- Kernel - mailweb.openeuler.org

[PATCH 1/2] mmap: export __do_mmap_mm symbol
by Wupeng Ma 19 Sep '23

19 Sep '23

From: Chunsheng Luo <luochunsheng(a)huawei.com> euleros inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I7WLVX --------------------------------------------- export __do_mmap_mm symbol by referring to the earlier version Signed-off-by: Chunsheng Luo <luochunsheng(a)huawei.com> --- mm/mmap.c | 1 + 1 file changed, 1 insertion(+) diff --git a/mm/mmap.c b/mm/mmap.c index 2aef07b8a85e..819c37e7eb1f 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -1469,6 +1469,7 @@ unsigned long __do_mmap_mm(struct mm_struct *mm, struct file *file, unsigned lon *populate = len; return addr; } +EXPORT_SYMBOL(__do_mmap_mm); unsigned long do_mmap(struct file *file, unsigned long addr, unsigned long len, unsigned long prot, unsigned long flags, -- 2.33.0

1 1

[PATCH OLK-5.10] sched/qos: Fix warning in CPU hotplug scenarios
by Xia Fukun 19 Sep '23

19 Sep '23

hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I7ZMCB CVE: NA -------------------------------- CPU hotplug callbacks race against distribute_cfs_runtime(), when the QOS_SCHED feature is enabled, there may be situations where the cfs_rq-> runtime_remaining == 1 and cfs_rq is QOS_THROTTLED. Turn off the Qos_throttle when the CPU is offline. No longer allocate time to cfs_rq in this scenario to fix the warning. Fixes: 4eb6eb7941dc ("sched/qos: Don't unthrottle cfs_rq when cfs_rq is throttled by qos") Signed-off-by: Xia Fukun <xiafukun(a)huawei.com> --- kernel/sched/fair.c | 17 +++++++++++++++++ 1 file changed, 17 insertions(+) diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index daa853b19853..b8bf7acb9f9a 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -5267,6 +5267,19 @@ static void distribute_cfs_runtime(struct cfs_bandwidth *cfs_b) if (!cfs_rq_throttled(cfs_rq)) goto next; + /* + * CPU hotplug callbacks race against distribute_cfs_runtime() + * when the QOS_SCHED feature is enabled, there may be + * situations where the runtime_remaining > 0. + * Qos_sched does not care whether the cfs_rq has time left, + * so no longer allocate time to cfs_rq in this scenario. + */ +#ifdef CONFIG_QOS_SCHED + if (cfs_rq->throttled == QOS_THROTTLED && + cfs_rq->runtime_remaining > 0) + goto next; +#endif + /* By the above check, this should never be true */ SCHED_WARN_ON(cfs_rq->runtime_remaining > 0); @@ -7923,6 +7936,10 @@ static __always_inline bool check_qos_cfs_rq(struct cfs_rq *cfs_rq) if (unlikely(cfs_rq && is_offline_level(cfs_rq->tg->qos_level) && !sched_idle_cpu(smp_processor_id()) && cfs_rq->h_nr_running == cfs_rq->idle_h_nr_running)) { + + if (!rq_of(cfs_rq)->online) + return false; + throttle_qos_cfs_rq(cfs_rq); return true; } -- 2.34.1

2 1

[PATCH openEuler-23.09] mm/mlock: return EINVAL for illegal user memory range in mlock
by Wupeng Ma 19 Sep '23

19 Sep '23

From: Ma Wupeng <mawupeng1(a)huawei.com> hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I7TW89 -------------------------------- While testing mlock, we have a problem if the len of mlock is ULONG_MAX. The return value of mlock is zero. But nothing will be locked since the len in do_mlock overflows to zero due to the following code in mlock: len = PAGE_ALIGN(len + (offset_in_page(start))); The same problem happens in munlock. Since TASK_SIZE is the maximum user space address. The start or len of mlock shouldn't be bigger than this. Function access_ok can be used to check this issue, so return -EINVAL if bigger. Signed-off-by: Ma Wupeng <mawupeng1(a)huawei.com> --- mm/mlock.c | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) diff --git a/mm/mlock.c b/mm/mlock.c index 40b43f8740df..e90139d42f88 100644 --- a/mm/mlock.c +++ b/mm/mlock.c @@ -479,8 +479,6 @@ static int apply_vma_lock_flags(unsigned long start, size_t len, end = start + len; if (end < start) return -EINVAL; - if (end == start) - return 0; vma = vma_iter_load(&vmi); if (!vma) return -ENOMEM; @@ -574,9 +572,15 @@ static __must_check int do_mlock(unsigned long start, size_t len, vm_flags_t fla if (!can_do_mlock()) return -EPERM; + if (!len) + return 0; + len = PAGE_ALIGN(len + (offset_in_page(start))); start &= PAGE_MASK; + if (!len) + return -EINVAL; + lock_limit = rlimit(RLIMIT_MEMLOCK); lock_limit >>= PAGE_SHIFT; locked = len >> PAGE_SHIFT; @@ -634,8 +638,13 @@ SYSCALL_DEFINE2(munlock, unsigned long, start, size_t, len) start = untagged_addr(start); + if (!len) + return 0; + len = PAGE_ALIGN(len + (offset_in_page(start))); start &= PAGE_MASK; + if (!len) + return -EINVAL; if (mmap_write_lock_killable(current->mm)) return -EINTR; -- 2.25.1

2 1

[PATCH OLK-5.10] uacce: modify the configuration mode of device isolation stragety
by Wenkai Lin 19 Sep '23

19 Sep '23

From: Qi Tao <taoqi10(a)huawei.com> driver inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I82ARL CVE: NA ------------------------------------------------------------------ cancel the reference counting of the accelerator device so that the device (with hardware errors) isolation stragety can also be modified while the accelerator is running a task. Signed-off-by: Qi Tao <taoqi10(a)huawei.com> Signed-off-by: JiangShui Yang <yangjiangshui(a)h-partners.com> --- drivers/misc/uacce/uacce.c | 5 ----- include/linux/uacce.h | 2 -- 2 files changed, 7 deletions(-) diff --git a/drivers/misc/uacce/uacce.c b/drivers/misc/uacce/uacce.c index 21a4f5892aff..7a64eed01df7 100644 --- a/drivers/misc/uacce/uacce.c +++ b/drivers/misc/uacce/uacce.c @@ -88,7 +88,6 @@ static int uacce_put_queue(struct uacce_queue *q) uacce->ops->put_queue(q); q->state = UACCE_Q_ZOMBIE; - atomic_dec(&uacce->ref); return 0; } @@ -348,7 +347,6 @@ static int uacce_fops_open(struct inode *inode, struct file *filep) goto out_with_bond; } - atomic_inc(&uacce->ref); init_waitqueue_head(&q->wait); filep->private_data = q; q->state = UACCE_Q_INIT; @@ -821,9 +819,6 @@ static ssize_t isolate_strategy_store(struct device *dev, struct device_attribut if (val > UACCE_MAX_ERR_THRESHOLD) return -EINVAL; - if (atomic_read(&uacce->ref)) - return -EBUSY; - ret = uacce->ops->isolate_err_threshold_write(uacce, val); if (ret) return ret; diff --git a/include/linux/uacce.h b/include/linux/uacce.h index 8187c1bda236..9b0c04a9cff7 100644 --- a/include/linux/uacce.h +++ b/include/linux/uacce.h @@ -148,7 +148,6 @@ struct uacce_queue { * @mutex: protects uacce operation * @priv: private pointer of the uacce * @queues: list of queues - * @ref: reference of the uacce */ struct uacce_device { const char *algs; @@ -164,7 +163,6 @@ struct uacce_device { struct device dev; struct mutex mutex; void *priv; - atomic_t ref; struct uacce_err_isolate *isolate; struct list_head queues; }; -- 2.30.0

2 1

[PATCH OLK-5.10 v2] jbd2: Fix potential data lost in recovering journal raced with synchronizing fs bdev
by Zhihao Cheng 19 Sep '23

19 Sep '23

maillist inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I828EV CVE: NA Reference: https://lore.kernel.org/linux-ext4/20230911155822.kbg2xayc7ys24kay@quack3/T… -------------------------------- JBD2 makes sure journal data is fallen on fs device by sync_blockdev(), however, other process could intercept the EIO information from bdev's mapping, which leads journal recovering successful even EIO occurs during data written back to fs device. We found this problem in our product, iscsi + multipath is chosen for block device of ext4. Unstable network may trigger kpartx to rescan partitions in device mapper layer. Detailed process is shown as following: mount kpartx irq jbd2_journal_recover do_one_pass memcpy(nbh->b_data, obh->b_data) // copy data to fs dev from journal mark_buffer_dirty // mark bh dirty vfs_read generic_file_read_iter // dio filemap_write_and_wait_range __filemap_fdatawrite_range do_writepages block_write_full_folio submit_bh_wbc >> EIO occurs in disk << end_buffer_async_write mark_buffer_write_io_error mapping_set_error set_bit(AS_EIO, &mapping->flags) // set! filemap_check_errors test_and_clear_bit(AS_EIO, &mapping->flags) // clear! err2 = sync_blockdev filemap_write_and_wait filemap_check_errors test_and_clear_bit(AS_EIO, &mapping->flags) // false err2 = 0 Filesystem is mounted successfully even data from journal is failed written into disk, and ext4/ocfs2 could become corrupted. Fix it by comparing the wb_err state in fs block device before recovering and after recovering. Fetch a reproducer in [Link]. Link: https://bugzilla.kernel.org/show_bug.cgi?id=217888 Cc: stable(a)vger.kernel.org Signed-off-by: Zhihao Cheng <chengzhihao1(a)huawei.com> Signed-off-by: Zhang Yi <yi.zhang(a)huawei.com> --- v1->v3: Initialize wb_err. Untialized wb_err could be same with mapping->wb_err(eg. EIO without ERRSEQ_SEEN). When EIO occurs again, mapping->wb_err won't be changed, and wb_err is still same with mapping->wb_err. fs/jbd2/recovery.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/fs/jbd2/recovery.c b/fs/jbd2/recovery.c index 9e4537349380..cdf4c553f058 100644 --- a/fs/jbd2/recovery.c +++ b/fs/jbd2/recovery.c @@ -292,6 +292,8 @@ int jbd2_journal_recover(journal_t *journal) journal_superblock_t * sb; struct recovery_info info; + errseq_t wb_err; + struct address_space *mapping; memset(&info, 0, sizeof(info)); sb = journal->j_superblock; @@ -309,6 +311,9 @@ int jbd2_journal_recover(journal_t *journal) return 0; } + wb_err = 0; + mapping = journal->j_fs_dev->bd_inode->i_mapping; + errseq_check_and_advance(&mapping->wb_err, &wb_err); err = do_one_pass(journal, &info, PASS_SCAN); if (!err) err = do_one_pass(journal, &info, PASS_REVOKE); @@ -327,6 +332,9 @@ int jbd2_journal_recover(journal_t *journal) jbd2_journal_clear_revoke(journal); err2 = sync_blockdev(journal->j_fs_dev); + if (!err) + err = err2; + err2 = errseq_check_and_advance(&mapping->wb_err, &wb_err); if (!err) err = err2; /* Make sure all replayed data is on permanent storage */ -- 2.31.1

2 1

[PATCH openEuler-23.09] config: Disable x86 IBT for kpatch
by Wei Li 19 Sep '23

19 Sep '23

hulk inclusion category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/I82GQG -------------------------------- The CONFIG_X86_KERNEL_IBT option causes the compiled symbols to lack the __prefix__ symbol, leading kpatch to incorrectly determine that the symbol has no padding. In reality, the symbol does have padding, resulting in a compilation failure for kpatch. Let's disable IBT for now. Signed-off-by: Wei Li <liwei391(a)huawei.com> --- arch/x86/configs/openeuler_defconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/x86/configs/openeuler_defconfig b/arch/x86/configs/openeuler_defconfig index 6bbaaaa82a0b..d3c0da3ddd64 100644 --- a/arch/x86/configs/openeuler_defconfig +++ b/arch/x86/configs/openeuler_defconfig @@ -457,7 +457,7 @@ CONFIG_MTRR_SANITIZER_SPARE_REG_NR_DEFAULT=1 CONFIG_X86_PAT=y CONFIG_ARCH_USES_PG_UNCACHED=y CONFIG_X86_UMIP=y -CONFIG_X86_KERNEL_IBT=y +# CONFIG_X86_KERNEL_IBT is not set CONFIG_X86_INTEL_MEMORY_PROTECTION_KEYS=y CONFIG_X86_INTEL_TSX_MODE_OFF=y # CONFIG_X86_INTEL_TSX_MODE_ON is not set -- 2.25.1

2 1

[PATCH openEuler-1.0-LTS] crypto:hisilicon/qm - cache write back before flr and poweroff
by w00416078 19 Sep '23

19 Sep '23

From: Yu'an Wang <wangyuan46(a)huawei.com> driver inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I4YKIJ CVE: NA -------------------------------- In order to prevent an error when the hardware writes back cache, the driver should write back the hardware cache before flr and poweroff. If vfs is enabled, we should process abnormal scenes of the whole vfs. Signed-off-by: Yu'an Wang <wangyuan46(a)huawei.com> Reviewed-by: Weili Qian <qianweili(a)huawei.com> Reviewed-by: Ling Mingqiang <lingmingqiang(a)huawei.com> Acked-by: Xie XiuQi <xiexiuqi(a)huawei.com> --- drivers/crypto/hisilicon/qm.c | 31 ++++++++++++++++++++++--------- 1 file changed, 22 insertions(+), 9 deletions(-) diff --git a/drivers/crypto/hisilicon/qm.c b/drivers/crypto/hisilicon/qm.c index 8fa855fd2387..e0fd83465dce 100644 --- a/drivers/crypto/hisilicon/qm.c +++ b/drivers/crypto/hisilicon/qm.c @@ -2772,6 +2772,8 @@ void hisi_qm_dev_shutdown(struct pci_dev *pdev) ret = hisi_qm_stop(qm, QM_NORMAL); if (ret) dev_err(&pdev->dev, "Fail to stop qm in shutdown!\n"); + + hisi_qm_cache_wb(qm); } EXPORT_SYMBOL_GPL(hisi_qm_dev_shutdown); @@ -3718,12 +3720,16 @@ static int qm_vf_reset_prepare(struct pci_dev *pdev, pci_save_state(dev); ret = hisi_qm_stop(qm, stop_reason); - if (ret) - goto prepare_fail; + if (ret) { + hisi_qm_set_hw_reset(qm, + QM_RESET_STOP_TX_OFFSET); + hisi_qm_set_hw_reset(qm, + QM_RESET_STOP_RX_OFFSET); + atomic_set(&qm->status.flags, QM_STOP); + } } } -prepare_fail: mutex_unlock(&qm_list->lock); return ret; } @@ -4117,19 +4123,26 @@ void hisi_qm_reset_prepare(struct pci_dev *pdev) if (qm->vfs_num) { ret = qm_vf_reset_prepare(pdev, qm->qm_list, QM_FLR); - if (ret) { - pci_err(pdev, "Fails to prepare reset!\n"); - return; - } + if (ret) + pci_err(pdev, "Failed to stop vfs!\n"); } ret = hisi_qm_stop(qm, QM_FLR); if (ret) { - pci_err(pdev, "Fails to stop QM!\n"); - return; + pci_err(pdev, "Failed to stop QM!\n"); + goto err_prepare; } + hisi_qm_cache_wb(qm); pci_info(pdev, "FLR resetting...\n"); + return; + +err_prepare: + pci_info(pdev, "FLR resetting prepare failed!\n"); + hisi_qm_set_hw_reset(qm, QM_RESET_STOP_TX_OFFSET); + hisi_qm_set_hw_reset(qm, QM_RESET_STOP_RX_OFFSET); + atomic_set(&qm->status.flags, QM_STOP); + hisi_qm_cache_wb(qm); } EXPORT_SYMBOL_GPL(hisi_qm_reset_prepare); -- 2.34.1

2 1

[PATCH openEuler-1.0-LTS v2 00/11] Fix booting failure on arm64
by Wei Li 18 Sep '23

18 Sep '23

The arm64 test machine boot failed when using hulk_defconfig after d63c76835476 ("arm64: efi: Execute runtime services from a dedicated stack"), revert this patch set first for weekly release. Wei Li (11): Revert "arm64: efi: Make efi_rt_lock a raw_spinlock" Revert "efi: rt-wrapper: Add missing include" Revert "arm64: efi: Recover from synchronous exceptions occurring in firmware" Revert "arm64: efi: Execute runtime services from a dedicated stack" Revert "efi: fix userspace infinite retry read efivars after EFI runtime services page fault" Revert "arm64: efi: Restore register x18 if it was corrupted" Revert "x86/efi: fix a -Wtype-limits compilation warning" Revert "efi: Fix build error due to enum collision between efi.h and ima.h" Revert "efi: Fix debugobjects warning on 'efi_rts_work'" Revert "efi/x86: Handle page faults occurring while running EFI runtime services" Revert "efi: Make efi_rts_work accessible to efi page fault handler" arch/arm64/include/asm/efi.h | 12 --- arch/arm64/kernel/efi-rt-wrapper.S | 49 +---------- arch/arm64/kernel/efi.c | 50 ----------- arch/arm64/mm/fault.c | 4 - arch/x86/include/asm/efi.h | 1 - arch/x86/mm/fault.c | 9 -- arch/x86/platform/efi/quirks.c | 78 ----------------- drivers/firmware/efi/runtime-wrappers.c | 109 +++++++++++++++--------- include/linux/efi.h | 42 --------- 9 files changed, 70 insertions(+), 284 deletions(-) -- 2.25.1

2 12

[PATCH openEuler-1.0-LTS] crypto:hisilicon/sec - modify hw endian config
by w00416078 18 Sep '23

18 Sep '23

From: Yu'an Wang <wangyuan46(a)huawei.com> driver inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I4YKLC CVE: NA -------------------------------- When the endian configuration of the hardware is abnormal, it will affect the normal function. Currently the soft configuration method can't restore the faulty device. The endian needs to be configured according to the system properties. So fix it. Signed-off-by: Yu'an Wang <wangyuan46(a)huawei.com> Reviewed-by: Kai Ye <yekai13(a)huawei.com> Reviewed-by: Ling Mingqiang <lingmingqiang(a)huawei.com> Acked-by: Xie XiuQi <xiexiuqi(a)huawei.com> --- drivers/crypto/hisilicon/sec2/sec_main.c | 36 +++++++----------------- 1 file changed, 10 insertions(+), 26 deletions(-) diff --git a/drivers/crypto/hisilicon/sec2/sec_main.c b/drivers/crypto/hisilicon/sec2/sec_main.c index 58f726ba022f..a568d5363c1e 100644 --- a/drivers/crypto/hisilicon/sec2/sec_main.c +++ b/drivers/crypto/hisilicon/sec2/sec_main.c @@ -255,33 +255,19 @@ static const struct pci_device_id sec_dev_ids[] = { }; MODULE_DEVICE_TABLE(pci, sec_dev_ids); -static u8 sec_get_endian(struct hisi_qm *qm) +static void sec_set_endian(struct hisi_qm *qm) { u32 reg; - /* - * As for VF, it is a wrong way to get endian setting by - * reading a register of the engine - */ - if (qm->pdev->is_virtfn) { - dev_err_ratelimited(&qm->pdev->dev, - "cannot access a register in VF!\n"); - return SEC_LE; - } - reg = readl_relaxed(qm->io_base + SEC_ENGINE_PF_CFG_OFF + - SEC_ACC_COMMON_REG_OFF + SEC_CONTROL_REG); - - /* BD little endian mode */ - if (!(reg & BIT(0))) - return SEC_LE; + reg = readl_relaxed(SEC_ADDR(qm, SEC_CONTROL_REG)); + reg &= ~(BIT(1) | BIT(0)); + if (IS_ENABLED(CONFIG_64BIT)) + reg |= BIT(1); - /* BD 32-bits big endian mode */ - else if (!(reg & BIT(1))) - return SEC_32BE; + if (IS_ENABLED(CONFIG_CPU_BIG_ENDIAN)) + reg |= BIT(0); - /* BD 64-bits big endian mode */ - else - return SEC_64BE; + writel_relaxed(reg, SEC_ADDR(qm, SEC_CONTROL_REG)); } static int sec_engine_init(struct hisi_qm *qm) @@ -331,9 +317,7 @@ static int sec_engine_init(struct hisi_qm *qm) SEC_ADDR(qm, SEC_BD_ERR_CHK_EN_REG3)); /* config endian */ - reg = readl_relaxed(SEC_ADDR(qm, SEC_CONTROL_REG)); - reg |= sec_get_endian(qm); - writel_relaxed(reg, SEC_ADDR(qm, SEC_CONTROL_REG)); + sec_set_endian(qm); return 0; } @@ -813,7 +797,7 @@ static int sec_qm_pre_init(struct hisi_qm *qm, struct pci_dev *pdev) { int ret; - qm->algs = "sec\ncipher\ndigest\naead\n"; + qm->algs = "cipher\ndigest\naead\n"; qm->uacce_mode = uacce_mode; qm->pdev = pdev; ret = hisi_qm_pre_init(qm, pf_q_num, SEC_PF_DEF_Q_BASE); -- 2.34.1

2 1

[PATCH OLK-5.10 0/2] Not clear ATA_PFLAG_EH_PENDING and not thaw the port twice in ata_eh_reset()
by Xingui Yang 18 Sep '23

18 Sep '23

Clear port pending interrupts before reset, as per AHCI specifications (Szuying). Followup fixes for this one are to not clear ATA_PFLAG_EH_PENDING in ata_eh_reset() to allow EH to continue on with other actions recorded with error interrupts triggered before EH completes. A~Nd an additional fix to avoid thawing a port twice in EH (Niklas). Niklas Cassel (2): ata: libata-eh: do not clear ATA_PFLAG_EH_PENDING in ata_eh_reset() ata: libata-eh: do not thaw the port twice in ata_eh_reset() drivers/ata/libata-eh.c | 16 +++------------- 1 file changed, 3 insertions(+), 13 deletions(-) -- 2.17.1

2 3