Re: Linux 5.11-rc2
On Fri, Jan 8, 2021 at 3:12 AM Christopher William Snowhill wrote: > > There appears to be a regression with the filesystem NLS modules. I cannot > load any of them. They all produce: > > modprobe: ERROR: could not insert 'nls_cp437': Invalid argument > > The system journal reports: > > Jan 08 02:04:56 mrgency kernel: BPF:No data > Jan 08 02:04:56 mrgency kernel: failed to validate module [nls_cp437] BTF: -22 > Seems like you are getting module BTF with absolutely no types and strings, and BTF verification fails for such case (currently). Can you please share your .config, and also let me know what version of pahole do you have (pahole --version)? Thanks! The fix is to allow such "empty" BTF, most probably, but I'd like to double-check nothing fishy is going on before that. > As a result, I cannot mount vfat filesystems with the usual CP437 default. Or > any other code page, for that matter.
Re: Linux 5.11-rc2
There appears to be a regression with the filesystem NLS modules. I cannot load any of them. They all produce: modprobe: ERROR: could not insert 'nls_cp437': Invalid argument The system journal reports: Jan 08 02:04:56 mrgency kernel: BPF:No data Jan 08 02:04:56 mrgency kernel: failed to validate module [nls_cp437] BTF: -22 As a result, I cannot mount vfat filesystems with the usual CP437 default. Or any other code page, for that matter.
Re: Linux 5.11-rc2
On Tue, Jan 5, 2021 at 12:57 PM Guenter Roeck wrote: > > NP. The test are running automatically anyway, so I figured I might > as well report the results. Does that make sense, or is it just noise ? Definitely not noise. I very much like seeing the results. In fact, even in situations like this, where the results are a subset of what the previous rc already had, I very much like seeing them, if only to make sure the issues don't get forgotten about. So please continue for as long as you have the cycles available.. Linus
Re: Linux 5.11-rc2
On Tue, Jan 05, 2021 at 11:55:20AM -0800, Linus Torvalds wrote: > On Tue, Jan 5, 2021 at 10:46 AM Guenter Roeck wrote: > > > > Problems are as already reported against v5.11-rc1. > > Yes. Thanks for keeping on top of this, I'm expecting to get the fixes > as people get back from vacations. > NP. The test are running automatically anyway, so I figured I might as well report the results. Does that make sense, or is it just noise ? Thanks, Guenter
Re: Linux 5.11-rc2
On Tue, Jan 5, 2021 at 10:46 AM Guenter Roeck wrote: > > Problems are as already reported against v5.11-rc1. Yes. Thanks for keeping on top of this, I'm expecting to get the fixes as people get back from vacations. Linus
Re: Linux 5.11-rc2
On Sun, Jan 03, 2021 at 04:13:08PM -0800, Linus Torvalds wrote: > Ok, let's be honest - not a lot has happened in the last week or two. > > The merge window itself may not have been hugely impacted by the > holiday season, but that's because all the new code should already > have been ready before the merge window even opened, so the holidays > just didn't end up affecting things all that much. > > But people have (rightly) mostly been offline since, presumably > over-eating and doing all the other traditional holiday things. And > just generally not being hugely active. That very much shows in a tiny > rc2 release. > > I expect next week to slowly start ramping up fixes, but I know some > people are still on vacation or just in an extended food coma, and > there's a delay from testing to fixes, so we'll see. Maybe rc3 ends up > being fairly small too. > > It's much too early to say whether this will then end up causing some > delays in the final release - it's possible, but with 5.11 not being a > particularly big release maybe it doesn't even matter that we had a > fairly quiet week or two in the early rc series. > > Anyway, for whatever reasons, the few fixes we _do_ have in rc2 tend > to be mostly in SCSI and block devices. But there's a random > smattering of other things too. For once, the shortlog is so small > that you might as well just read it. > > Time to slowly crawl out from under all the xmas wrapping paper piles > and go test... > Build results: total: 153 pass: 151 fail: 2 Failed builds: arm64:allmodconfig ia64:defconfig Qemu test results: total: 430 pass: 428 fail: 2 Failed tests: arm:raspi2:multi_v7_defconfig:bcm2836-rpi-2-b:initrd arm:raspi2:multi_v7_defconfig:sd:bcm2836-rpi-2-b:rootfs arm64:allmodconfig: ERROR: modpost: "irq_check_status_bit" [drivers/perf/arm_spe_pmu.ko] undefined! ia64:defconfig: include/linux/mmzone.h:1156:2: error: #error Allocator MAX_ORDER exceeds SECTION_SIZE arm:raspi2 boot tests stall during boot. Problems are as already reported against v5.11-rc1. Guenter
Linux 5.11-rc2
Ok, let's be honest - not a lot has happened in the last week or two. The merge window itself may not have been hugely impacted by the holiday season, but that's because all the new code should already have been ready before the merge window even opened, so the holidays just didn't end up affecting things all that much. But people have (rightly) mostly been offline since, presumably over-eating and doing all the other traditional holiday things. And just generally not being hugely active. That very much shows in a tiny rc2 release. I expect next week to slowly start ramping up fixes, but I know some people are still on vacation or just in an extended food coma, and there's a delay from testing to fixes, so we'll see. Maybe rc3 ends up being fairly small too. It's much too early to say whether this will then end up causing some delays in the final release - it's possible, but with 5.11 not being a particularly big release maybe it doesn't even matter that we had a fairly quiet week or two in the early rc series. Anyway, for whatever reasons, the few fixes we _do_ have in rc2 tend to be mostly in SCSI and block devices. But there's a random smattering of other things too. For once, the shortlog is so small that you might as well just read it. Time to slowly crawl out from under all the xmas wrapping paper piles and go test... Linus --- Adrian Hunter (4): scsi: ufs-pci: Fix restore from S4 for Intel controllers scsi: ufs-pci: Ensure UFS device is in PowerDown mode for suspend-to-disk ->poweroff() scsi: ufs-pci: Fix recovery from hibernate exit errors for Intel controllers scsi: ufs-pci: Enable UFSHCD_CAP_RPM_AUTOSUSPEND for Intel controllers Alan Stern (1): scsi: block: Do not accept any requests while suspended Andres Freund (1): block: add debugfs stanza for QUEUE_FLAG_NOWAIT Artem Bityutskiy (1): intel_idle: add SnowRidge C-state table Baoquan He (1): mm: memmap defer init doesn't work as expected Bart Van Assche (7): scsi: block: Fix a race in the runtime power management code scsi: block: Introduce BLK_MQ_REQ_PM scsi: ide: Do not set the RQF_PREEMPT flag for sense requests scsi: ide: Mark power management requests with RQF_PM instead of RQF_PREEMPT scsi: scsi_transport_spi: Set RQF_PM for domain validation commands scsi: core: Only process PM requests if rpm_status != RPM_ACTIVE scsi: block: Remove RQF_PREEMPT and BLK_MQ_REQ_PREEMPT Bean Huo (2): scsi: ufs: Remove unused macro definition POWER_DESC_MAX_SIZE scsi: ufs: Fix wrong print message in dev_err() Bhaskar Chowdhury (1): kernel: cgroup: Mundane spelling fixes throughout the file Dan Carpenter (1): scsi: mpt3sas: Signedness bug in _base_get_diag_triggers() Harish (1): selftests/vm: fix building protection keys test Heiko Carstens (2): s390: update defconfigs s390/Kconfig: sort config S390 select list once again Huang Shijie (2): sizes.h: add SZ_8G/SZ_16G/SZ_32G macros lib/genalloc: fix the overflow when size is too big Hui Su (1): cgroup/cgroup.c: replace 'of->kn->priv' with of_cft() Ilya Dryomov (4): ceph: reencode gid_list when reconnecting libceph: fix auth_signature buffer allocation in secure mode libceph: align session_key and con_secret to 16 bytes libceph: add __maybe_unused to DEFINE_MSGR2_FEATURE Ilya Leoshkevich (1): lib/zlib: fix inflating zlib streams on s390 Jens Axboe (1): io_uring: don't assume mm is constant across submits Joe Perches (1): checkpatch: prefer strscpy to strlcpy Josh Poimboeuf (1): kdev_t: always inline major/minor helper functions Kalesh Singh (1): mm/mremap.c: fix extent calculation Linus Torvalds (2): depmod: handle the case of /sbin/depmod without /sbin in PATH Linux 5.11-rc2 Lukas Bulwahn (1): MAINTAINERS: include governors into CPU IDLE TIME MANAGEMENT FRAMEWORK Masahiro Yamada (1): Revert "kbuild: avoid static_assert for genksyms" Mike Kravetz (1): mm/hugetlb: fix deadlock in hugetlb_cow error path Mike Snitzer (1): Revert "dm crypt: export sysfs of kcryptd workqueue" Nicholas Piggin (1): mm: generalise COW SMC TLB flushing race comment Pavel Begunkov (3): io_uring: add a helper for setting a ref node io_uring: fix io_sqe_files_unregister() hangs kernel/io_uring: cancel io_uring before task works Qinglang Miao (1): cgroup: Fix memory leak when parsing multiple source parameters Quanyang Wang (1): opp: fix memory leak in _allocate_opp_table Rafael J. Wysocki (1): cpufreq: intel_pstate: Fix fast-switch fallback path Randall Huang (1): scsi: ufs: Clear UAC for RPMB after ufshcd resets Randy Dunlap (4): scsi: cxgb4i: Fix TLS dependency local64.h: make mandatory zlib: move EXPORT_SYMBOL() and MODULE_LICENSE() out of dfltcc_syms