\nSome architectures use thread_saved_pc() in their arch-specific code.\nLeave their thread_saved_pc() intact.\n\nSigned-off-by: Tobias Klauser \u003ctklauser@distanz.ch\u003e\nAcked-by: Geert Uytterhoeven \u003cgeert@linux-m68k.org\u003e\nCc: Ingo Molnar \u003cmingo@kernel.org\u003e\nSigned-off-by: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\n" }, { "commit": "9ae3b3f52c62ddd5eb12c57f195f4f38121faa01", "tree": "19a4a6c51bc46e32f80237253b7c4ad8605ebb78", "parents": [ "ebef7368571d88f0f80b817e6898075c62265b4e" ], "author": { "name": "Jens Axboe", "email": "axboe@kernel.dk", "time": "Wed Jun 28 15:30:13 2017 -0600" }, "committer": { "name": "Jens Axboe", "email": "axboe@kernel.dk", "time": "Wed Jun 28 15:30:13 2017 -0600" }, "message": "block: provide bio_uninit() free freeing integrity/task associations\n\nWen reports significant memory leaks with DIF and O_DIRECT:\n\n\"With nvme devive + T10 enabled, On a system it has 256GB and started\nlogging /proc/meminfo \u0026 /proc/slabinfo for every minute and in an hour\nit increased by 15968128 kB or ~15+GB.. Approximately 256 MB / minute\nleaking.\n\n/proc/meminfo | grep SUnreclaim...\n\nSUnreclaim: 6752128 kB\nSUnreclaim: 6874880 kB\nSUnreclaim: 7238080 kB\n....\nSUnreclaim: 22307264 kB\nSUnreclaim: 22485888 kB\nSUnreclaim: 22720256 kB\n\nWhen testcases with T10 enabled call into __blkdev_direct_IO_simple,\ncode doesn\u0027t free memory allocated by bio_integrity_alloc. The patch\nfixes the issue. HTX has been run with +60 hours without failure.\"\n\nSince __blkdev_direct_IO_simple() allocates the bio on the stack, it\ndoesn\u0027t go through the regular bio free. This means that any ancillary\ndata allocated with the bio through the stack is not freed. Hence, we\ncan leak the integrity data associated with the bio, if the device is\nusing DIF/DIX.\n\nFix this by providing a bio_uninit() and export it, so that we can use\nit to free this data. Note that this is a minimal fix for this issue.\nAny current user of bio\u0027s that are allocated outside of\nbio_alloc_bioset() suffers from this issue, most notably some drivers.\nWe will fix those in a more comprehensive patch for 4.13. This also\nmeans that the commit marked as being fixed by this isn\u0027t the real\nculprit, it\u0027s just the most obvious one out there.\n\nFixes: 542ff7bf18c6 (\"block: new direct I/O implementation\")\nReported-by: Wen Xiong \u003cwenxiong@linux.vnet.ibm.com\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Jens Axboe \u003caxboe@kernel.dk\u003e\n" }, { "commit": "e547204f1f4f314a7352f7b93332fdc1a2e4b24c", "tree": "2a6d1cdd9224b4a402f479c8d3b25944ae0a111e", "parents": [ "5a37be4b51bbf85e37cdb5831899c330505faf03", "2e31b4cb895ae78db31dffb860cd255d86c6561c" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Jun 28 13:27:15 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Jun 28 13:27:15 2017 -0700" }, "message": "Merge tag \u0027nfs-for-4.12-3\u0027 of git://git.linux-nfs.org/projects/trondmy/linux-nfs\n\nPull NFS client bugfixes from Trond Myklebust:\n \"Bugfixes include:\n\n - stable fix for exclusive create if the server supports the umask\n attribute\n\n - trunking detection should handle ERESTARTSYS/EINTR\n\n - stable fix for a race in the LAYOUTGET function\n\n - stable fix to revert \"nfs_rename() handle -ERESTARTSYS dentry left\n behind\"\n\n - nfs4_callback_free_slot() cannot call nfs4_slot_tbl_drain_complete()\"\n\n* tag \u0027nfs-for-4.12-3\u0027 of git://git.linux-nfs.org/projects/trondmy/linux-nfs:\n NFSv4.1: nfs4_callback_free_slot() cannot call nfs4_slot_tbl_drain_complete()\n Revert \"NFS: nfs_rename() handle -ERESTARTSYS dentry left behind\"\n NFSv4.1: Fix a race in nfs4_proc_layoutget\n NFS: Trunking detection should handle ERESTARTSYS/EINTR\n NFSv4.2: Don\u0027t send mode again in post-EXCLUSIVE4_1 SETATTR with umask\n" }, { "commit": "5a37be4b51bbf85e37cdb5831899c330505faf03", "tree": "24ec48d353a9c6f1169f454357c7fdd59a0b2ca1", "parents": [ "cf723497f2b0dfd688dd27949f1d6e26a5aec44c", "9ff1beb1d19ffe2b26bf9cd2d33e6073d4f4b5fe" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Jun 28 13:22:26 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Jun 28 13:22:26 2017 -0700" }, "message": "Merge branch \u0027drm-fixes\u0027 of git://people.freedesktop.org/~airlied/linux\n\nPull drm fixes from Dave Airlie:\n \"This is the final set of fixes for -rc8, just a few i915 and one\n vmwgfx ones.\n\n I\u0027m off on holidays for a week, so if anything shows up for fixes I\u0027ve\n asked Daniel or Sean Paul to herd it in the right direction\"\n\n[ The additional etnaviv fixes were already herded towards me as seen in\n my previous pull - Linus ]\n\n* \u0027drm-fixes\u0027 of git://people.freedesktop.org/~airlied/linux:\n drm/vmwgfx: Free hash table allocated by cmdbuf managed res mgr\n drm/i915: Disable EXEC_OBJECT_ASYNC when doing relocations\n drm/i915: Hold struct_mutex for per-file stats in debugfs/i915_gem_object\n drm/i915: Retire the VMA\u0027s fence tracker before unbinding\n" }, { "commit": "cf723497f2b0dfd688dd27949f1d6e26a5aec44c", "tree": "f1c760a700ff5fe97f9169eace297b923009a237", "parents": [ "3c2bfbaadff6e0c257bb6b16c9c97f43618b13dc", "426ef1bb40a3bd99ccd16d629bd1294805e96fc6" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Jun 28 13:13:48 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Wed Jun 28 13:13:48 2017 -0700" }, "message": "Merge branch \u0027etnaviv/fixes\u0027 of git://git.pengutronix.de/git/lst/linux\n\nPull drm/etnaviv fixes from Lucas Stach:\n \"I realized I just missed the cut-off point for the final drm fixes\n pull, but I have 2 more etnaviv fixes that need to go into 4.12, as\n they fix fallout from the explicit sync work introduced in the last\n merge window\"\n\n[ Pulling directly because Dave is on vacation. Noted by Daniel Vetter,\n and acked by Dave Airlie - Linus ]\n\n* \u0027etnaviv/fixes\u0027 of git://git.pengutronix.de/git/lst/linux:\n drm/etnaviv: Fix implicit/explicit sync sense inversion\n drm/etnaviv: fix submit flags getting overwritten by BO content\n" }, { "commit": "84a21dbdef0b96d773599c33c2afbb002198d303", "tree": "e6056e1911ed1f54c808f99937d41bcae2088488", "parents": [ "7304e8f28bb2f3cb2baf37c1bb2caa0f54ca4e43" ], "author": { "name": "Suravee Suthikulpanit", "email": "suravee.suthikulpanit@amd.com", "time": "Mon Jun 26 04:28:04 2017 -0500" }, "committer": { "name": "Joerg Roedel", "email": "jroedel@suse.de", "time": "Wed Jun 28 14:44:56 2017 +0200" }, "message": "iommu/amd: Fix interrupt remapping when disable guest_mode\n\nPass-through devices to VM guest can get updated IRQ affinity\ninformation via irq_set_affinity() when not running in guest mode.\nCurrently, AMD IOMMU driver in GA mode ignores the updated information\nif the pass-through device is setup to use vAPIC regardless of guest_mode.\nThis could cause invalid interrupt remapping.\n\nAlso, the guest_mode bit should be set and cleared only when\nSVM updates posted-interrupt interrupt remapping information.\n\nSigned-off-by: Suravee Suthikulpanit \u003csuravee.suthikulpanit@amd.com\u003e\nCc: Joerg Roedel \u003cjroedel@suse.de\u003e\nFixes: d98de49a53e48 (\u0027iommu/amd: Enable vAPIC interrupt remapping mode by default\u0027)\nSigned-off-by: Joerg Roedel \u003cjroedel@suse.de\u003e\n" }, { "commit": "fbaf94ee3cd5cfa85fe2d4d04be968820cb00556", "tree": "bacfa95214738bc69f996ce708bbb3afc789b3d6", "parents": [ "e85f82ff9b8ef503923a3be8ca6b5fd1908a7f3f" ], "author": { "name": "Miklos Szeredi", "email": "mszeredi@redhat.com", "time": "Wed Jun 28 13:41:22 2017 +0200" }, "committer": { "name": "Miklos Szeredi", "email": "mszeredi@redhat.com", "time": "Wed Jun 28 13:41:22 2017 +0200" }, "message": "ovl: don\u0027t set origin on broken lower hardlink\n\nWhen copying up a file that has multiple hard links we need to break any\nassociation with the origin file. This makes copy-up be essentially an\natomic replace.\n\nThe new file has nothing to do with the old one (except having the same\ndata and metadata initially), so don\u0027t set the overlay.origin attribute.\n\nWe can relax this in the future when we are able to index upper object by\norigin.\n\nSigned-off-by: Miklos Szeredi \u003cmszeredi@redhat.com\u003e\nFixes: 3a1e819b4e80 (\"ovl: store file handle of lower inode on copy up\")\n" }, { "commit": "e85f82ff9b8ef503923a3be8ca6b5fd1908a7f3f", "tree": "7c0b05231e998562d9fe83c41568f8f11ea480a6", "parents": [ "c0bc126f97fb929b3ae02c1c62322645d70eb408" ], "author": { "name": "Miklos Szeredi", "email": "mszeredi@redhat.com", "time": "Wed Jun 28 13:41:22 2017 +0200" }, "committer": { "name": "Miklos Szeredi", "email": "mszeredi@redhat.com", "time": "Wed Jun 28 13:41:22 2017 +0200" }, "message": "ovl: copy-up: don\u0027t unlock between lookup and link\n\nNothing prevents mischief on upper layer while we are busy copying up the\ndata.\n\nMove the lookup right before the looked up dentry is actually used.\n\nSigned-off-by: Miklos Szeredi \u003cmszeredi@redhat.com\u003e\nFixes: 01ad3eb8a073 (\"ovl: concurrent copy up of regular files\")\nCc: \u003cstable@vger.kernel.org\u003e # v4.11\n" }, { "commit": "d94815f917da770d42c377786dc428f542e38f71", "tree": "ef94286153e3104457ee51ed22a9aee1dd378605", "parents": [ "a8f20fd25bdce81a8e41767c39f456d346b63427" ], "author": { "name": "Takashi Iwai", "email": "tiwai@suse.de", "time": "Wed Jun 28 12:02:02 2017 +0200" }, "committer": { "name": "Takashi Iwai", "email": "tiwai@suse.de", "time": "Wed Jun 28 12:10:05 2017 +0200" }, "message": "ALSA: hda - Fix endless loop of codec configure\n\nazx_codec_configure() loops over the codecs found on the given\ncontroller via a linked list. The code used to work in the past, but\nin the current version, this may lead to an endless loop when a codec\nbinding returns an error.\n\nThe culprit is that the snd_hda_codec_configure() unregisters the\ndevice upon error, and this eventually deletes the given codec object\nfrom the bus. Since the list is initialized via list_del_init(), the\nnext object points to the same device itself. This behavior change\nwas introduced at splitting the HD-audio code code, and forgotten to\nadapt it here.\n\nFor fixing this bug, just use a *_safe() version of list iteration.\n\nFixes: d068ebc25e6e (\"ALSA: hda - Move some codes up to hdac_bus struct\")\nReported-by: Daniel Vetter \u003cdaniel.vetter@ffwll.ch\u003e\nCc: \u003cstable@vger.kernel.org\u003e\nSigned-off-by: Takashi Iwai \u003ctiwai@suse.de\u003e\n" }, { "commit": "426ef1bb40a3bd99ccd16d629bd1294805e96fc6", "tree": "5b78b761eae592038aca6cc0d690b6ea351edb9d", "parents": [ "f4a4381ba4944603905167d044d865990356ba22" ], "author": { "name": "Daniel Stone", "email": "daniels@collabora.com", "time": "Thu Jun 22 12:22:22 2017 +0100" }, "committer": { "name": "Lucas Stach", "email": "l.stach@pengutronix.de", "time": "Wed Jun 28 10:35:53 2017 +0200" }, "message": "drm/etnaviv: Fix implicit/explicit sync sense inversion\n\nWe were reading the no-implicit sync flag the wrong way around,\nsynchronizing too much for the explicit case, and not at all for the\nimplicit case. Oops.\n\nSigned-off-by: Daniel Stone \u003cdaniels@collabora.com\u003e\nSigned-off-by: Lucas Stach \u003cl.stach@pengutronix.de\u003e\n" }, { "commit": "f4a4381ba4944603905167d044d865990356ba22", "tree": "de20643711176489e23c3277a3722ec51fe5cd8f", "parents": [ "657314b7a5d16961e7e0ecdae4a59d28123e74c0" ], "author": { "name": "Lucas Stach", "email": "l.stach@pengutronix.de", "time": "Tue Jun 27 16:02:51 2017 +0200" }, "committer": { "name": "Lucas Stach", "email": "l.stach@pengutronix.de", "time": "Wed Jun 28 10:35:46 2017 +0200" }, "message": "drm/etnaviv: fix submit flags getting overwritten by BO content\n\nThe addition of the flags member to etnaviv_gem_submit structure didn\u0027t\ntake into account that the last member of this structure is a variable\nlength array.\n\nSigned-off-by: Lucas Stach \u003cl.stach@pengutronix.de\u003e\n" }, { "commit": "9ff1beb1d19ffe2b26bf9cd2d33e6073d4f4b5fe", "tree": "0d154cdf42dfeeb1afa11ee05395f5c16e5fa0f0", "parents": [ "5193c08c7e82248cb6b1e5d672d261c4da6bb47b", "611cdf3695a307fdca3ff3779a1e6cca935e2d31" ], "author": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Wed Jun 28 17:07:15 2017 +1000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Wed Jun 28 17:07:15 2017 +1000" }, "message": "Merge tag \u0027drm-intel-fixes-2017-06-27\u0027 of git://anongit.freedesktop.org/git/drm-intel into drm-fixes\n\nJust a few minor fixes. Important one is the execbuf async fix (aka\nANDROID_native_sync). There was another patch for a display coherency\ncorner case on APL, but we\u0027ve random-walked in that space too much,\nand the cherry-pick looked really invasive.\n\n* tag \u0027drm-intel-fixes-2017-06-27\u0027 of git://anongit.freedesktop.org/git/drm-intel:\n drm/i915: Disable EXEC_OBJECT_ASYNC when doing relocations\n drm/i915: Hold struct_mutex for per-file stats in debugfs/i915_gem_object\n drm/i915: Retire the VMA\u0027s fence tracker before unbinding\n" }, { "commit": "5193c08c7e82248cb6b1e5d672d261c4da6bb47b", "tree": "419c297a9d91661a3b27b74ec85e2fa7d8d68202", "parents": [ "c0bc126f97fb929b3ae02c1c62322645d70eb408", "82fcee526ba8ca2c5d378bdf51b21b7eb058fe3a" ], "author": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Wed Jun 28 17:06:58 2017 +1000" }, "committer": { "name": "Dave Airlie", "email": "airlied@redhat.com", "time": "Wed Jun 28 17:06:58 2017 +1000" }, "message": "Merge branch \u0027vmwgfx-fixes-4.12\u0027 of git://people.freedesktop.org/~thomash/linux into drm-fixes\n\nSingle vmwgfx fix\n* \u0027vmwgfx-fixes-4.12\u0027 of git://people.freedesktop.org/~thomash/linux:\n drm/vmwgfx: Free hash table allocated by cmdbuf managed res mgr\n" }, { "commit": "a8f20fd25bdce81a8e41767c39f456d346b63427", "tree": "2e90fa8be4832f7ddbea4c247b0a9e8916a1a48e", "parents": [ "c7ecb9068e6772c43941ce609f08bc53f36e1dce" ], "author": { "name": "Hui Wang", "email": "hui.wang@canonical.com", "time": "Wed Jun 28 08:59:16 2017 +0800" }, "committer": { "name": "Takashi Iwai", "email": "tiwai@suse.de", "time": "Wed Jun 28 07:09:19 2017 +0200" }, "message": "ALSA: hda - set input_path bitmap to zero after moving it to new place\n\nRecently we met a problem, the codec has valid adcs and input pins,\nand they can form valid input paths, but the driver does not build\nvalid controls for them like \"Mic boost\", \"Capture Volume\" and\n\"Capture Switch\".\n\nThrough debugging, I found the driver needs to shrink the invalid\nadcs and input paths for this machine, so it will move the whole\ncolumn bitmap value to the previous column, after moving it, the\ndriver forgets to set the original column bitmap value to zero, as a\nresult, the driver will invalidate the path whose index value is the\noriginal colume bitmap value. After executing this function, all\nvalid input paths are invalidated by a mistake, there are no any\nvalid input paths, so the driver won\u0027t build controls for them.\n\nFixes: 3a65bcdc577a (\"ALSA: hda - Fix inconsistent input_paths after ADC reduction\")\nCc: \u003cstable@vger.kernel.org\u003e\nSigned-off-by: Hui Wang \u003chui.wang@canonical.com\u003e\nSigned-off-by: Takashi Iwai \u003ctiwai@suse.de\u003e\n" }, { "commit": "2e31b4cb895ae78db31dffb860cd255d86c6561c", "tree": "7152005780458fa4f1927e754bc3d97a48025125", "parents": [ "d9f2950006f110f54444a10442752372ee568289" ], "author": { "name": "Trond Myklebust", "email": "trond.myklebust@primarydata.com", "time": "Tue Jun 27 17:40:50 2017 -0400" }, "committer": { "name": "Trond Myklebust", "email": "trond.myklebust@primarydata.com", "time": "Tue Jun 27 22:26:23 2017 -0400" }, "message": "NFSv4.1: nfs4_callback_free_slot() cannot call nfs4_slot_tbl_drain_complete()\n\nThe current code works only for the case where we have exactly one slot,\nwhich is no longer true.\nnfs4_free_slot() will automatically declare the callback channel to be\ndrained when all slots have been returned.\n\nSigned-off-by: Trond Myklebust \u003ctrond.myklebust@primarydata.com\u003e\n" }, { "commit": "d9f2950006f110f54444a10442752372ee568289", "tree": "c81e3a280f628cb8d4cf648ba34150c04079b2ee", "parents": [ "bd171930e6a3de4f5cffdafbb944e50093dfb59b" ], "author": { "name": "Benjamin Coddington", "email": "bcodding@redhat.com", "time": "Fri Jun 16 11:12:59 2017 -0400" }, "committer": { "name": "Trond Myklebust", "email": "trond.myklebust@primarydata.com", "time": "Tue Jun 27 21:58:14 2017 -0400" }, "message": "Revert \"NFS: nfs_rename() handle -ERESTARTSYS dentry left behind\"\n\nThis reverts commit 920b4530fb80430ff30ef83efe21ba1fa5623731 which could\ncall d_move() without holding the directory\u0027s i_mutex, and reverts commit\nd4ea7e3c5c0e341c15b073016dbf3ab6c65f12f3 \"NFS: Fix old dentry rehash after\nmove\", which was a follow-up fix.\n\nSigned-off-by: Benjamin Coddington \u003cbcodding@redhat.com\u003e\nFixes: 920b4530fb80 (\"NFS: nfs_rename() handle -ERESTARTSYS dentry left behind\")\nCc: stable@vger.kernel.org # v4.10+\nReviewed-by: Jeff Layton \u003cjlayton@redhat.com\u003e\nSigned-off-by: Trond Myklebust \u003ctrond.myklebust@primarydata.com\u003e\n" }, { "commit": "bd171930e6a3de4f5cffdafbb944e50093dfb59b", "tree": "4a6fba8ee37663af6c2405c631ec612c9ba4651a", "parents": [ "898fc11bb2bd4fbcefb685872d9fffaba2c8edaf" ], "author": { "name": "Trond Myklebust", "email": "trond.myklebust@primarydata.com", "time": "Tue Jun 27 17:33:38 2017 -0400" }, "committer": { "name": "Trond Myklebust", "email": "trond.myklebust@primarydata.com", "time": "Tue Jun 27 21:44:58 2017 -0400" }, "message": "NFSv4.1: Fix a race in nfs4_proc_layoutget\n\nIf the task calling layoutget is signalled, then it is possible for the\ncalls to nfs4_sequence_free_slot() and nfs4_layoutget_prepare() to race,\nin which case we leak a slot.\nThe fix is to move the call to nfs4_sequence_free_slot() into the\nnfs4_layoutget_release() so that it gets called at task teardown time.\n\nFixes: 2e80dbe7ac51 (\"NFSv4.1: Close callback races for OPEN, LAYOUTGET...\")\nCc: stable@vger.kernel.org # v4.8+\nSigned-off-by: Trond Myklebust \u003ctrond.myklebust@primarydata.com\u003e\n" }, { "commit": "898fc11bb2bd4fbcefb685872d9fffaba2c8edaf", "tree": "b0750f37005b8852f4efe24c86181cf3dbf661bd", "parents": [ "501e7a4689378f8b1690089bfdd4f1e12ec22903" ], "author": { "name": "Trond Myklebust", "email": "trond.myklebust@primarydata.com", "time": "Wed Jun 21 10:16:56 2017 -0400" }, "committer": { "name": "Trond Myklebust", "email": "trond.myklebust@primarydata.com", "time": "Tue Jun 27 21:44:58 2017 -0400" }, "message": "NFS: Trunking detection should handle ERESTARTSYS/EINTR\n\nCurrently, it will return EIO in those cases.\n\nSigned-off-by: Trond Myklebust \u003ctrond.myklebust@primarydata.com\u003e\n" }, { "commit": "ddbfff7429a75d954bf5bdff9f2222bceb4c236a", "tree": "89a95f0e4b2d3859fd900e9b1e74a25887bb2355", "parents": [ "25d8b92e0af75d72ce8b99e63e5a449cc0888efa" ], "author": { "name": "Aleksandar Markovic", "email": "aleksandar.markovic@imgtec.com", "time": "Mon Jun 19 17:50:12 2017 +0200" }, "committer": { "name": "Ralf Baechle", "email": "ralf@linux-mips.org", "time": "Wed Jun 28 02:54:30 2017 +0200" }, "message": "MIPS: math-emu: Handle zero accumulator case in MADDF and MSUBF separately\n\nIf accumulator value is zero, just return the value of previously\ncalculated product. This brings logic in MADDF/MSUBF implementation\ncloser to the logic in ADD/SUB case.\n\nSigned-off-by: Miodrag Dinic \u003cmiodrag.dinic@imgtec.com\u003e\nSigned-off-by: Goran Ferenc \u003cgoran.ferenc@imgtec.com\u003e\nSigned-off-by: Aleksandar Markovic \u003caleksandar.markovic@imgtec.com\u003e\nCc: James.Hogan@imgtec.com\nCc: Paul.Burton@imgtec.com\nCc: Raghu.Gandham@imgtec.com\nCc: Leonid.Yegoshin@imgtec.com\nCc: Douglas.Leung@imgtec.com\nCc: Petar.Jovanovic@imgtec.com\nCc: linux-mips@linux-mips.org\nPatchwork: https://patchwork.linux-mips.org/patch/16512/\nSigned-off-by: Ralf Baechle \u003cralf@linux-mips.org\u003e\n" }, { "commit": "ebef7368571d88f0f80b817e6898075c62265b4e", "tree": "d38abb27b86a8f58077047372cf531bdfd7ddb2b", "parents": [ "3c2bfbaadff6e0c257bb6b16c9c97f43618b13dc" ], "author": { "name": "Keith Busch", "email": "keith.busch@intel.com", "time": "Tue Jun 27 17:44:05 2017 -0600" }, "committer": { "name": "Jens Axboe", "email": "axboe@kernel.dk", "time": "Tue Jun 27 17:44:05 2017 -0600" }, "message": "nvme/pci: Fix stuck nvme reset\n\nThe controller state is set to resetting prior to disabling the\ncontroller, so this patch accounts for that state when deciding if it\nneeds to freeze the queues. Without this, an \u0027nvme reset /dev/nvme0\u0027\nblocks forever because the queues were never frozen.\n\nFixes: 82b057caefaf (\"nvme-pci: fix multiple ctrl removal scheduling\")\nSigned-off-by: Keith Busch \u003ckeith.busch@intel.com\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Jens Axboe \u003caxboe@kernel.dk\u003e\n" }, { "commit": "25d8b92e0af75d72ce8b99e63e5a449cc0888efa", "tree": "72fa5534d97c8d3b7d03570278a7529cfc370aab", "parents": [ "c0bc126f97fb929b3ae02c1c62322645d70eb408" ], "author": { "name": "Karl Beldan", "email": "karl.beldan@gmail.com", "time": "Tue Jun 27 19:22:16 2017 +0000" }, "committer": { "name": "Ralf Baechle", "email": "ralf@linux-mips.org", "time": "Tue Jun 27 23:35:21 2017 +0200" }, "message": "MIPS: head: Reorder instructions missing a delay slot\n\nIn this sequence the \u0027move\u0027 is assumed in the delay slot of the \u0027beq\u0027,\nbut head.S is in reorder mode and the former gets pushed one \u0027nop\u0027\nfarther by the assembler.\n\nThe corrected behavior made booting with an UHI supplied dtb erratic.\n\nFixes: 15f37e158892 (\"MIPS: store the appended dtb address in a variable\")\nSigned-off-by: Karl Beldan \u003ckarl.beldan+oss@gmail.com\u003e\nReviewed-by: James Hogan \u003cjames.hogan@imgtec.com\u003e\nCc: Jonas Gorski \u003cjogo@openwrt.org\u003e\nCc: linux-mips@linux-mips.org\nCc: linux-kernel@vger.kernel.org\nCc: stable@vger.kernel.org\nPatchwork: https://patchwork.linux-mips.org/patch/16614/\nSigned-off-by: Ralf Baechle \u003cralf@linux-mips.org\u003e\n" }, { "commit": "e20bd60bf62a2448be873653c7febca1d4d73afc", "tree": "a8fbb45e02c8b0c9b55978959c01ed949a53fa79", "parents": [ "85688d9adf68557233d974ef28971e69b89fb690" ], "author": { "name": "Andrew F. Davis", "email": "afd@ti.com", "time": "Mon Jun 26 12:41:20 2017 -0500" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Tue Jun 27 15:46:07 2017 -0400" }, "message": "net: usb: asix88179_178a: Add support for the Belkin B2B128\n\nThe Belkin B2B128 is a USB 3.0 Hub + Gigabit Ethernet Adapter, the\nEthernet adapter uses the ASIX AX88179 USB 3.0 to Gigabit Ethernet\nchip supported by this driver, add the USB ID for the same.\n\nThis patch is based on work by Geoffrey Tran \u003cgeoffrey.tran@gmail.com\u003e\nwho has indicated they would like this upstreamed by someone more\nfamiliar with the upstreaming process.\n\nSigned-off-by: Andrew F. Davis \u003cafd@ti.com\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "85688d9adf68557233d974ef28971e69b89fb690", "tree": "10e7848012ff3066392dabc06f26e0354a19c6a7", "parents": [ "6f64ec74515925cced6df4571638b5a099a49aae" ], "author": { "name": "Madalin Bucur", "email": "madalin.bucur@nxp.com", "time": "Mon Jun 26 18:47:00 2017 +0300" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Tue Jun 27 15:42:30 2017 -0400" }, "message": "fsl/fman: add dependency on HAS_DMA\n\nA previous commit (5567e989198b5a8d) inserted a dependency on DMA\nAPI that requires HAS_DMA to be added in Kconfig.\n\nSigned-off-by: Madalin Bucur \u003cmadalin.bucur@nxp.com\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "00a0ea33b495ee6149bf5a77ac5807ce87323abb", "tree": "b33f6834021fb03250a7fd78e9d2f2d836dac4f7", "parents": [ "c4d097d13052d1e6f29b8798264aed6135d99568" ], "author": { "name": "Vallish Vaidyeshwara", "email": "vallish@amazon.com", "time": "Fri Jun 23 18:53:06 2017 +0000" }, "committer": { "name": "Mike Snitzer", "email": "snitzer@redhat.com", "time": "Tue Jun 27 15:14:34 2017 -0400" }, "message": "dm thin: do not queue freed thin mapping for next stage processing\n\nprocess_prepared_discard_passdown_pt1() should cleanup\ndm_thin_new_mapping in cases of error.\n\ndm_pool_inc_data_range() can fail trying to get a block reference:\n\nmetadata operation \u0027dm_pool_inc_data_range\u0027 failed: error \u003d -61\n\nWhen dm_pool_inc_data_range() fails, dm thin aborts current metadata\ntransaction and marks pool as PM_READ_ONLY. Memory for thin mapping\nis released as well. However, current thin mapping will be queued\nonto next stage as part of queue_passdown_pt2() or passdown_endio().\nThis dangling thin mapping memory when processed and accessed in\nnext stage will lead to device mapper crashing.\n\nCode flow without fix:\n-\u003e process_prepared_discard_passdown_pt1(m)\n -\u003e dm_thin_remove_range()\n -\u003e discard passdown\n --\u003e passdown_endio(m) queues m onto next stage\n -\u003e dm_pool_inc_data_range() fails, frees memory m\n but does not remove it from next stage queue\n\n-\u003e process_prepared_discard_passdown_pt2(m)\n -\u003e processes freed memory m and crashes\n\nOne such stack:\n\nCall Trace:\n[\u003cffffffffa037a46f\u003e] dm_cell_release_no_holder+0x2f/0x70 [dm_bio_prison]\n[\u003cffffffffa039b6dc\u003e] cell_defer_no_holder+0x3c/0x80 [dm_thin_pool]\n[\u003cffffffffa039b88b\u003e] process_prepared_discard_passdown_pt2+0x4b/0x90 [dm_thin_pool]\n[\u003cffffffffa0399611\u003e] process_prepared+0x81/0xa0 [dm_thin_pool]\n[\u003cffffffffa039e735\u003e] do_worker+0xc5/0x820 [dm_thin_pool]\n[\u003cffffffff8152bf54\u003e] ? __schedule+0x244/0x680\n[\u003cffffffff81087e72\u003e] ? pwq_activate_delayed_work+0x42/0xb0\n[\u003cffffffff81089f53\u003e] process_one_work+0x153/0x3f0\n[\u003cffffffff8108a71b\u003e] worker_thread+0x12b/0x4b0\n[\u003cffffffff8108a5f0\u003e] ? rescuer_thread+0x350/0x350\n[\u003cffffffff8108fd6a\u003e] kthread+0xca/0xe0\n[\u003cffffffff8108fca0\u003e] ? kthread_park+0x60/0x60\n[\u003cffffffff81530b45\u003e] ret_from_fork+0x25/0x30\n\nThe fix is to first take the block ref count for discarded block and\nthen do a passdown discard of this block. If block ref count fails,\nthen bail out aborting current metadata transaction, mark pool as\nPM_READ_ONLY and also free current thin mapping memory (existing error\nhandling code) without queueing this thin mapping onto next stage of\nprocessing. If block ref count succeeds, then passdown discard of this\nblock. Discard callback of passdown_endio() will queue this thin mapping\nonto next stage of processing.\n\nCode flow with fix:\n-\u003e process_prepared_discard_passdown_pt1(m)\n -\u003e dm_thin_remove_range()\n -\u003e dm_pool_inc_data_range()\n --\u003e if fails, free memory m and bail out\n -\u003e discard passdown\n --\u003e passdown_endio(m) queues m onto next stage\n\nCc: stable \u003cstable@vger.kernel.org\u003e # v4.9+\nReviewed-by: Eduardo Valentin \u003ceduval@amazon.com\u003e\nReviewed-by: Cristian Gafton \u003cgafton@amazon.com\u003e\nReviewed-by: Anchal Agarwal \u003canchalag@amazon.com\u003e\nSigned-off-by: Vallish Vaidyeshwara \u003cvallish@amazon.com\u003e\nReviewed-by: Joe Thornber \u003cejt@redhat.com\u003e\nSigned-off-by: Mike Snitzer \u003csnitzer@redhat.com\u003e\n" }, { "commit": "6f64ec74515925cced6df4571638b5a099a49aae", "tree": "527fdc4e1a62d9a0bd43a83f1790db5b0bad5255", "parents": [ "d747a7a51b00984127a88113cdbbc26f91e9d815" ], "author": { "name": "Eric Dumazet", "email": "edumazet@google.com", "time": "Tue Jun 27 07:02:20 2017 -0700" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Tue Jun 27 14:45:12 2017 -0400" }, "message": "net: prevent sign extension in dev_get_stats()\n\nSimilar to the fix provided by Dominik Heidler in commit\n9b3dc0a17d73 (\"l2tp: cast l2tp traffic counter to unsigned\")\nwe need to take care of 32bit kernels in dev_get_stats().\n\nWhen using atomic_long_read(), we add a \u0027long\u0027 to u64 and\nmight misinterpret high order bit, unless we cast to unsigned.\n\nFixes: caf586e5f23ce (\"net: add a core netdev-\u003erx_dropped counter\")\nFixes: 015f0688f57ca (\"net: net: add a core netdev-\u003etx_dropped counter\")\nFixes: 6e7333d315a76 (\"net: add rx_nohandler stat counter\")\nSigned-off-by: Eric Dumazet \u003cedumazet@google.com\u003e\nCc: Jarod Wilson \u003cjarod@redhat.com\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "3c2bfbaadff6e0c257bb6b16c9c97f43618b13dc", "tree": "e395b50dab446e0f19468924565d8a90112d54a4", "parents": [ "da8b14e45c8b8234cb9a609c947f7e4d15f8dbef", "d360a687d99577110c181e67ebfb9a1b6fed63a2" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 27 08:56:52 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Tue Jun 27 08:56:52 2017 -0700" }, "message": "Merge branch \u0027fixes\u0027 of git://git.armlinux.org.uk/~rmk/linux-arm\n\nPull ARM fixes from Russell King:\n \"Three more fixes:\n\n - Fix the previous fix merged in the last pull for the Thumb2\n decompressor.\n\n - A fix from Vladimir to correctly identify the V7M cache type.\n\n - The optimised 3G vmsplit case does not work with LPAE, so don\u0027t\n allow this to be selected for LPAE configurations\"\n\n* \u0027fixes\u0027 of git://git.armlinux.org.uk/~rmk/linux-arm:\n ARM: 8682/1: V7M: Set cacheid iff DminLine or IminLine is nonzero\n ARM: 8681/1: make VMSPLIT_3G_OPT depends on !ARM_LPAE\n ARM: 8680/1: boot/compressed: fix inappropriate Thumb2 mnemonic for __nop\n" }, { "commit": "e3c2c4fb524b548e58108a8820c0ac7787667584", "tree": "ee172d607ccd73ca627b6df319d6aced16c428a1", "parents": [ "c0bc126f97fb929b3ae02c1c62322645d70eb408", "3f938ee2f6c4fff8d95b24636a0964b5a93cf547" ], "author": { "name": "Ingo Molnar", "email": "mingo@kernel.org", "time": "Tue Jun 27 09:17:02 2017 +0200" }, "committer": { "name": "Ingo Molnar", "email": "mingo@kernel.org", "time": "Tue Jun 27 09:17:02 2017 +0200" }, "message": "Merge tag \u0027perf-urgent-for-mingo-4.12-20170626\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/urgent\n\nPull perf/urgent fix from Arnaldo Carvalho de Melo:\n\n - Fix segfault for kernel.kptr_restrict\u003d2 (Jiri Olsa)\n\nSigned-off-by: Arnaldo Carvalho de Melo \u003cacme@redhat.com\u003e\nSigned-off-by: Ingo Molnar \u003cmingo@kernel.org\u003e\n" }, { "commit": "da8b14e45c8b8234cb9a609c947f7e4d15f8dbef", "tree": "94c6bbe28a4e5fdecc7808136ecbc8b10804f1c5", "parents": [ "9d646c97e1abf777cf229031365860e7bc1c588c", "91ebcd1b97aea87172e8ae8871f1f328b9fa9c05" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 26 12:25:59 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 26 12:25:59 2017 -0700" }, "message": "Merge tag \u0027for-linus\u0027 of git://linux-c6x.org/git/projects/linux-c6x-upstreaming\n\nPull c6x fixlet from Mark Salter:\n \"Update maintainer email\"\n\n* tag \u0027for-linus\u0027 of git://linux-c6x.org/git/projects/linux-c6x-upstreaming:\n MAINTAINERS: update email address for C6x maintainer\n" }, { "commit": "9d646c97e1abf777cf229031365860e7bc1c588c", "tree": "25a0f2700c0ffeae5b99a43b71bf84172c48a2c0", "parents": [ "c0bc126f97fb929b3ae02c1c62322645d70eb408", "4130b28f568688f79539b732797a1dc04b048442" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 26 11:58:21 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Mon Jun 26 11:58:21 2017 -0700" }, "message": "Merge branch \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux\n\nPull s390 bugfix from Martin Schwidefsky:\n \"One last s390 patch for 4.12\n\n Revert the re-IPL semantics back to the v4.7 state. It turned out that\n the memory layout may change due to memory hotplug if load-normal is\n used\"\n\n* \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux:\n s390/ipl: revert Load Normal semantics for LPAR CCW-type re-IPL\n" }, { "commit": "3f938ee2f6c4fff8d95b24636a0964b5a93cf547", "tree": "fccf1eaf673ecc1126891f0fbe27bf47fdce5bc4", "parents": [ "977282ed17a0213c84d59498a1e3f7088e211470" ], "author": { "name": "Jiri Olsa", "email": "jolsa@kernel.org", "time": "Mon Jun 26 11:51:53 2017 +0200" }, "committer": { "name": "Arnaldo Carvalho de Melo", "email": "acme@redhat.com", "time": "Mon Jun 26 11:52:37 2017 -0300" }, "message": "perf machine: Fix segfault for kernel.kptr_restrict\u003d2\n\nMichael reported the segfault when kernel.kptr_restrict\u003d2 is set.\n\n $ perf record ls\n ...\n perf: Segmentation fault\n Obtained 16 stack frames.\n ./perf(dump_stack+0x2d) [0x5068df]\n ./perf(sighandler_dump_stack+0x2d) [0x5069bf]\n ./perf() [0x43e47b]\n /lib64/libc.so.6(+0x3594f) [0x7f762004794f]\n /lib64/libc.so.6(strlen+0x26) [0x7f762009ef86]\n /lib64/libc.so.6(__strdup+0xd) [0x7f762009ecbd]\n ./perf(maps__set_kallsyms_ref_reloc_sym+0x4d) [0x51590f]\n ./perf(machine__create_kernel_maps+0x136) [0x50a7de]\n ./perf(perf_session__create_kernel_maps+0x2c) [0x510a81]\n ./perf(perf_session__new+0x13d) [0x510e23]\n ./perf() [0x43fd61]\n ./perf(cmd_record+0x704) [0x441823]\n ./perf() [0x4bc1a0]\n ./perf() [0x4bc40d]\n ./perf() [0x4bc55f]\n ./perf(main+0x2d5) [0x4bc939]\n Segmentation fault (core dumped)\n\nThe reason is that with kernel.kptr_restrict\u003d2, we don\u0027t get\nthe symbol from machine__get_running_kernel_start, which we\nwant to use in maps__set_kallsyms_ref_reloc_sym and we crash.\n\nCheck the symbol name value before calling\nmaps__set_kallsyms_ref_reloc_sym() and succeed without ref_reloc_sym\nbeing set. It\u0027s safe because we check its existence before we use it.\n\nReported-by: Michael Petlan \u003cmpetlan@redhat.com\u003e\nSigned-off-by: Jiri Olsa \u003cjolsa@kernel.org\u003e\nTested-by: Arnaldo Carvalho de Melo \u003cacme@redhat.com\u003e\nCc: David Ahern \u003cdsahern@gmail.com\u003e\nCc: Namhyung Kim \u003cnamhyung@kernel.org\u003e\nCc: Peter Zijlstra \u003ca.p.zijlstra@chello.nl\u003e\nLink: http://lkml.kernel.org/r/20170626095153.553-1-jolsa@kernel.org\nSigned-off-by: Arnaldo Carvalho de Melo \u003cacme@redhat.com\u003e\n" }, { "commit": "d6bd8194e2867e85ac2de63486d3b83ccfae4e62", "tree": "34ccc874a6d83ca2c494974c49427637d2a6c901", "parents": [ "797625deaedd9a0621376817db2813244b3246e3" ], "author": { "name": "Michael Ellerman", "email": "mpe@ellerman.id.au", "time": "Mon Jun 26 11:30:57 2017 +1000" }, "committer": { "name": "Michael Ellerman", "email": "mpe@ellerman.id.au", "time": "Mon Jun 26 23:25:08 2017 +1000" }, "message": "powerpc/32: Avoid miscompilation w/GCC 4.6.3 - don\u0027t inline copy_to/from_user()\n\nLarry Finger reported that his Powerbook G4 was no longer booting with v4.12-rc,\nuserspace was up but giving weird errors such as:\n\n udevd[64]: starting version 175\n udevd[64]: Unable to receive ctrl message: Bad address.\n modprobe: chdir(4.12-rc1): No such file or directory\n\nHe bisected the problem to commit 3448890c32c3 (\"powerpc: get rid of zeroing,\nswitch to RAW_COPY_USER\").\n\nAl identified that the problem is actually a miscompilation by GCC 4.6.3, which\nis exposed by the above commit.\n\nAl also pointed out that inlining copy_to/from_user() is probably of little or\nno benefit, which is correct. Using Anton\u0027s copy_to_user benchmark, with a\npathological single byte copy, we see a small increase in performance\nby *removing* inlining:\n\n Before (inlined):\n # time ./copy_to_user -w -l 1 -i 10000000\t( x 3 )\n real\t0m22.063s\n real\t0m22.059s\n real\t0m22.076s\n\n After:\n # time ./copy_to_user -w -l 1 -i 10000000\t( x 3 )\n real\t0m21.325s\n real\t0m21.299s\n real\t0m21.364s\n\nSo as a small performance improvement and to avoid the miscompilation, drop\ninlining copy_to/from_user() on 32-bit.\n\nFixes: 3448890c32c3 (\"powerpc: get rid of zeroing, switch to RAW_COPY_USER\")\nReported-by: Larry Finger \u003cLarry.Finger@lwfinger.net\u003e\nSuggested-by: Al Viro \u003cviro@zeniv.linux.org.uk\u003e\nSigned-off-by: Michael Ellerman \u003cmpe@ellerman.id.au\u003e\n" }, { "commit": "82fcee526ba8ca2c5d378bdf51b21b7eb058fe3a", "tree": "c50304358472d78e48c6b9eb100444de9d908cd9", "parents": [ "1929e6610bddf0cc44f0859fc72d4016cba0c1fa" ], "author": { "name": "Deepak Rawat", "email": "drawat@vmware.com", "time": "Mon Jun 26 14:39:08 2017 +0200" }, "committer": { "name": "Thomas Hellstrom", "email": "thellstrom@vmware.com", "time": "Mon Jun 26 14:39:08 2017 +0200" }, "message": "drm/vmwgfx: Free hash table allocated by cmdbuf managed res mgr\n\nThe hash table created during vmw_cmdbuf_res_man_create was\nnever freed. This causes memory leak in context creation.\nAdded the corresponding drm_ht_remove in vmw_cmdbuf_res_man_destroy.\n\nTested for memory leak by running piglit overnight and kernel\nmemory is not inflated which earlier was.\n\nCc: \u003cstable@vger.kernel.org\u003e\nSigned-off-by: Deepak Rawat \u003cdrawat@vmware.com\u003e\nReviewed-by: Sinclair Yeh \u003csyeh@vmware.com\u003e\nSigned-off-by: Thomas Hellstrom \u003cthellstrom@vmware.com\u003e\n" }, { "commit": "98fe3633c5a44e5ee3d642907739eb0210407886", "tree": "3408d8dfefe07787adc5bb32946b2a0ae22167e2", "parents": [ "c0bc126f97fb929b3ae02c1c62322645d70eb408" ], "author": { "name": "Jérôme Glisse", "email": "jglisse@redhat.com", "time": "Sat Jun 24 14:05:14 2017 -0400" }, "committer": { "name": "Ingo Molnar", "email": "mingo@kernel.org", "time": "Mon Jun 26 11:44:19 2017 +0200" }, "message": "x86/mm/hotplug: Fix BUG_ON() after hot-remove by not freeing PUD\n\nSince commit:\n\n af2cf278ef4f (\"x86/mm/hotplug: Don\u0027t remove PGD entries in remove_pagetable()\")\n\nwe no longer free PUDs so that we do not have to synchronize\nall PGDs on hot-remove/vfree().\n\nBut the new 5-level page table patchset reverted that for 4-level\npage tables, in the following commit:\n\n f2a6a7050109: (\"x86: Convert the rest of the code to support p4d_t\")\n\nThis patch restores the damage and disables free_pud() if we are in the\n4-level page table case, thus avoiding BUG_ON() after hot-remove.\n\nSigned-off-by: Jérôme Glisse \u003cjglisse@redhat.com\u003e\n[ Clarified the changelog and the code comments. ]\nReviewed-by: Kirill A. Shutemov \u003ckirill.shutemov@linux.intel.com\u003e\nCc: Andrew Morton \u003cakpm@linux-foundation.org\u003e\nCc: Andy Lutomirski \u003cluto@kernel.org\u003e\nCc: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\nCc: Logan Gunthorpe \u003clogang@deltatee.com\u003e\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nCc: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: linux-mm@kvack.org\nLink: http://lkml.kernel.org/r/20170624180514.3821-1-jglisse@redhat.com\nSigned-off-by: Ingo Molnar \u003cmingo@kernel.org\u003e\n" }, { "commit": "611cdf3695a307fdca3ff3779a1e6cca935e2d31", "tree": "b25b091e31a3c8725a78a64356c3c32dac7d8f21", "parents": [ "2c7367626733e27d6f6d9906db7a31ada587566b" ], "author": { "name": "Chris Wilson", "email": "chris@chris-wilson.co.uk", "time": "Fri Jun 16 15:05:18 2017 +0100" }, "committer": { "name": "Daniel Vetter", "email": "daniel.vetter@ffwll.ch", "time": "Mon Jun 26 10:43:26 2017 +0200" }, "message": "drm/i915: Disable EXEC_OBJECT_ASYNC when doing relocations\n\nIf we write a relocation into the buffer, we require our own implicit\nsynchronisation added after the start of the execbuf, outside of the\nuser\u0027s control. As we may end up clflushing, or doing the patch itself\non the GPU, asynchronously we need to look at the implicit serialisation\non obj-\u003eresv and hence need to disable EXEC_OBJECT_ASYNC for this\nobject.\n\nIf the user does trigger a stall for relocations, we make sure the stall\nis complete enough so that the batch is not submitted before we complete\nthose relocations.\n\nFixes: 77ae9957897d (\"drm/i915: Enable userspace to opt-out of implicit fencing\")\nSigned-off-by: Chris Wilson \u003cchris@chris-wilson.co.uk\u003e\nCc: Joonas Lahtinen \u003cjoonas.lahtinen@linux.intel.com\u003e\nCc: Jason Ekstrand \u003cjason@jlekstrand.net\u003e\nReviewed-by: Joonas Lahtinen \u003cjoonas.lahtinen@linux.intel.com\u003e\n(cherry picked from commit 071750e550af46b5d3a84ad56c2a108c3e136284)\n[danvet: Resolve conflicts, resolution reviewed by Tvrtko on irc.]\nSigned-off-by: Daniel Vetter \u003cdaniel.vetter@ffwll.ch\u003e\n" }, { "commit": "2c7367626733e27d6f6d9906db7a31ada587566b", "tree": "cc5d79d456e4e457b8f378258acc2749796cc7ce", "parents": [ "d8462d0ad350ccd399899de051060ee562a5e537" ], "author": { "name": "Chris Wilson", "email": "chris@chris-wilson.co.uk", "time": "Sat Jun 17 12:57:44 2017 +0100" }, "committer": { "name": "Daniel Vetter", "email": "daniel.vetter@ffwll.ch", "time": "Mon Jun 26 09:53:47 2017 +0200" }, "message": "drm/i915: Hold struct_mutex for per-file stats in debugfs/i915_gem_object\n\nAs we walk the obj-\u003evma_list in per_file_stats(), we need to hold\nstruct_mutex to prevent alteration of that list.\n\nFixes: 1d2ac403ae3b (\"drm: Protect dev-\u003efilelist with its own mutex\")\nFixes: c84455b4bacc (\"drm/i915: Move debug only per-request pid tracking from request to ctx\")\nBugzilla: https://bugs.freedesktop.org/show_bug.cgi?id\u003d101460\nSigned-off-by: Chris Wilson \u003cchris@chris-wilson.co.uk\u003e\nCc: Daniel Vetter \u003cdaniel.vetter@intel.com\u003e\nCc: Joonas Lahtinen \u003cjoonas.lahtinen@linux.intel.com\u003e\nLink: http://patchwork.freedesktop.org/patch/msgid/20170617115744.4452-1-chris@chris-wilson.co.uk\nReviewed-by: Tvrtko Ursulin \u003ctvrtko.ursulin@intel.com\u003e\n(cherry picked from commit 0caf81b5c53d9bd332a95dbcb44db8de0b397a7c)\nSigned-off-by: Daniel Vetter \u003cdaniel.vetter@ffwll.ch\u003e\n" }, { "commit": "d8462d0ad350ccd399899de051060ee562a5e537", "tree": "7df755e7ff2fe6d89336fa85293621c8f5a7a225", "parents": [ "c0bc126f97fb929b3ae02c1c62322645d70eb408" ], "author": { "name": "Chris Wilson", "email": "chris@chris-wilson.co.uk", "time": "Tue Jun 20 13:43:19 2017 +0100" }, "committer": { "name": "Daniel Vetter", "email": "daniel.vetter@ffwll.ch", "time": "Mon Jun 26 09:53:39 2017 +0200" }, "message": "drm/i915: Retire the VMA\u0027s fence tracker before unbinding\n\nSince we may track unfenced access (GPU access to the vma that\nexplicitly requires no fence), vma-\u003elast_fence may be set without any\nattached fence (vma-\u003efence) and so will not be flushed when we call\ni915_vma_put_fence(). Since we stopped doing a full retire of the\nactivity trackers for unbind, we need to explicitly retire each tracker.\n\nFixes: b0decaf75bd9 (\"drm/i915: Track active vma requests\")\nSigned-off-by: Chris Wilson \u003cchris@chris-wilson.co.uk\u003e\nCc: Joonas Lahtinen \u003cjoonas.lahtinen@linux.intel.com\u003e\nCc: Tvrtko Ursulin \u003ctvrtko.ursulin@linux.intel.com\u003e\nLink: http://patchwork.freedesktop.org/patch/msgid/20170620124321.1108-1-chris@chris-wilson.co.uk\nReviewed-by: Tvrtko Ursulin \u003ctvrtko.ursulin@linux.intel.com\u003e\n(cherry picked from commit 760a898d8069111704e1bd43f00ebf369ae46e57)\nSigned-off-by: Daniel Vetter \u003cdaniel.vetter@ffwll.ch\u003e\n" }, { "commit": "c0bc126f97fb929b3ae02c1c62322645d70eb408", "tree": "311ebe1faa69d3dfd667cef5c4e9b44c80f51805", "parents": [ "a4fd8b3accf43d407472e34403d4b0a4df5c0e71" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jun 25 18:30:05 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jun 25 18:30:05 2017 -0700" }, "message": "Linux 4.12-rc7\n" }, { "commit": "a4fd8b3accf43d407472e34403d4b0a4df5c0e71", "tree": "6562da0e4b50f623faf0551617063441a7a70d9a", "parents": [ "5f4b37d8784da4217ede8f8bcd301686853dcf8c", "26fcd952d5c977a94ac64bb44ed409e37607b2c9" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jun 25 12:01:56 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jun 25 12:01:56 2017 -0700" }, "message": "Merge branch \u0027x86-urgent-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull x86 fix from Thomas Gleixner:\n \"A single fix to unbreak the vdso32 build for 64bit kernels caused by\n excess #includes in the mshyperv header\"\n\n* \u0027x86-urgent-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n x86/mshyperv: Remove excess #includes from mshyperv.h\n" }, { "commit": "5f4b37d8784da4217ede8f8bcd301686853dcf8c", "tree": "0ce7799c0133a4707628b128473f22db89ff7502", "parents": [ "35d8d5d47c0ef066197f629f71cdd6039693b86d", "8e6cec1c7c5afa489687c90be15d6ed82c742975" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jun 25 11:59:19 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jun 25 11:59:19 2017 -0700" }, "message": "Merge branch \u0027timers-urgent-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull timer fixes from Thomas Gleixner:\n \"A few fixes for timekeeping and timers:\n\n - Plug a subtle race due to a missing READ_ONCE() in the timekeeping\n code where reloading of a pointer results in an inconsistent\n callback argument being supplied to the clocksource-\u003eread function.\n\n - Correct the CLOCK_MONOTONIC_RAW sub-nanosecond accounting in the\n time keeping core code, to prevent a possible discontuity.\n\n - Apply a similar fix to the arm64 vdso clock_gettime()\n implementation\n\n - Add missing includes to clocksource drivers, which relied on\n indirect includes which fails in certain configs.\n\n - Use the proper iomem pointer for read/iounmap in a probe function\"\n\n* \u0027timers-urgent-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n arm64/vdso: Fix nsec handling for CLOCK_MONOTONIC_RAW\n time: Fix CLOCK_MONOTONIC_RAW sub-nanosecond accounting\n time: Fix clock-\u003eread(clock) race around clocksource changes\n clocksource: Explicitly include linux/clocksource.h when needed\n clocksource/drivers/arm_arch_timer: Fix read and iounmap of incorrect variable\n" }, { "commit": "35d8d5d47c0ef066197f629f71cdd6039693b86d", "tree": "0156f16eed2f2f91b8b9f65143503ec0efdb1926", "parents": [ "1a8cca1880e6127a22cf6604b3c4ccd8b27cbf65", "977282ed17a0213c84d59498a1e3f7088e211470" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jun 25 11:55:21 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jun 25 11:55:21 2017 -0700" }, "message": "Merge branch \u0027perf-urgent-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull perf fixes from Thomas Gleixner:\n \"Three fixlets for perf:\n\n - Return the proper error code if aux buffers for a event are not\n supported.\n\n - Calculate the probe offset for inlined functions correctly\n\n - Update the Skylake DTLB load/store miss event so it can count 1G\n TLB entries as well\"\n\n* \u0027perf-urgent-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n perf probe: Fix probe definition for inlined functions\n perf/x86/intel: Add 1G DTLB load/store miss support for SKL\n perf/aux: Correct return code of rb_alloc_aux() if !has_aux(ev)\n" }, { "commit": "1a8cca1880e6127a22cf6604b3c4ccd8b27cbf65", "tree": "7d9384f0b9092a6fca4a715e7a8dc6eb9ab3a6ee", "parents": [ "412572bffa9d1bc0f3ebae39291a815dad1d0418", "9f93d87cba63e3d18629261243b1f633519eabb5" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jun 25 11:53:36 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jun 25 11:53:36 2017 -0700" }, "message": "Merge branch \u0027irq-urgent-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip\n\nPull irq fix from Thomas Gleixner:\n \"A single fix for the MIPS GIC to prevent ftrace recursion\"\n\n* \u0027irq-urgent-for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:\n irqchip/mips-gic: Mark count and compare accessors notrace\n" }, { "commit": "412572bffa9d1bc0f3ebae39291a815dad1d0418", "tree": "583c4af60b143e682a00596b0700ff160f13ab90", "parents": [ "d5d5c1825e85a144e67448ad777eff441ddb907f", "9768935264c4f0e4afd788a185d8e8d89c28e41d" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jun 25 10:39:43 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jun 25 10:39:43 2017 -0700" }, "message": "Merge branch \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input\n\nPull input fixes from Dmitry Torokhov:\n\n - a quirk to i8042 to ignore timeout bit on Lifebook AH544\n\n - a fixup to Synaptics RMI function 54 that was breaking some Dells\n\n - a fix for memory leak in soc_button_array driver\n\n* \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:\n Input: synaptics-rmi4 - only read the F54 query registers which are used\n Input: i8042 - add Fujitsu Lifebook AH544 to notimeout list\n Input: soc_button_array - fix leaking the ACPI button descriptor buffer\n" }, { "commit": "d5d5c1825e85a144e67448ad777eff441ddb907f", "tree": "56d65b81ace7e253e5709fd90363e90824a407d2", "parents": [ "bb9b8fd26b50c9ec822805b9d877bec9d88acff4", "abb85a9b512e8ca7ad04a5a8a6db9664fe644974" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jun 25 10:36:44 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sun Jun 25 10:36:44 2017 -0700" }, "message": "Merge git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending\n\nPull SCSI target fixes from Nicholas Bellinger:\n \"Here are the target-pending fixes for v4.12-rc7 that have been queued\n up for the last 2 weeks. This includes:\n\n - Fix a TMR related kref underflow detected by the recent refcount_t\n conversion in upstream.\n\n - Fix a iscsi-target corner case during explicit connection logout\n timeout failure.\n\n - Address last fallout in iscsi-target immediate data handling from\n v4.4 target-core now allowing control CDB payload underflow\"\n\n* git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending:\n iscsi-target: Reject immediate data underflow larger than SCSI transfer length\n iscsi-target: Fix delayed logout processing greater than SECONDS_FOR_LOGOUT_COMP\n target: Fix kref-\u003erefcount underflow in transport_cmd_finish_abort\n" }, { "commit": "d747a7a51b00984127a88113cdbbc26f91e9d815", "tree": "aef4eb636d2b95553a7907ae986a3b96e195bdd9", "parents": [ "85cb73ff9b74785a7fc752875d7f0fe17ca3ea7c" ], "author": { "name": "WANG Cong", "email": "xiyou.wangcong@gmail.com", "time": "Sat Jun 24 23:50:30 2017 -0700" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Sun Jun 25 12:23:07 2017 -0400" }, "message": "tcp: reset sk_rx_dst in tcp_disconnect()\n\nWe have to reset the sk-\u003esk_rx_dst when we disconnect a TCP\nconnection, because otherwise when we re-connect it this\ndst reference is simply overridden in tcp_finish_connect().\n\nThis fixes a dst leak which leads to a loopback dev refcnt\nleak. It is a long-standing bug, Kevin reported a very similar\n(if not same) bug before. Thanks to Andrei for providing such\na reliable reproducer which greatly narrows down the problem.\n\nFixes: 41063e9dd119 (\"ipv4: Early TCP socket demux.\")\nReported-by: Andrei Vagin \u003cavagin@gmail.com\u003e\nReported-by: Kevin Xu \u003ckaiwen.xu@hulu.com\u003e\nSigned-off-by: Cong Wang \u003cxiyou.wangcong@gmail.com\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "85cb73ff9b74785a7fc752875d7f0fe17ca3ea7c", "tree": "00339cc580169297c2ae02f372d7f90cc0f7e652", "parents": [ "d0c32a16235aeacd32c9de6ff90f9219614d7e4e" ], "author": { "name": "Wei Wang", "email": "weiwan@google.com", "time": "Fri Jun 23 15:25:37 2017 -0700" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Sun Jun 25 11:46:56 2017 -0400" }, "message": "net: ipv6: reset daddr and dport in sk if connect() fails\n\nIn __ip6_datagram_connect(), reset sk-\u003esk_v6_daddr and inet-\u003edport if\nerror occurs.\nIn udp_v6_early_demux(), check for sk_state to make sure it is in\nTCP_ESTABLISHED state.\nTogether, it makes sure unconnected UDP socket won\u0027t be considered as a\nvalid candidate for early demux.\n\nv3: add TCP_ESTABLISHED state check in udp_v6_early_demux()\nv2: fix compilation error\n\nFixes: 5425077d73e0 (\"net: ipv6: Add early demux handler for UDP unicast\")\nSigned-off-by: Wei Wang \u003cweiwan@google.com\u003e\nAcked-by: Maciej Żenczykowski \u003cmaze@google.com\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "d0c32a16235aeacd32c9de6ff90f9219614d7e4e", "tree": "80add3385fe6533da4b74db9d350319480c52336", "parents": [ "92cc8a5105e1236584e92288595ddbd2149bb6cf" ], "author": { "name": "Mintz, Yuval", "email": "Yuval.Mintz@cavium.com", "time": "Sat Jun 24 15:37:00 2017 +0300" }, "committer": { "name": "David S. Miller", "email": "davem@davemloft.net", "time": "Sun Jun 25 11:36:56 2017 -0400" }, "message": "bnx2x: Don\u0027t log mc removal needlessly\n\nWhen mc configuration changes bnx2x_config_mcast() can return 0 for\nsuccess, negative for failure and positive for benign reason preventing\nits immediate work, e.g., when the command awaits the completion of\na previously sent command.\n\nWhen removing all configured macs on a 578xx adapter, if a positive\nvalue would be returned driver would errneously log it as an error.\n\nFixes: c7b7b483ccc9 (\"bnx2x: Don\u0027t flush multicast MACs\")\nSigned-off-by: Yuval Mintz \u003cYuval.Mintz@cavium.com\u003e\nSigned-off-by: David S. Miller \u003cdavem@davemloft.net\u003e\n" }, { "commit": "bb9b8fd26b50c9ec822805b9d877bec9d88acff4", "tree": "64e562f6a22a13fc3b63f2f9289ca688c50ad52b", "parents": [ "f65013d655ac9639f37d3b54189f6468f672e60b", "ad8181060788c80c0ad75b583f24c22fa962a7a6" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Jun 24 16:18:00 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Jun 24 16:18:00 2017 -0700" }, "message": "Merge tag \u0027kbuild-fixes-v4.12-2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild\n\nPull Kbuild fixes from Masahiro Yamada:\n \"Nothing scary, just some random fixes:\n\n - fix warnings of host programs\n\n - fix \"make tags\" when COMPILED_SOURCE\u003d1 is specified along with O\u003d\n\n - clarify help message of C\u003d1 option\n\n - fix dependency for ncurses compatibility check\n\n - fix \"make headers_install\" for fakechroot environment\"\n\n* tag \u0027kbuild-fixes-v4.12-2\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:\n kconfig: fix sparse warnings in nconfig\n kbuild: fix header installation under fakechroot environment\n kconfig: Check for libncurses before menuconfig\n Kbuild: tiny correction on `make help`\n tags: honor COMPILED_SOURCE with apart output directory\n genksyms: add printf format attribute to error_with_pos()\n" }, { "commit": "f65013d655ac9639f37d3b54189f6468f672e60b", "tree": "395f8f44cada4613934e309491421b0701e03613", "parents": [ "94a6df251dd08c6436ebd6d10c68f03659148ce1", "57db7e4a2d92c2d3dfbca4ef8057849b2682436b" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Jun 24 02:24:53 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Sat Jun 24 02:24:53 2017 -0700" }, "message": "Merge branch \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/user-namespace\n\nPull timer fix from Eric Biederman:\n \"This fixes an issue of confusing injected signals with the signals\n from posix timers that has existed since posix timers have been in the\n kernel.\n\n This patch is slightly simpler than my earlier version of this patch\n as I discovered in testing that I had misspelled \"#ifdef\n CONFIG_POSIX_TIMERS\". So I deleted that unnecessary test and made\n setting of resched_timer uncondtional.\n\n I have tested this and verified that without this patch there is a\n nasty hang that is easy to trigger, and with this patch everything\n works properly\"\n\nThomas Gleixner dixit:\n \"It fixes the problem at hand and covers the ptrace case as well, which\n I missed.\n\n Reviewed-and-tested-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\"\n\n* \u0027for-linus\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/user-namespace:\n signal: Only reschedule timers on signals timers have sent\n" }, { "commit": "26fcd952d5c977a94ac64bb44ed409e37607b2c9", "tree": "ce32288b58ab87c450ecef24a7bdd45df5d9e931", "parents": [ "8d829b9bb878904f3419af8003176a12910b94e8" ], "author": { "name": "Thomas Gleixner", "email": "tglx@linutronix.de", "time": "Fri Jun 23 10:50:38 2017 +0200" }, "committer": { "name": "Ingo Molnar", "email": "mingo@kernel.org", "time": "Sat Jun 24 08:48:51 2017 +0200" }, "message": "x86/mshyperv: Remove excess #includes from mshyperv.h\n\nA recent commit included linux/slab.h in linux/irq.h. This breaks the build\nof vdso32 on a 64-bit kernel.\n\nThe reason is that linux/irq.h gets included into the vdso code via\nlinux/interrupt.h which is included from asm/mshyperv.h. That makes the\n32-bit vdso compile fail, because slab.h includes the pgtable headers for\n64-bit on a 64-bit build.\n\nNeither linux/clocksource.h nor linux/interrupt.h are needed in the\nmshyperv.h header file itself - it has a dependency on \u003clinux/atomic.h\u003e.\n\nRemove the includes and unbreak the build.\n\nReported-by: Ingo Molnar \u003cmingo@kernel.org\u003e\nSigned-off-by: Thomas Gleixner \u003ctglx@linutronix.de\u003e\nCc: K. Y. Srinivasan \u003ckys@microsoft.com\u003e\nCc: Linus Torvalds \u003ctorvalds@linux-foundation.org\u003e\nCc: Peter Zijlstra \u003cpeterz@infradead.org\u003e\nCc: Vitaly Kuznetsov \u003cvkuznets@redhat.com\u003e\nCc: devel@linuxdriverproject.org\nFixes: dee863b571b0 (\"hv: export current Hyper-V clocksource\")\nLink: http://lkml.kernel.org/r/alpine.DEB.2.20.1706231038460.2647@nanos\nSigned-off-by: Ingo Molnar \u003cmingo@kernel.org\u003e\n" }, { "commit": "94a6df251dd08c6436ebd6d10c68f03659148ce1", "tree": "51aca5d25d4bffa9860d8c1b1d96d3a64cd28a38", "parents": [ "cd5545ae87ed9ca76dba7753b436713ce8df2872", "34f19ff1b5a0d11e46df479623d6936460105c9f" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 23 17:53:16 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 23 17:53:16 2017 -0700" }, "message": "Merge tag \u0027powerpc-4.12-7\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux\n\nPull powerpc fixes from Michael Ellerman:\n \"Some more powerpc fixes for 4.12. Most of these actually came in last\n week but got held up for some more testing.\n\n - three fixes for kprobes/ftrace/livepatch interactions.\n\n - properly handle data breakpoints when using the Radix MMU.\n\n - fix for perf sampling of registers during call_usermodehelper().\n\n - properly initialise the thread_info on our emergency stacks\n\n - add an explicit flush when doing TLB invalidations for a process\n using NPU2.\n\n Thanks to: Alistair Popple, Naveen N. Rao, Nicholas Piggin, Ravi\n Bangoria, Masami Hiramatsu\"\n\n* tag \u0027powerpc-4.12-7\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:\n powerpc/64: Initialise thread_info for emergency stacks\n powerpc/powernv/npu-dma: Add explicit flush when sending an ATSD\n powerpc/perf: Fix oops when kthread execs user process\n powerpc/64s: Handle data breakpoints in Radix mode\n powerpc/kprobes: Skip livepatch_handler() for jprobes\n powerpc/ftrace: Pass the correct stack pointer for DYNAMIC_FTRACE_WITH_REGS\n powerpc/kprobes: Pause function_graph tracing during jprobes handling\n" }, { "commit": "cd5545ae87ed9ca76dba7753b436713ce8df2872", "tree": "3b3dc7f0e421dac137f02c5be012d20744c0423e", "parents": [ "ba6cbdb673c287cee87a965266419fd2111e20c4", "e4330d8bf669139a983255d1801733b64c2ae841" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 23 17:49:12 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 23 17:49:12 2017 -0700" }, "message": "Merge tag \u0027acpi-4.12-rc7\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm\n\nPull ACPI fix from Rafael Wysocki:\n \"This fixes the ACPI-based enumeration of some I2C and SPI devices\n broken in 4.11.\n\n Specifics:\n\n - I2C and SPI devices are expected to be enumerated by the I2C and\n SPI subsystems, respectively, but due to a change made during the\n 4.11 cycle, in some cases the ACPI core marks them as already\n enumerated which causes the I2C and SPI subsystems to overlook\n them, so fix that (Jarkko Nikula)\"\n\n* tag \u0027acpi-4.12-rc7\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:\n ACPI / scan: Fix enumeration for special SPI and I2C devices\n" }, { "commit": "ba6cbdb673c287cee87a965266419fd2111e20c4", "tree": "aef4f30552353259b27421cc2d179ca762f3a89a", "parents": [ "25b2398f5c0167accd6d0ab18db0561ac22ca6cc", "6c782a5ea56a799658e213a78dc1455264938afa" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 23 17:46:10 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 23 17:46:10 2017 -0700" }, "message": "Merge branch \u0027i2c/for-current-fixed\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux\n\nPull i2c fix from Wolfram Sang.\n\n* \u0027i2c/for-current-fixed\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:\n i2c: imx: Use correct function to write to register\n" }, { "commit": "25b2398f5c0167accd6d0ab18db0561ac22ca6cc", "tree": "fa00cc450f3f8e6b4d6c9d7c4ce338d6b744bfe0", "parents": [ "51c933f208c51b3e5371779ba3c246bde144ed11", "6c7515c61ffa0985c57abd8892c7928b52b9a306" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 23 17:40:41 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 23 17:40:41 2017 -0700" }, "message": "Merge tag \u0027gpio-v4.12-3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio\n\nPull GPIO fix from Linus Walleij:\n \"A single GPIO patch fixing the compatible string for the MVEBU PWM\n controller embedded in the GPIO controller before we release v4.12.\n Hopefully\"\n\n* tag \u0027gpio-v4.12-3\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio:\n gpio: mvebu: change compatible string for PWM support\n" }, { "commit": "51c933f208c51b3e5371779ba3c246bde144ed11", "tree": "8c756afbd1dbd1b3a5b890f89a97384cb2e97ae3", "parents": [ "311548f173ff53bd4c7407eaf52097d00d2ea4de", "c7ecb9068e6772c43941ce609f08bc53f36e1dce" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 23 17:37:56 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 23 17:37:56 2017 -0700" }, "message": "Merge tag \u0027sound-4.12-rc7\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound\n\nPull sound fixes from Takashi Iwai:\n \"Nothing exciting here, just a few stable fixes:\n\n - suppress spurious kernel WARNING in PCM core\n\n - fix potential spin deadlock at error handling in firewire\n\n - HD-audio PCI ID addition / fixup\"\n\n* tag \u0027sound-4.12-rc7\u0027 of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:\n ALSA: hda - Apply quirks to Broxton-T, too\n ALSA: firewire-lib: Fix stall of process context at packet error\n ALSA: pcm: Don\u0027t treat NULL chmap as a fatal error\n ALSA: hda - Add Coffelake PCI ID\n" }, { "commit": "311548f173ff53bd4c7407eaf52097d00d2ea4de", "tree": "306415dd24627509dcd2978313edd0f9c367bdca", "parents": [ "7139a06b1637fa09d826997d6b869c59d40ac5c1", "33ce7563a4858430fdf38bb867513d3fbc3dda65" ], "author": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 23 17:35:57 2017 -0700" }, "committer": { "name": "Linus Torvalds", "email": "torvalds@linux-foundation.org", "time": "Fri Jun 23 17:35:57 2017 -0700" }, "message": "Merge tag \u0027drm-fixes-for-v4.12-rc7\u0027 of git://people.freedesktop.org/~airlied/linux\n\nPull drm fixes from Dave Airlie:\n \"A varied bunch of fixes, one for an API regression with connectors.\n\n Otherwise amdgpu and i915 have a bunch of varied fixes, the shrinker\n ones being the most important\"\n\n* tag \u0027drm-fixes-for-v4.12-rc7\u0027 of git://people.freedesktop.org/~airlied/linux:\n drm: Fix GETCONNECTOR regression\n drm/radeon: add a quirk for Toshiba Satellite L20-183\n drm/radeon: add a PX quirk for another K53TK variant\n drm/amdgpu: adjust default display clock\n drm/amdgpu/atom: fix ps allocation size for EnableDispPowerGating\n drm/amdgpu: add Polaris12 DID\n drm/i915: Don\u0027t enable backlight at setup time.\n drm/i915: Plumb the correct acquire ctx into intel_crtc_disable_noatomic()\n drm/i915: Fix deadlock witha the pipe A quirk during resume\n drm/i915: Remove __GFP_NORETRY from our buffer allocator\n drm/i915: Encourage our shrinker more when our shmemfs allocations fails\n drm/i915: Differentiate between sw write location into ring and last hw read\n" } ], "next": "7139a06b1637fa09d826997d6b869c59d40ac5c1" }