- Kernel - mailweb.openeuler.org

[PATCH OLK-5.10 00/18] xfs: rework feature flags
by Long Li 06 Jun '23

06 Jun '23

1 18

[PATCH OLK-5.10 1/3] xfs: fix the problem of mount failure caused by not refreshing mp->m_sb
by Zhihao Cheng 06 Jun '23

06 Jun '23

From: Wu Guanghao <wuguanghao3(a)huawei.com> hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I77NBB -------------------------------- After testing xfs_growfs + fsstress + fault injection, the following stack appeared when mounting the filesystem: [ 149.902032] XFS (loop0): xfs_buf_map_verify: daddr 0x200001 out of range, EOFS 0x200000 [ 149.902072] WARNING: CPU: 12 PID: 3045 at fs/xfs/xfs_buf.c:535 xfs_buf_get_map+0x5ae/0x650 [xfs] ... [ 149.902473] xfs_buf_read_map+0x59/0x330 [xfs] [ 149.902621] ? xlog_recover_items_pass2+0x55/0xd0 [xfs] [ 149.902809] xlog_recover_buf_commit_pass2+0xff/0x640 [xfs] [ 149.902959] ? xlog_recover_items_pass2+0x55/0xd0 [xfs] [ 149.903104] xlog_recover_items_pass2+0x55/0xd0 [xfs] [ 149.903247] xlog_recover_commit_trans+0x2e0/0x330 [xfs] [ 149.903390] xlog_recovery_process_trans+0x8e/0xf0 [xfs] [ 149.903531] xlog_recover_process_data+0x9c/0x130 [xfs] [ 149.903687] xlog_do_recovery_pass+0x3cc/0x5d0 [xfs] [ 149.903843] xlog_do_log_recovery+0x5c/0x80 [xfs] [ 149.903984] xlog_do_recover+0x33/0x1c0 [xfs] [ 149.904125] xlog_recover+0xdd/0x190 [xfs] [ 149.904265] xfs_log_mount+0x125/0x2f0 [xfs] [ 149.904410] xfs_mountfs+0x41a/0x910 [xfs] [ 149.904558] ? __pfx_xfs_fstrm_free_func+0x10/0x10 [xfs] [ 149.904725] xfs_fs_fill_super+0x4b7/0x940 [xfs] [ 149.904873] ? __pfx_xfs_fs_fill_super+0x10/0x10 [xfs] [ 149.905016] get_tree_bdev+0x19a/0x280 [ 149.905020] vfs_get_tree+0x29/0xd0 [ 149.905023] path_mount+0x69e/0x9b0 [ 149.905026] do_mount+0x7d/0xa0 [ 149.905029] __x64_sys_mount+0xdc/0x100 [ 149.905032] do_syscall_64+0x3e/0x90 [ 149.905035] entry_SYSCALL_64_after_hwframe+0x72/0xdc The trigger process is as follows: 1. Growfs size from 0x200000 to 0x300000 2. Using the space range of 0x200000~0x300000 3. The above operations have only been written to the log area on disk 4. Fault injection and shutdown filesystem 5. Mount the filesystem and replay the log about growfs, but only modify the superblock buffer without modifying the mp->m_sb structure in memory 6. Continuing the log replay, at this point we are replaying operation 2, then it was discovered that the blocks used more than mp->m_sb.sb_dblocks Therefore, during log replay, if there are any modifications made to the superblock, we should refresh the information recorded in the mp->m_sb. Signed-off-by: Wu Guanghao <wuguanghao3(a)huawei.com> Signed-off-by: Zhihao Cheng <chengzhihao1(a)huawei.com> --- fs/xfs/xfs_buf_item_recover.c | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+) diff --git a/fs/xfs/xfs_buf_item_recover.c b/fs/xfs/xfs_buf_item_recover.c index d892125eb022..d78060a262bd 100644 --- a/fs/xfs/xfs_buf_item_recover.c +++ b/fs/xfs/xfs_buf_item_recover.c @@ -22,6 +22,8 @@ #include "xfs_inode.h" #include "xfs_dir2.h" #include "xfs_quota.h" +#include "xfs_sb.h" +#include "xfs_ag.h" /* * This is the number of entries in the l_buf_cancel_table used during @@ -969,6 +971,28 @@ xlog_recover_buf_commit_pass2( goto out_release; } else { xlog_recover_do_reg_buffer(mp, item, bp, buf_f, current_lsn); + /* + * If the superblock buffer is modified, we also need to modify the + * content of the mp. + */ + if (bp->b_maps[0].bm_bn == XFS_SB_DADDR && bp->b_ops) { + struct xfs_dsb *sb = bp->b_addr; + + bp->b_ops->verify_write(bp); + error = bp->b_error; + if (error) + goto out_release; + + if (be32_to_cpu(sb->sb_agcount) > mp->m_sb.sb_agcount) { + error = xfs_initialize_perag(mp, + be32_to_cpu(sb->sb_agcount), + &mp->m_maxagi); + if (error) + goto out_release; + } + + xfs_sb_from_disk(&mp->m_sb, sb); + } } /* -- 2.31.1

1 0

[PATCH OLK-5.10 0/2] jbd2: fix checkpoint inconsistent
by Zhihao Cheng 06 Jun '23

06 Jun '23

Zhang Yi (2): jbd2: recheck chechpointing non-dirty buffer jbd2: remove t_checkpoint_io_list fs/jbd2/checkpoint.c | 144 ++++++++++--------------------------------- fs/jbd2/commit.c | 3 +- 2 files changed, 32 insertions(+), 115 deletions(-) -- 2.31.1

2 4

[PATCH OLK-5.10] dm thin: Fix ABBA deadlock by resetting dm_bufio_client
by Li Lingfeng 05 Jun '23

05 Jun '23

hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I79ZEK CVE: NA -------------------------------- As described in commit d0dcee7d1cf8 ("dm thin: Fix ABBA deadlock between shrink_slab and dm_pool_abort_metadata"), ABBA deadlock will be triggered since shrinker_rwsem need to be held when operations failed on dm pool metadata. We have noticed the following three problem scenarios: 1) Described by commit d0dcee7d1cf8 ("dm thin: Fix ABBA deadlock between shrink_slab and dm_pool_abort_metadata") 2) shrinker_rwsem and throttle->lock P1(drop cache) P2(kworker) drop_caches_sysctl_handler drop_slab shrink_slab down_read(&shrinker_rwsem) - LOCK A do_shrink_slab super_cache_scan prune_icache_sb dispose_list evict ext4_evict_inode ext4_clear_inode ext4_discard_preallocations ext4_mb_load_buddy_gfp ext4_mb_init_cache ext4_wait_block_bitmap __ext4_error ext4_handle_error ext4_commit_super ... dm_submit_bio do_worker throttle_work_update down_write(&t->lock) -- LOCK B process_deferred_bios commit metadata_operation_failed dm_pool_abort_metadata dm_block_manager_create dm_bufio_client_create register_shrinker down_write(&shrinker_rwsem) -- LOCK A thin_map thin_bio_map thin_defer_bio_with_throttle throttle_lock down_read(&t->lock) - LOCK B 3) shrinker_rwsem and wait_on_buffer P1(drop cache) P2(kworker) drop_caches_sysctl_handler drop_slab shrink_slab down_read(&shrinker_rwsem) - LOCK A do_shrink_slab ... ext4_wait_block_bitmap __ext4_error ext4_handle_error jbd2_journal_abort jbd2_journal_update_sb_errno jbd2_write_superblock submit_bh // LOCK B // RELEASE B do_worker throttle_work_update down_write(&t->lock) - LOCK B process_deferred_bios process_bio commit metadata_operation_failed dm_pool_abort_metadata dm_block_manager_create dm_bufio_client_create register_shrinker register_shrinker_prepared down_write(&shrinker_rwsem) - LOCK A bio_endio wait_on_buffer __wait_on_buffer Fix these by resetting dm_bufio_client without holding shrinker_rwsem. Signed-off-by: Li Lingfeng <lilingfeng3(a)huawei.com> --- drivers/md/dm-bufio.c | 7 +++ drivers/md/dm-thin-metadata.c | 59 ++++++++----------- drivers/md/persistent-data/dm-block-manager.c | 6 ++ drivers/md/persistent-data/dm-block-manager.h | 1 + drivers/md/persistent-data/dm-space-map.h | 3 +- .../persistent-data/dm-transaction-manager.c | 3 + include/linux/dm-bufio.h | 2 + 7 files changed, 47 insertions(+), 34 deletions(-) diff --git a/drivers/md/dm-bufio.c b/drivers/md/dm-bufio.c index 50f3e673729c..7e3db7618916 100644 --- a/drivers/md/dm-bufio.c +++ b/drivers/md/dm-bufio.c @@ -1883,6 +1883,13 @@ void dm_bufio_client_destroy(struct dm_bufio_client *c) } EXPORT_SYMBOL_GPL(dm_bufio_client_destroy); +void dm_bufio_client_reset(struct dm_bufio_client *c) +{ + drop_buffers(c); + flush_work(&c->shrink_work); +} +EXPORT_SYMBOL_GPL(dm_bufio_client_reset); + void dm_bufio_set_sector_offset(struct dm_bufio_client *c, sector_t start) { c->start = start; diff --git a/drivers/md/dm-thin-metadata.c b/drivers/md/dm-thin-metadata.c index 8f4d149bb99d..85e80fc17641 100644 --- a/drivers/md/dm-thin-metadata.c +++ b/drivers/md/dm-thin-metadata.c @@ -574,6 +574,8 @@ static int __format_metadata(struct dm_pool_metadata *pmd) r = dm_tm_create_with_sm(pmd->bm, THIN_SUPERBLOCK_LOCATION, &pmd->tm, &pmd->metadata_sm); if (r < 0) { + pmd->tm = NULL; + pmd->metadata_sm = NULL; DMERR("tm_create_with_sm failed"); return r; } @@ -582,6 +584,7 @@ static int __format_metadata(struct dm_pool_metadata *pmd) if (IS_ERR(pmd->data_sm)) { DMERR("sm_disk_create failed"); r = PTR_ERR(pmd->data_sm); + pmd->data_sm = NULL; goto bad_cleanup_tm; } @@ -612,11 +615,15 @@ static int __format_metadata(struct dm_pool_metadata *pmd) bad_cleanup_nb_tm: dm_tm_destroy(pmd->nb_tm); + pmd->nb_tm = NULL; bad_cleanup_data_sm: dm_sm_destroy(pmd->data_sm); + pmd->data_sm = NULL; bad_cleanup_tm: dm_tm_destroy(pmd->tm); + pmd->tm = NULL; dm_sm_destroy(pmd->metadata_sm); + pmd->metadata_sm = NULL; return r; } @@ -682,6 +689,8 @@ static int __open_metadata(struct dm_pool_metadata *pmd) sizeof(disk_super->metadata_space_map_root), &pmd->tm, &pmd->metadata_sm); if (r < 0) { + pmd->tm = NULL; + pmd->metadata_sm = NULL; DMERR("tm_open_with_sm failed"); goto bad_unlock_sblock; } @@ -691,6 +700,7 @@ static int __open_metadata(struct dm_pool_metadata *pmd) if (IS_ERR(pmd->data_sm)) { DMERR("sm_disk_open failed"); r = PTR_ERR(pmd->data_sm); + pmd->data_sm = NULL; goto bad_cleanup_tm; } @@ -717,9 +727,12 @@ static int __open_metadata(struct dm_pool_metadata *pmd) bad_cleanup_data_sm: dm_sm_destroy(pmd->data_sm); + pmd->data_sm = NULL; bad_cleanup_tm: dm_tm_destroy(pmd->tm); + pmd->tm = NULL; dm_sm_destroy(pmd->metadata_sm); + pmd->metadata_sm = NULL; bad_unlock_sblock: dm_bm_unlock(sblock); @@ -766,9 +779,13 @@ static void __destroy_persistent_data_objects(struct dm_pool_metadata *pmd, bool destroy_bm) { dm_sm_destroy(pmd->data_sm); + pmd->data_sm = NULL; dm_sm_destroy(pmd->metadata_sm); + pmd->metadata_sm = NULL; dm_tm_destroy(pmd->nb_tm); + pmd->nb_tm = NULL; dm_tm_destroy(pmd->tm); + pmd->tm = NULL; if (destroy_bm) dm_block_manager_destroy(pmd->bm); } @@ -976,8 +993,7 @@ int dm_pool_metadata_close(struct dm_pool_metadata *pmd) __func__, r); } pmd_write_unlock(pmd); - if (!pmd->fail_io) - __destroy_persistent_data_objects(pmd, true); + __destroy_persistent_data_objects(pmd, true); kfree(pmd); return 0; @@ -1884,53 +1900,30 @@ static void __set_abort_with_changes_flags(struct dm_pool_metadata *pmd) int dm_pool_abort_metadata(struct dm_pool_metadata *pmd) { int r = -EINVAL; - struct dm_block_manager *old_bm = NULL, *new_bm = NULL; /* fail_io is double-checked with pmd->root_lock held below */ if (unlikely(pmd->fail_io)) return r; - /* - * Replacement block manager (new_bm) is created and old_bm destroyed outside of - * pmd root_lock to avoid ABBA deadlock that would result (due to life-cycle of - * shrinker associated with the block manager's bufio client vs pmd root_lock). - * - must take shrinker_rwsem without holding pmd->root_lock - */ - new_bm = dm_block_manager_create(pmd->bdev, THIN_METADATA_BLOCK_SIZE << SECTOR_SHIFT, - THIN_MAX_CONCURRENT_LOCKS); - pmd_write_lock(pmd); if (pmd->fail_io) { pmd_write_unlock(pmd); - goto out; + return r; } - __set_abort_with_changes_flags(pmd); + + /* destroy data_sm/metadata_sm/nb_tm/tm */ __destroy_persistent_data_objects(pmd, false); - old_bm = pmd->bm; - if (IS_ERR(new_bm)) { - DMERR("could not create block manager during abort"); - pmd->bm = NULL; - r = PTR_ERR(new_bm); - goto out_unlock; - } - pmd->bm = new_bm; + /* reset bm */ + dm_block_manager_reset(pmd->bm); + + /* rebuild data_sm/metadata_sm/nb_tm/tm */ r = __open_or_format_metadata(pmd, false); - if (r) { - pmd->bm = NULL; - goto out_unlock; - } - new_bm = NULL; -out_unlock: if (r) pmd->fail_io = true; - pmd_write_unlock(pmd); - dm_block_manager_destroy(old_bm); -out: - if (new_bm && !IS_ERR(new_bm)) - dm_block_manager_destroy(new_bm); + pmd_write_unlock(pmd); return r; } diff --git a/drivers/md/persistent-data/dm-block-manager.c b/drivers/md/persistent-data/dm-block-manager.c index 54c089a50b15..a9db18a73dc2 100644 --- a/drivers/md/persistent-data/dm-block-manager.c +++ b/drivers/md/persistent-data/dm-block-manager.c @@ -414,6 +414,12 @@ void dm_block_manager_destroy(struct dm_block_manager *bm) } EXPORT_SYMBOL_GPL(dm_block_manager_destroy); +void dm_block_manager_reset(struct dm_block_manager *bm) +{ + dm_bufio_client_reset(bm->bufio); +} +EXPORT_SYMBOL_GPL(dm_block_manager_reset); + unsigned dm_bm_block_size(struct dm_block_manager *bm) { return dm_bufio_get_block_size(bm->bufio); diff --git a/drivers/md/persistent-data/dm-block-manager.h b/drivers/md/persistent-data/dm-block-manager.h index e728937f376a..644cd7015e28 100644 --- a/drivers/md/persistent-data/dm-block-manager.h +++ b/drivers/md/persistent-data/dm-block-manager.h @@ -35,6 +35,7 @@ struct dm_block_manager *dm_block_manager_create( struct block_device *bdev, unsigned block_size, unsigned max_held_per_thread); void dm_block_manager_destroy(struct dm_block_manager *bm); +void dm_block_manager_reset(struct dm_block_manager *bm); unsigned dm_bm_block_size(struct dm_block_manager *bm); dm_block_t dm_bm_nr_blocks(struct dm_block_manager *bm); diff --git a/drivers/md/persistent-data/dm-space-map.h b/drivers/md/persistent-data/dm-space-map.h index 3e6d1153b7c4..f300aece842a 100644 --- a/drivers/md/persistent-data/dm-space-map.h +++ b/drivers/md/persistent-data/dm-space-map.h @@ -76,7 +76,8 @@ struct dm_space_map { static inline void dm_sm_destroy(struct dm_space_map *sm) { - sm->destroy(sm); + if (sm) + sm->destroy(sm); } static inline int dm_sm_extend(struct dm_space_map *sm, dm_block_t extra_blocks) diff --git a/drivers/md/persistent-data/dm-transaction-manager.c b/drivers/md/persistent-data/dm-transaction-manager.c index abe2c5dd0993..73f57977b2e3 100644 --- a/drivers/md/persistent-data/dm-transaction-manager.c +++ b/drivers/md/persistent-data/dm-transaction-manager.c @@ -197,6 +197,9 @@ EXPORT_SYMBOL_GPL(dm_tm_create_non_blocking_clone); void dm_tm_destroy(struct dm_transaction_manager *tm) { + if (!tm) + return; + if (!tm->is_clone) wipe_shadow_table(tm); diff --git a/include/linux/dm-bufio.h b/include/linux/dm-bufio.h index 90bd558a17f5..c93917b2f54b 100644 --- a/include/linux/dm-bufio.h +++ b/include/linux/dm-bufio.h @@ -31,6 +31,8 @@ dm_bufio_client_create(struct block_device *bdev, unsigned block_size, */ void dm_bufio_client_destroy(struct dm_bufio_client *c); +void dm_bufio_client_reset(struct dm_bufio_client *c); + /* * Set the sector range. * When this function is called, there must be no I/O in progress on the bufio -- 2.31.1

2 1

[PATCH OLK-5.10 v4] Revert "ext4: dio take shared inode lock when overwriting preallocated blocks"
by Baokun Li 05 Jun '23

05 Jun '23

hulk inclusion category: perf bugzilla: 188836, https://gitee.com/openeuler/kernel/issues/I7AYNZ -------------------------------- This reverts commit 5193a88ea78c9d4a61752443cd7e7457eeeed3c6. This commit may cause performance degradation, so it is being reverted temporarily. Signed-off-by: Baokun Li <libaokun1(a)huawei.com> --- V1->V2: Modify patch description. V2->V3: Add maillist kernel(a)openeuler.org. V3->V4: subscribe kernel(a)openeuler.org. fs/ext4/file.c | 34 ++++++++++++---------------------- 1 file changed, 12 insertions(+), 22 deletions(-) diff --git a/fs/ext4/file.c b/fs/ext4/file.c index 43076e72cd76..aeb169354a53 100644 --- a/fs/ext4/file.c +++ b/fs/ext4/file.c @@ -190,9 +190,8 @@ ext4_extending_io(struct inode *inode, loff_t offset, size_t len) return false; } -/* Is IO overwriting allocated or initialized blocks? */ -static bool ext4_overwrite_io(struct inode *inode, - loff_t pos, loff_t len, bool *unwritten) +/* Is IO overwriting allocated and initialized blocks? */ +static bool ext4_overwrite_io(struct inode *inode, loff_t pos, loff_t len) { struct ext4_map_blocks map; unsigned int blkbits = inode->i_blkbits; @@ -206,15 +205,12 @@ static bool ext4_overwrite_io(struct inode *inode, blklen = map.m_len; err = ext4_map_blocks(NULL, inode, &map, 0); - if (err != blklen) - return false; /* * 'err==len' means that all of the blocks have been preallocated, - * regardless of whether they have been initialized or not. We need to - * check m_flags to distinguish the unwritten extents. + * regardless of whether they have been initialized or not. To exclude + * unwritten extents, we need to check m_flags. */ - *unwritten = !(map.m_flags & EXT4_MAP_MAPPED); - return true; + return err == blklen && (map.m_flags & EXT4_MAP_MAPPED); } static ssize_t ext4_generic_write_checks(struct kiocb *iocb, @@ -425,16 +421,11 @@ static const struct iomap_dio_ops ext4_dio_write_ops = { * - For extending writes case we don't take the shared lock, since it requires * updating inode i_disksize and/or orphan handling with exclusive lock. * - * - shared locking will only be true mostly with overwrites, including - * initialized blocks and unwritten blocks. For overwrite unwritten blocks - * we protect splitting extents by i_data_sem in ext4_inode_info, so we can - * also release exclusive i_rwsem lock. - * - * - Otherwise we will switch to exclusive i_rwsem lock. + * - shared locking will only be true mostly with overwrites. Otherwise we will + * switch to exclusive i_rwsem lock. */ static ssize_t ext4_dio_write_checks(struct kiocb *iocb, struct iov_iter *from, - bool *ilock_shared, bool *extend, - bool *unwritten) + bool *ilock_shared, bool *extend) { struct file *file = iocb->ki_filp; struct inode *inode = file_inode(file); @@ -458,7 +449,7 @@ static ssize_t ext4_dio_write_checks(struct kiocb *iocb, struct iov_iter *from, * in file_modified(). */ if (*ilock_shared && (!IS_NOSEC(inode) || *extend || - !ext4_overwrite_io(inode, offset, count, unwritten))) { + !ext4_overwrite_io(inode, offset, count))) { if (iocb->ki_flags & IOCB_NOWAIT) { ret = -EAGAIN; goto out; @@ -490,7 +481,7 @@ static ssize_t ext4_dio_write_iter(struct kiocb *iocb, struct iov_iter *from) loff_t offset = iocb->ki_pos; size_t count = iov_iter_count(from); const struct iomap_ops *iomap_ops = &ext4_iomap_ops; - bool extend = false, unaligned_io = false, unwritten = false; + bool extend = false, unaligned_io = false; bool ilock_shared = true; /* @@ -533,8 +524,7 @@ static ssize_t ext4_dio_write_iter(struct kiocb *iocb, struct iov_iter *from) return ext4_buffered_write_iter(iocb, from); } - ret = ext4_dio_write_checks(iocb, from, - &ilock_shared, &extend, &unwritten); + ret = ext4_dio_write_checks(iocb, from, &ilock_shared, &extend); if (ret <= 0) return ret; @@ -584,7 +574,7 @@ static ssize_t ext4_dio_write_iter(struct kiocb *iocb, struct iov_iter *from) ext4_journal_stop(handle); } - if (ilock_shared && !unwritten) + if (ilock_shared) iomap_ops = &ext4_iomap_overwrite_ops; ret = iomap_dio_rw(iocb, from, iomap_ops, &ext4_dio_write_ops, is_sync_kiocb(iocb) || unaligned_io || extend); -- 2.31.1

2 1

[PATCH openEuler-1.0-LTS 0/2] Balance initial LPI affinity across CPUs
by Zhang Changzhong 05 Jun '23

05 Jun '23

Marc Zyngier (2): irqchip/gic-v3-its: Track LPI distribution on a per CPU basis irqchip/gic-v3-its: Balance initial LPI affinity across CPUs drivers/irqchip/irq-gic-v3-its.c | 173 ++++++++++++++++++++++++++++++++------- 1 file changed, 145 insertions(+), 28 deletions(-) -- 2.9.5

1 2

[PATCH OLK-5.10] ipv6: Add lwtunnel encap size of all siblings in nexthop calculation
by Lu Wei 05 Jun '23

05 Jun '23

stable inclusion from stable-v5.10.173 commit da26369377f0b671c14692e2d65ceb38131053e1 category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I6GT9T Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit 4cc59f386991ec9374cb4bc83dbe1c0b5a95033f ] In function rt6_nlmsg_size(), the length of nexthop is calculated by multipling the nexthop length of fib6_info and the number of siblings. However if the fib6_info has no lwtunnel but the siblings have lwtunnels, the nexthop length is less than it should be, and it will trigger a warning in inet6_rt_notify() as follows: WARNING: CPU: 0 PID: 6082 at net/ipv6/route.c:6180 inet6_rt_notify+0x120/0x130 ...... Call Trace: <TASK> fib6_add_rt2node+0x685/0xa30 fib6_add+0x96/0x1b0 ip6_route_add+0x50/0xd0 inet6_rtm_newroute+0x97/0xa0 rtnetlink_rcv_msg+0x156/0x3d0 netlink_rcv_skb+0x5a/0x110 netlink_unicast+0x246/0x350 netlink_sendmsg+0x250/0x4c0 sock_sendmsg+0x66/0x70 ___sys_sendmsg+0x7c/0xd0 __sys_sendmsg+0x5d/0xb0 do_syscall_64+0x3f/0x90 entry_SYSCALL_64_after_hwframe+0x72/0xdc This bug can be reproduced by script: ip -6 addr add 2002::2/64 dev ens2 ip -6 route add 100::/64 via 2002::1 dev ens2 metric 100 for i in 10 20 30 40 50 60 70; do ip link add link ens2 name ipv_$i type ipvlan ip -6 addr add 2002::$i/64 dev ipv_$i ifconfig ipv_$i up done for i in 10 20 30 40 50 60; do ip -6 route append 100::/64 encap ip6 dst 2002::$i via 2002::1 dev ipv_$i metric 100 done ip -6 route append 100::/64 via 2002::1 dev ipv_70 metric 100 This patch fixes it by adding nexthop_len of every siblings using rt6_nh_nlmsg_size(). Fixes: beb1afac518d ("net: ipv6: Add support to dump multipath routes via RTA_MULTIPATH attribute") Signed-off-by: Lu Wei <luwei32(a)huawei.com> Reviewed-by: David Ahern <dsahern(a)kernel.org> Link: https://lore.kernel.org/r/20230222083629.335683-2-luwei32@huawei.com Signed-off-by: Paolo Abeni <pabeni(a)redhat.com> Signed-off-by: Sasha Levin <sashal(a)kernel.org> Signed-off-by: Lu Wei <luwei32(a)huawei.com> --- net/ipv6/route.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) diff --git a/net/ipv6/route.c b/net/ipv6/route.c index c18a6c17ce09..b5e7ce4909d5 100644 --- a/net/ipv6/route.c +++ b/net/ipv6/route.c @@ -5439,16 +5439,17 @@ static size_t rt6_nlmsg_size(struct fib6_info *f6i) nexthop_for_each_fib6_nh(f6i->nh, rt6_nh_nlmsg_size, &nexthop_len); } else { + struct fib6_info *sibling, *next_sibling; struct fib6_nh *nh = f6i->fib6_nh; nexthop_len = 0; if (f6i->fib6_nsiblings) { - nexthop_len = nla_total_size(0) /* RTA_MULTIPATH */ - + NLA_ALIGN(sizeof(struct rtnexthop)) - + nla_total_size(16) /* RTA_GATEWAY */ - + lwtunnel_get_encap_size(nh->fib_nh_lws); + rt6_nh_nlmsg_size(nh, &nexthop_len); - nexthop_len *= f6i->fib6_nsiblings; + list_for_each_entry_safe(sibling, next_sibling, + &f6i->fib6_siblings, fib6_siblings) { + rt6_nh_nlmsg_size(sibling->fib6_nh, &nexthop_len); + } } nexthop_len += lwtunnel_get_encap_size(nh->fib_nh_lws); } -- 2.31.1

1 0

[PATCH OLK-5.10 0/4] Accelerator some bugs round
by Weili Qian 05 Jun '23

05 Jun '23

From: JiangShui Yang <yangjiangshui(a)h-partners.com> Misc fixes for Kunpeng accelerator drivers. Bjorn Helgaas (1): crypto: hisilicon/qm - remove unnecessary aer.h include Weili Qian (3): crypto: hisilicon/qm - prevent soft lockup in qm_poll_req_cb()'s loop crypto: hisilicon/hpre - ensure private key less than n crypto: hisilicon/qm - stop function and write data to memory drivers/crypto/hisilicon/hpre/hpre_crypto.c | 10 ++++++++-- drivers/crypto/hisilicon/qm.c | 16 +++++++++------- include/linux/hisi_acc_qm.h | 2 +- 3 files changed, 18 insertions(+), 10 deletions(-) -- 2.30.0

1 4

[PATCH OLK-5.10] tcp/dccp: Add another way to allocate local ports in connect()
by Liu Jian 05 Jun '23

05 Jun '23

From: Lu Wei <luwei32(a)huawei.com> hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I7AO8G CVE: NA -------------------------------- Commit 07f4c90062f8 ("tcp/dccp: try to not exhaust ip_local_port_range in connect()") allocates even ports for connect() first while leaving odd ports for bind() and this works well in busy servers. But this strategy causes severe performance degradation in busy clients. when a client has used more than half of the local ports setted in proc/sys/net/ipv4/ip_local_port_range, if this client trys to connect to a server again, the connect time increases rapidly since it will traverse all the even ports though they are exhausted. So this path provides another strategy by introducing a system option: local_port_allocation. If it is a busy client, users should set it to 1 to use sequential allocation while it should be set to 0 in other situations. Its default value is 0. Signed-off-by: Lu Wei <luwei32(a)huawei.com> Signed-off-by: Liu Jian <liujian56(a)huawei.com> --- include/net/tcp.h | 1 + net/ipv4/inet_hashtables.c | 11 ++++++++--- net/ipv4/sysctl_net_ipv4.c | 8 ++++++++ 3 files changed, 17 insertions(+), 3 deletions(-) diff --git a/include/net/tcp.h b/include/net/tcp.h index 9c86ff1e3ed1..1b832683ccf8 100644 --- a/include/net/tcp.h +++ b/include/net/tcp.h @@ -258,6 +258,7 @@ extern long sysctl_tcp_mem[3]; extern atomic_long_t tcp_memory_allocated; extern struct percpu_counter tcp_sockets_allocated; extern unsigned long tcp_memory_pressure; +extern int sysctl_local_port_allocation; /* optimized version of sk_under_memory_pressure() for TCP sockets */ static inline bool tcp_under_memory_pressure(const struct sock *sk) diff --git a/net/ipv4/inet_hashtables.c b/net/ipv4/inet_hashtables.c index ed8324ee4b1b..396a58d9708c 100644 --- a/net/ipv4/inet_hashtables.c +++ b/net/ipv4/inet_hashtables.c @@ -742,7 +742,7 @@ int __inet_hash_connect(struct inet_timewait_death_row *death_row, struct net *net = sock_net(sk); struct inet_bind_bucket *tb; u32 remaining, offset; - int ret, i, low, high; + int ret, i, low, high, span_size; int l3mdev; u32 index; @@ -762,6 +762,11 @@ int __inet_hash_connect(struct inet_timewait_death_row *death_row, local_bh_enable(); return ret; } + /* local_port_allocation 0 means even and odd port allocation strategy + * will be applied, so span size is 2; otherwise sequential allocation + * will be used and span size is 1. Default value is 0. + */ + span_size = sysctl_local_port_allocation ? 1 : 2; l3mdev = inet_sk_bound_l3mdev(sk); @@ -783,7 +788,7 @@ int __inet_hash_connect(struct inet_timewait_death_row *death_row, offset &= ~1U; other_parity_scan: port = low + offset; - for (i = 0; i < remaining; i += 2, port += 2) { + for (i = 0; i < remaining; i += span_size, port += span_size) { if (unlikely(port >= high)) port -= remaining; if (inet_is_local_reserved_port(net, port)) @@ -824,7 +829,7 @@ int __inet_hash_connect(struct inet_timewait_death_row *death_row, } offset++; - if ((offset & 1) && remaining > 1) + if ((offset & 1) && remaining > 1 && span_size == 2) goto other_parity_scan; return -EADDRNOTAVAIL; diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c index dee47ef05655..d82c0eceec91 100644 --- a/net/ipv4/sysctl_net_ipv4.c +++ b/net/ipv4/sysctl_net_ipv4.c @@ -51,6 +51,7 @@ static int comp_sack_nr_max = 255; static u32 u32_max_div_HZ = UINT_MAX / HZ; static int one_day_secs = 24 * 3600; +int sysctl_local_port_allocation; /* obsolete */ static int sysctl_tcp_low_latency __read_mostly; @@ -575,6 +576,13 @@ static struct ctl_table ipv4_table[] = { .extra2 = SYSCTL_ONE, }, #endif + { + .procname = "local_port_allocation", + .data = &sysctl_local_port_allocation, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = proc_dointvec, + }, { } }; -- 2.34.1

1 0

[patch OLK-5.10 0/2] support ACPI for MPAM 2.0
by Yu Liao 03 Jun '23

03 Jun '23

This patch series support ACPI for MPAM 2.0 Hesham Almatary (1): ACPICA: Add support for Arm's MPAM ACPI table version 2 Yu Liao (1): ACPI/MPAM: Adapt to Arm's MPAM ACPI table version 2 arch/arm64/kernel/mpam/mpam_device.c | 2 +- drivers/acpi/arm64/Makefile | 2 +- drivers/acpi/arm64/mpam.c | 19 ++- drivers/acpi/arm64/mpam_v2.c | 188 +++++++++++++++++++++++++++ include/acpi/actbl2.h | 111 ++++++++++++++++ include/linux/arm_mpam.h | 2 +- 6 files changed, 318 insertions(+), 6 deletions(-) create mode 100644 drivers/acpi/arm64/mpam_v2.c -- 2.25.1

1 2