- Kernel - mailweb.openeuler.org

[PATCH OLK-5.10 0/3] Some bugfixs for ubi/fs
by Zhihao Cheng 15 Oct '24

15 Oct '24

From: Zhihao Cheng <chengzhihao1(a)huawei.com> Amir Goldstein (1): locks: fix TOCTOU race when granting write lease Li Nan (1): ubi: block: fix null-pointer-dereference in ubiblock_create() Li zeming (1): ubi: block: Remove in vain semicolon drivers/mtd/ubi/block.c | 9 +++++---- fs/file_table.c | 7 +------ fs/internal.h | 10 ++++++++++ fs/open.c | 11 ++++------- 4 files changed, 20 insertions(+), 17 deletions(-) -- 2.39.2

2 4

[PATCH v2 OLK-5.10] atm: idt77252: prevent use after free in dequeue_rx()
by Peng Zhang 15 Oct '24

15 Oct '24

From: Dan Carpenter <dan.carpenter(a)linaro.org> stable inclusion from stable-v5.10.225 commit 1cece837e387c039225f19028df255df87a97c0d category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/IAOXZG CVE: CVE-2024-44998 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit a9a18e8f770c9b0703dab93580d0b02e199a4c79 ] We can't dereference "skb" after calling vcc->push() because the skb is released. Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Signed-off-by: Dan Carpenter <dan.carpenter(a)linaro.org> Signed-off-by: David S. Miller <davem(a)davemloft.net> Signed-off-by: Sasha Levin <sashal(a)kernel.org> Signed-off-by: ZhangPeng <zhangpeng362(a)huawei.com> --- drivers/atm/idt77252.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/drivers/atm/idt77252.c b/drivers/atm/idt77252.c index e616e33c8a20..25fd73fafb37 100644 --- a/drivers/atm/idt77252.c +++ b/drivers/atm/idt77252.c @@ -1118,8 +1118,8 @@ dequeue_rx(struct idt77252_dev *card, struct rsq_entry *rsqe) rpp->len += skb->len; if (stat & SAR_RSQE_EPDU) { + unsigned int len, truesize; unsigned char *l1l2; - unsigned int len; l1l2 = (unsigned char *) ((unsigned long) skb->data + skb->len - 6); @@ -1189,14 +1189,15 @@ dequeue_rx(struct idt77252_dev *card, struct rsq_entry *rsqe) ATM_SKB(skb)->vcc = vcc; __net_timestamp(skb); + truesize = skb->truesize; vcc->push(vcc, skb); atomic_inc(&vcc->stats->rx); - if (skb->truesize > SAR_FB_SIZE_3) + if (truesize > SAR_FB_SIZE_3) add_rx_skb(card, 3, SAR_FB_SIZE_3, 1); - else if (skb->truesize > SAR_FB_SIZE_2) + else if (truesize > SAR_FB_SIZE_2) add_rx_skb(card, 2, SAR_FB_SIZE_2, 1); - else if (skb->truesize > SAR_FB_SIZE_1) + else if (truesize > SAR_FB_SIZE_1) add_rx_skb(card, 1, SAR_FB_SIZE_1, 1); else add_rx_skb(card, 0, SAR_FB_SIZE_0, 1); -- 2.25.1

2 1

[PATCH v2 openEuler-22.03-LTS-SP1] atm: idt77252: prevent use after free in dequeue_rx()
by Peng Zhang 15 Oct '24

15 Oct '24

From: Dan Carpenter <dan.carpenter(a)linaro.org> stable inclusion from stable-v5.10.225 commit 1cece837e387c039225f19028df255df87a97c0d category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/IAOXZG CVE: CVE-2024-44998 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit a9a18e8f770c9b0703dab93580d0b02e199a4c79 ] We can't dereference "skb" after calling vcc->push() because the skb is released. Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Signed-off-by: Dan Carpenter <dan.carpenter(a)linaro.org> Signed-off-by: David S. Miller <davem(a)davemloft.net> Signed-off-by: Sasha Levin <sashal(a)kernel.org> Signed-off-by: ZhangPeng <zhangpeng362(a)huawei.com> --- drivers/atm/idt77252.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/drivers/atm/idt77252.c b/drivers/atm/idt77252.c index e616e33c8a20..25fd73fafb37 100644 --- a/drivers/atm/idt77252.c +++ b/drivers/atm/idt77252.c @@ -1118,8 +1118,8 @@ dequeue_rx(struct idt77252_dev *card, struct rsq_entry *rsqe) rpp->len += skb->len; if (stat & SAR_RSQE_EPDU) { + unsigned int len, truesize; unsigned char *l1l2; - unsigned int len; l1l2 = (unsigned char *) ((unsigned long) skb->data + skb->len - 6); @@ -1189,14 +1189,15 @@ dequeue_rx(struct idt77252_dev *card, struct rsq_entry *rsqe) ATM_SKB(skb)->vcc = vcc; __net_timestamp(skb); + truesize = skb->truesize; vcc->push(vcc, skb); atomic_inc(&vcc->stats->rx); - if (skb->truesize > SAR_FB_SIZE_3) + if (truesize > SAR_FB_SIZE_3) add_rx_skb(card, 3, SAR_FB_SIZE_3, 1); - else if (skb->truesize > SAR_FB_SIZE_2) + else if (truesize > SAR_FB_SIZE_2) add_rx_skb(card, 2, SAR_FB_SIZE_2, 1); - else if (skb->truesize > SAR_FB_SIZE_1) + else if (truesize > SAR_FB_SIZE_1) add_rx_skb(card, 1, SAR_FB_SIZE_1, 1); else add_rx_skb(card, 0, SAR_FB_SIZE_0, 1); -- 2.25.1

2 1

[PATCH OLK-5.10 v6 00/12] xfs: some fix for forcealign
by Long Li 15 Oct '24

15 Oct '24

This patch set fix some bug for forcealign: patch 1 ~ 6 : fix tail alignment issue while approach to no space. patch 7 : fix forcealign not compatible with reflink and realtime. patch 8 : only datafork need bunmap algin for focealign patch 9 ~ 12 : fix truncate for forcealign Dave Chinner (3): xfs: only allow minlen allocations when near ENOSPC xfs: always tail align maxlen allocations xfs: align args->minlen for forced allocation alignment John Garry (1): xfs: Don't revert allocated offset for forcealign Long Li (5): xfs: don't attempting non-aligned fallbacks alloc for forcealign xfs: simplify extent allocation alignment xfs: forcealign not compatible with reflink and realtime device xfs: only bunmap align in datafork for forcealign xfs: correct the truncate blocksize of forcealign Zhang Yi (3): math64: add rem_u64() to just return the remainder iomap: pass blocksize to iomap_truncate_page() xfs: refactor the truncating order fs/iomap/buffered-io.c | 8 +-- fs/xfs/libxfs/xfs_alloc.c | 31 +++++---- fs/xfs/libxfs/xfs_bmap.c | 140 ++++++++++++++++++++------------------ fs/xfs/xfs_iops.c | 124 +++++++++++++++++---------------- fs/xfs/xfs_super.c | 19 +++++- include/linux/iomap.h | 4 +- include/linux/math64.h | 24 +++++++ 7 files changed, 208 insertions(+), 142 deletions(-) -- 2.39.2

2 13

[openeuler:OLK-5.10 22909/30000] kernel/sched/topology.c:1749:9: error: implicit declaration of function 'register_sysctl_init'; did you mean 'register_sysctl'?
by kernel test robot 15 Oct '24

15 Oct '24

tree: https://gitee.com/openeuler/kernel.git OLK-5.10 head: 5f264976b651040f3fac9a10fca21c542662709b commit: f4128ae8e6ff03b4c805707fe75d0345797f7f53 [22909/30000] sched: Add cluster scheduler level for x86 config: x86_64-buildonly-randconfig-002-20241013 (https://download.01.org/0day-ci/archive/20241015/202410150906.0KpoaIkq-lkp@…) compiler: gcc-12 (Debian 12.2.0-14) 12.2.0 reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20241015/202410150906.0KpoaIkq-lkp@…) If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags | Reported-by: kernel test robot <lkp(a)intel.com> | Closes: https://lore.kernel.org/oe-kbuild-all/202410150906.0KpoaIkq-lkp@intel.com/ All errors (new ones prefixed by >>): kernel/sched/topology.c: In function 'sched_cluster_sysctl_init': >> kernel/sched/topology.c:1749:9: error: implicit declaration of function 'register_sysctl_init'; did you mean 'register_sysctl'? [-Werror=implicit-function-declaration] 1749 | register_sysctl_init("kernel", sched_cluster_sysctls); | ^~~~~~~~~~~~~~~~~~~~ | register_sysctl cc1: some warnings being treated as errors vim +1749 kernel/sched/topology.c 8ce3e706b31409 Tim Chen 2021-12-03 1746 8ce3e706b31409 Tim Chen 2021-12-03 1747 static int __init sched_cluster_sysctl_init(void) 8ce3e706b31409 Tim Chen 2021-12-03 1748 { 8ce3e706b31409 Tim Chen 2021-12-03 @1749 register_sysctl_init("kernel", sched_cluster_sysctls); 8ce3e706b31409 Tim Chen 2021-12-03 1750 return 0; 8ce3e706b31409 Tim Chen 2021-12-03 1751 } 8ce3e706b31409 Tim Chen 2021-12-03 1752 late_initcall(sched_cluster_sysctl_init); 9e68cc2bf535a2 Tim Chen 2021-12-03 1753 :::::: The code at line 1749 was first introduced by commit :::::: 8ce3e706b31409147f035c037055caa68e450ce5 scheduler: Add runtime knob sysctl_sched_cluster :::::: TO: Tim Chen <tim.c.chen(a)linux.intel.com> :::::: CC: Jie Liu <liujie375(a)h-partners.com> -- 0-DAY CI Kernel Test Service https://github.com/intel/lkp-tests/wiki

1 0

[PATCH OLK-5.10 0/2] Backport "ima: Avoid blocking in RCU read-side critical section"
by Gu Bowen 15 Oct '24

15 Oct '24

Revert self-developed solution and backport mainline one. GUO Zihua (1): ima: Avoid blocking in RCU read-side critical section Gu Bowen (1): Revert "ima: Avoid blocking in RCU read-side critical section" include/linux/lsm_hook_defs.h | 2 +- include/linux/security.h | 5 +++-- kernel/auditfilter.c | 5 +++-- security/apparmor/audit.c | 6 +++--- security/apparmor/include/audit.h | 2 +- security/integrity/ima/ima.h | 2 +- security/integrity/ima/ima_policy.c | 15 +++++++++------ security/security.c | 6 ++++-- security/selinux/include/audit.h | 4 +++- security/selinux/ss/services.c | 5 +++-- security/smack/smack_lsm.c | 4 +++- 11 files changed, 34 insertions(+), 22 deletions(-) -- 2.25.1

2 3

[PATCH openEuler-22.03-LTS-SP1] spi: hisi-kunpeng: Add verification for the max_frequency provided by the firmware
by Yongjian Sun 14 Oct '24

14 Oct '24

From: Devyn Liu <liudingyuan(a)huawei.com> mainline inclusion from mainline-v6.11-rc3 commit 5127c42c77de18651aa9e8e0a3ced190103b449c category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/IAVU80 CVE: CVE-2024-47664 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- If the value of max_speed_hz is 0, it may cause a division by zero error in hisi_calc_effective_speed(). The value of max_speed_hz is provided by firmware. Firmware is generally considered as a trusted domain. However, as division by zero errors can cause system failure, for defense measure, the value of max_speed is validated here. So 0 is regarded as invalid and an error code is returned. Signed-off-by: Devyn Liu <liudingyuan(a)huawei.com> Reviewed-by: Jay Fang <f.fangjian(a)huawei.com> Link: https://patch.msgid.link/20240730032040.3156393-3-liudingyuan@huawei.com Signed-off-by: Mark Brown <broonie(a)kernel.org> Conflicts: drivers/spi/spi-hisi-kunpeng.c [The author changed the variable naming, replacing all instances of "master" with "host", which aligns with modern naming conventions.] Signed-off-by: Yongjian Sun <sunyongjian1(a)huawei.com> --- drivers/spi/spi-hisi-kunpeng.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/spi/spi-hisi-kunpeng.c b/drivers/spi/spi-hisi-kunpeng.c index 525cc0143a30..54e05fb0387e 100644 --- a/drivers/spi/spi-hisi-kunpeng.c +++ b/drivers/spi/spi-hisi-kunpeng.c @@ -483,6 +483,9 @@ static int hisi_spi_probe(struct platform_device *pdev) return -EINVAL; } + if (master->max_speed_hz == 0) + return dev_err_probe(dev, -EINVAL, "spi-max-frequency can't be 0\n"); + ret = device_property_read_u16(dev, "num-cs", &master->num_chipselect); if (ret) -- 2.39.2

2 1

[openeuler:OLK-5.10 15571/30000] mm/memcontrol.c:5402:40: error: implicit declaration of function 'ksm_process_profit'
by kernel test robot 14 Oct '24

14 Oct '24

Hi Nanyong, FYI, the error/warning still remains. tree: https://gitee.com/openeuler/kernel.git OLK-5.10 head: 748eeb7da1ecddfb1ace5ec61a512befe217702e commit: 0f6fb3570be40b783d8c756d64297be1a8e3d2b3 [15571/30000] memcg: support ksm merge any mode per cgroup config: x86_64-buildonly-randconfig-002-20241013 (https://download.01.org/0day-ci/archive/20241014/202410142209.nmkFxV8P-lkp@…) compiler: gcc-12 (Debian 12.2.0-14) 12.2.0 reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20241014/202410142209.nmkFxV8P-lkp@…) If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags | Reported-by: kernel test robot <lkp(a)intel.com> | Closes: https://lore.kernel.org/oe-kbuild-all/202410142209.nmkFxV8P-lkp@intel.com/ All errors (new ones prefixed by >>): mm/memcontrol.c: In function 'memory_ksm_show': >> mm/memcontrol.c:5402:40: error: implicit declaration of function 'ksm_process_profit' [-Werror=implicit-function-declaration] 5402 | ksm_process_profits += ksm_process_profit(mm); | ^~~~~~~~~~~~~~~~~~ cc1: some warnings being treated as errors vim +/ksm_process_profit +5402 mm/memcontrol.c 5379 5380 static int memory_ksm_show(struct seq_file *m, void *v) 5381 { 5382 unsigned long ksm_merging_pages = 0; 5383 unsigned long ksm_rmap_items = 0; 5384 long ksm_process_profits = 0; 5385 unsigned int tasks = 0; 5386 struct task_struct *task; 5387 struct mm_struct *mm; 5388 struct css_task_iter it; 5389 struct mem_cgroup *memcg = mem_cgroup_from_seq(m); 5390 5391 css_task_iter_start(&memcg->css, CSS_TASK_ITER_PROCS, &it); 5392 while ((task = css_task_iter_next(&it))) { 5393 mm = get_task_mm(task); 5394 if (!mm) 5395 continue; 5396 5397 if (test_bit(MMF_VM_MERGE_ANY, &mm->flags)) 5398 tasks++; 5399 5400 ksm_rmap_items += mm->ksm_rmap_items; 5401 ksm_merging_pages += mm->ksm_merging_pages; > 5402 ksm_process_profits += ksm_process_profit(mm); 5403 mmput(mm); 5404 } 5405 css_task_iter_end(&it); 5406 5407 seq_printf(m, "merge any state: %d\n", READ_ONCE(memcg->ksm_merge_any)); 5408 seq_printf(m, "merge any tasks: %u\n", tasks); 5409 seq_printf(m, "ksm_rmap_items %lu\n", ksm_rmap_items); 5410 seq_printf(m, "ksm_merging_pages %lu\n", ksm_merging_pages); 5411 seq_printf(m, "ksm_process_profits %ld\n", ksm_process_profits); 5412 return 0; 5413 } 5414 -- 0-DAY CI Kernel Test Service https://github.com/intel/lkp-tests/wiki

1 0

[PATCH OLK-5.10] netfilter: nf_tables: use timestamp to check for set element timeout
by Liu Jian 14 Oct '24

14 Oct '24

From: Pablo Neira Ayuso <pablo(a)netfilter.org> mainline inclusion from mainline-v6.8-rc4 commit 7395dfacfff65e9938ac0889dafa1ab01e987d15 category: bugfix bugzilla: 189924 CVE: CVE-2024-27397 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… --------------------------- Add a timestamp field at the beginning of the transaction, store it in the nftables per-netns area. Update set backend .insert, .deactivate and sync gc path to use the timestamp, this avoids that an element expires while control plane transaction is still unfinished. .lookup and .update, which are used from packet path, still use the current time to check if the element has expired. And .get path and dump also since this runs lockless under rcu read size lock. Then, there is async gc which also needs to check the current time since it runs asynchronously from a workqueue. Fixes: c3e1b005ed1c ("netfilter: nf_tables: add set element timeout support") Signed-off-by: Pablo Neira Ayuso <pablo(a)netfilter.org> Conflicts: include/net/netfilter/nf_tables.h net/netfilter/nf_tables_api.c net/netfilter/nft_set_pipapo.c net/netfilter/nft_set_rbtree.c [Do not backport nft_rbtree_gc() part, it is a async gc in our repository. Others caused by we did not backport commit ab482c6b66a4, b2d306542ff9, 8837ba3e58ea, 5f68718b34a5, d59d2f82f984.] Signed-off-by: Liu Jian <liujian56(a)huawei.com> --- include/net/netfilter/nf_tables.h | 24 ++++++++++++++++++++++-- net/netfilter/nf_tables_api.c | 1 + net/netfilter/nft_set_hash.c | 8 +++++++- net/netfilter/nft_set_pipapo.c | 18 +++++++++++------- net/netfilter/nft_set_rbtree.c | 6 ++++-- 5 files changed, 45 insertions(+), 12 deletions(-) diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h index 17659138e15a..a6e221de31ef 100644 --- a/include/net/netfilter/nf_tables.h +++ b/include/net/netfilter/nf_tables.h @@ -13,6 +13,7 @@ #include <net/netfilter/nf_flow_table.h> #include <net/netlink.h> #include <net/flow_offload.h> +#include <net/netns/generic.h> #define NFT_MAX_HOOKS (NF_INET_INGRESS + 1) @@ -700,10 +701,16 @@ static inline struct nft_expr *nft_set_ext_expr(const struct nft_set_ext *ext) return nft_set_ext(ext, NFT_SET_EXT_EXPR); } -static inline bool nft_set_elem_expired(const struct nft_set_ext *ext) +static inline bool __nft_set_elem_expired(const struct nft_set_ext *ext, + u64 tstamp) { return nft_set_ext_exists(ext, NFT_SET_EXT_EXPIRATION) && - time_is_before_eq_jiffies64(*nft_set_ext_expiration(ext)); + time_after_eq64(tstamp, *nft_set_ext_expiration(ext)); +} + +static inline bool nft_set_elem_expired(const struct nft_set_ext *ext) +{ + return __nft_set_elem_expired(ext, get_jiffies_64()); } static inline struct nft_set_ext *nft_set_elem_ext(const struct nft_set *set, @@ -1592,9 +1599,22 @@ struct nftables_pernet { struct list_head module_list; struct list_head notify_list; struct mutex commit_mutex; + u64 tstamp; unsigned int base_seq; u8 validate_state; unsigned int gc_seq; }; +extern unsigned int nf_tables_net_id; + +static inline struct nftables_pernet *nft_pernet(const struct net *net) +{ + return net_generic(net, nf_tables_net_id); +} + +static inline u64 nft_net_tstamp(const struct net *net) +{ + return nft_pernet(net)->tstamp; +} + #endif /* _NET_NF_TABLES_H */ diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c index 82526a768e15..81caffd53912 100644 --- a/net/netfilter/nf_tables_api.c +++ b/net/netfilter/nf_tables_api.c @@ -9074,6 +9074,7 @@ static bool nf_tables_valid_genid(struct net *net, u32 genid) bool genid_ok; mutex_lock(&nft_net->commit_mutex); + nft_net->tstamp = get_jiffies_64(); genid_ok = genid == 0 || nft_net->base_seq == genid; if (!genid_ok) diff --git a/net/netfilter/nft_set_hash.c b/net/netfilter/nft_set_hash.c index f4a76ae8a11c..1c52e0559011 100644 --- a/net/netfilter/nft_set_hash.c +++ b/net/netfilter/nft_set_hash.c @@ -38,6 +38,7 @@ struct nft_rhash_cmp_arg { const struct nft_set *set; const u32 *key; u8 genmask; + u64 tstamp; }; static inline u32 nft_rhash_key(const void *data, u32 len, u32 seed) @@ -64,7 +65,7 @@ static inline int nft_rhash_cmp(struct rhashtable_compare_arg *arg, return 1; if (nft_set_elem_is_dead(&he->ext)) return 1; - if (nft_set_elem_expired(&he->ext)) + if (__nft_set_elem_expired(&he->ext, x->tstamp)) return 1; if (!nft_set_elem_active(&he->ext, x->genmask)) return 1; @@ -88,6 +89,7 @@ static bool nft_rhash_lookup(const struct net *net, const struct nft_set *set, .genmask = nft_genmask_cur(net), .set = set, .key = key, + .tstamp = get_jiffies_64(), }; he = rhashtable_lookup(&priv->ht, &arg, nft_rhash_params); @@ -106,6 +108,7 @@ static void *nft_rhash_get(const struct net *net, const struct nft_set *set, .genmask = nft_genmask_cur(net), .set = set, .key = elem->key.val.data, + .tstamp = get_jiffies_64(), }; he = rhashtable_lookup(&priv->ht, &arg, nft_rhash_params); @@ -129,6 +132,7 @@ static bool nft_rhash_update(struct nft_set *set, const u32 *key, .genmask = NFT_GENMASK_ANY, .set = set, .key = key, + .tstamp = get_jiffies_64(), }; he = rhashtable_lookup(&priv->ht, &arg, nft_rhash_params); @@ -172,6 +176,7 @@ static int nft_rhash_insert(const struct net *net, const struct nft_set *set, .genmask = nft_genmask_next(net), .set = set, .key = elem->key.val.data, + .tstamp = nft_net_tstamp(net), }; struct nft_rhash_elem *prev; @@ -214,6 +219,7 @@ static void *nft_rhash_deactivate(const struct net *net, .genmask = nft_genmask_next(net), .set = set, .key = elem->key.val.data, + .tstamp = nft_net_tstamp(net), }; rcu_read_lock(); diff --git a/net/netfilter/nft_set_pipapo.c b/net/netfilter/nft_set_pipapo.c index 2a147e969c66..65ecf486f8d1 100644 --- a/net/netfilter/nft_set_pipapo.c +++ b/net/netfilter/nft_set_pipapo.c @@ -504,6 +504,7 @@ bool nft_pipapo_lookup(const struct net *net, const struct nft_set *set, * @set: nftables API set representation * @data: Key data to be matched against existing elements * @genmask: If set, check that element is active in given genmask + * @tstamp: timestamp to check for expired elements * * This is essentially the same as the lookup function, except that it matches * key data against the uncommitted copy and doesn't use preallocated maps for @@ -513,7 +514,8 @@ bool nft_pipapo_lookup(const struct net *net, const struct nft_set *set, */ static struct nft_pipapo_elem *pipapo_get(const struct net *net, const struct nft_set *set, - const u8 *data, u8 genmask) + const u8 *data, u8 genmask, + u64 tstamp) { struct nft_pipapo_elem *ret = ERR_PTR(-ENOENT); struct nft_pipapo *priv = nft_set_priv(set); @@ -566,7 +568,7 @@ static struct nft_pipapo_elem *pipapo_get(const struct net *net, goto out; if (last) { - if (nft_set_elem_expired(&f->mt[b].e->ext)) + if (__nft_set_elem_expired(&f->mt[b].e->ext, tstamp)) goto next_match; if ((genmask && !nft_set_elem_active(&f->mt[b].e->ext, genmask))) @@ -603,7 +605,7 @@ static void *nft_pipapo_get(const struct net *net, const struct nft_set *set, const struct nft_set_elem *elem, unsigned int flags) { return pipapo_get(net, set, (const u8 *)elem->key.val.data, - nft_genmask_cur(net)); + nft_genmask_cur(net), get_jiffies_64()); } /** @@ -1197,6 +1199,7 @@ static int nft_pipapo_insert(const struct net *net, const struct nft_set *set, struct nft_pipapo *priv = nft_set_priv(set); struct nft_pipapo_match *m = priv->clone; u8 genmask = nft_genmask_next(net); + u64 tstamp = nft_net_tstamp(net); struct nft_pipapo_field *f; const u8 *start_p, *end_p; int i, bsize_max, err = 0; @@ -1206,7 +1209,7 @@ static int nft_pipapo_insert(const struct net *net, const struct nft_set *set, else end = start; - dup = pipapo_get(net, set, start, genmask); + dup = pipapo_get(net, set, start, genmask, tstamp); if (!IS_ERR(dup)) { /* Check if we already have the same exact entry */ const struct nft_data *dup_key, *dup_end; @@ -1228,7 +1231,7 @@ static int nft_pipapo_insert(const struct net *net, const struct nft_set *set, if (PTR_ERR(dup) == -ENOENT) { /* Look for partially overlapping entries */ - dup = pipapo_get(net, set, end, nft_genmask_next(net)); + dup = pipapo_get(net, set, end, nft_genmask_next(net), tstamp); } if (PTR_ERR(dup) != -ENOENT) { @@ -1580,6 +1583,7 @@ static void pipapo_gc(const struct nft_set *_set, struct nft_pipapo_match *m) struct nft_set *set = (struct nft_set *) _set; struct nft_pipapo *priv = nft_set_priv(set); struct net *net = read_pnet(&set->net); + u64 tstamp = nft_net_tstamp(net); int rules_f0, first_rule = 0; struct nft_trans_gc *gc; @@ -1613,7 +1617,7 @@ static void pipapo_gc(const struct nft_set *_set, struct nft_pipapo_match *m) /* synchronous gc never fails, there is no need to set on * NFT_SET_ELEM_DEAD_BIT. */ - if (nft_set_elem_expired(&e->ext)) { + if (__nft_set_elem_expired(&e->ext, tstamp)) { priv->dirty = true; gc = nft_trans_gc_queue_sync(gc, GFP_ATOMIC); @@ -1772,7 +1776,7 @@ static void *pipapo_deactivate(const struct net *net, const struct nft_set *set, { struct nft_pipapo_elem *e; - e = pipapo_get(net, set, data, nft_genmask_next(net)); + e = pipapo_get(net, set, data, nft_genmask_next(net), nft_net_tstamp(net)); if (IS_ERR(e)) return NULL; diff --git a/net/netfilter/nft_set_rbtree.c b/net/netfilter/nft_set_rbtree.c index 323cbb446d19..2d29cf8d51e0 100644 --- a/net/netfilter/nft_set_rbtree.c +++ b/net/netfilter/nft_set_rbtree.c @@ -314,6 +314,7 @@ static int __nft_rbtree_insert(const struct net *net, const struct nft_set *set, struct rb_node *node, *next, *parent, **p, *first = NULL; struct nft_rbtree *priv = nft_set_priv(set); u8 genmask = nft_genmask_next(net); + u64 tstamp = nft_net_tstamp(net); int d, err; /* Descend the tree to search for an existing element greater than the @@ -359,7 +360,7 @@ static int __nft_rbtree_insert(const struct net *net, const struct nft_set *set, continue; /* perform garbage collection to avoid bogus overlap reports. */ - if (nft_set_elem_expired(&rbe->ext)) { + if (__nft_set_elem_expired(&rbe->ext, tstamp)) { err = nft_rbtree_gc_elem(set, priv, rbe); if (err < 0) return err; @@ -533,6 +534,7 @@ static void *nft_rbtree_deactivate(const struct net *net, const struct rb_node *parent = priv->root.rb_node; struct nft_rbtree_elem *rbe, *this = elem->priv; u8 genmask = nft_genmask_next(net); + u64 tstamp = nft_net_tstamp(net); int d; while (parent != NULL) { @@ -553,7 +555,7 @@ static void *nft_rbtree_deactivate(const struct net *net, nft_rbtree_interval_end(this)) { parent = parent->rb_right; continue; - } else if (nft_set_elem_expired(&rbe->ext)) { + } else if (__nft_set_elem_expired(&rbe->ext, tstamp)) { break; } else if (!nft_set_elem_active(&rbe->ext, genmask)) { parent = parent->rb_left; -- 2.34.1

2 1

[PATCH OLK-6.6] netfilter: nf_tables: use timestamp to check for set element timeout
by Liu Jian 14 Oct '24

14 Oct '24

From: Pablo Neira Ayuso <pablo(a)netfilter.org> mainline inclusion from mainline-v6.8-rc4 commit 7395dfacfff65e9938ac0889dafa1ab01e987d15 category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/I9O0MR CVE: CVE-2024-27397 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… --------------------------- Add a timestamp field at the beginning of the transaction, store it in the nftables per-netns area. Update set backend .insert, .deactivate and sync gc path to use the timestamp, this avoids that an element expires while control plane transaction is still unfinished. .lookup and .update, which are used from packet path, still use the current time to check if the element has expired. And .get path and dump also since this runs lockless under rcu read size lock. Then, there is async gc which also needs to check the current time since it runs asynchronously from a workqueue. Fixes: c3e1b005ed1c ("netfilter: nf_tables: add set element timeout support") Signed-off-by: Pablo Neira Ayuso <pablo(a)netfilter.org> Conflicts: net/netfilter/nft_set_pipapo.c net/netfilter/nft_set_rbtree.c net/netfilter/nf_tables_api.c [Do not backport nft_rbtree_gc() part, it is a async gc in our repository. The sync gc need backport 7d259f021aaa. Another caused by we did not backport commit 9dad402b89e8.] Signed-off-by: Liu Jian <liujian56(a)huawei.com> --- include/net/netfilter/nf_tables.h | 16 ++++++++++++++-- net/netfilter/nf_tables_api.c | 4 +++- net/netfilter/nft_set_hash.c | 8 +++++++- net/netfilter/nft_set_pipapo.c | 18 +++++++++++------- net/netfilter/nft_set_rbtree.c | 6 ++++-- 5 files changed, 39 insertions(+), 13 deletions(-) diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h index 8af2543520b9..45dc6922dec0 100644 --- a/include/net/netfilter/nf_tables.h +++ b/include/net/netfilter/nf_tables.h @@ -794,10 +794,16 @@ static inline struct nft_set_elem_expr *nft_set_ext_expr(const struct nft_set_ex return nft_set_ext(ext, NFT_SET_EXT_EXPRESSIONS); } -static inline bool nft_set_elem_expired(const struct nft_set_ext *ext) +static inline bool __nft_set_elem_expired(const struct nft_set_ext *ext, + u64 tstamp) { return nft_set_ext_exists(ext, NFT_SET_EXT_EXPIRATION) && - time_is_before_eq_jiffies64(*nft_set_ext_expiration(ext)); + time_after_eq64(tstamp, *nft_set_ext_expiration(ext)); +} + +static inline bool nft_set_elem_expired(const struct nft_set_ext *ext) +{ + return __nft_set_elem_expired(ext, get_jiffies_64()); } static inline struct nft_set_ext *nft_set_elem_ext(const struct nft_set *set, @@ -1740,6 +1746,7 @@ struct nftables_pernet { struct list_head notify_list; struct mutex commit_mutex; u64 table_handle; + u64 tstamp; unsigned int base_seq; unsigned int gc_seq; u8 validate_state; @@ -1752,6 +1759,11 @@ static inline struct nftables_pernet *nft_pernet(const struct net *net) return net_generic(net, nf_tables_net_id); } +static inline u64 nft_net_tstamp(const struct net *net) +{ + return nft_pernet(net)->tstamp; +} + #define __NFT_REDUCE_READONLY 1UL #define NFT_REDUCE_READONLY (void *)__NFT_REDUCE_READONLY diff --git a/net/netfilter/nf_tables_api.c b/net/netfilter/nf_tables_api.c index ea139fca74cb..1899ec55b872 100644 --- a/net/netfilter/nf_tables_api.c +++ b/net/netfilter/nf_tables_api.c @@ -9773,6 +9773,7 @@ struct nft_trans_gc *nft_trans_gc_catchall_async(struct nft_trans_gc *gc, struct nft_trans_gc *nft_trans_gc_catchall_sync(struct nft_trans_gc *gc) { struct nft_set_elem_catchall *catchall, *next; + u64 tstamp = nft_net_tstamp(gc->net); const struct nft_set *set = gc->set; struct nft_set_elem elem; struct nft_set_ext *ext; @@ -9782,7 +9783,7 @@ struct nft_trans_gc *nft_trans_gc_catchall_sync(struct nft_trans_gc *gc) list_for_each_entry_safe(catchall, next, &set->catchall_list, list) { ext = nft_set_elem_ext(set, catchall->elem); - if (!nft_set_elem_expired(ext)) + if (!__nft_set_elem_expired(ext, tstamp)) continue; gc = nft_trans_gc_queue_sync(gc, GFP_KERNEL); @@ -10581,6 +10582,7 @@ static bool nf_tables_valid_genid(struct net *net, u32 genid) bool genid_ok; mutex_lock(&nft_net->commit_mutex); + nft_net->tstamp = get_jiffies_64(); genid_ok = genid == 0 || nft_net->base_seq == genid; if (!genid_ok) diff --git a/net/netfilter/nft_set_hash.c b/net/netfilter/nft_set_hash.c index 2013de934cef..1fd3b413350d 100644 --- a/net/netfilter/nft_set_hash.c +++ b/net/netfilter/nft_set_hash.c @@ -35,6 +35,7 @@ struct nft_rhash_cmp_arg { const struct nft_set *set; const u32 *key; u8 genmask; + u64 tstamp; }; static inline u32 nft_rhash_key(const void *data, u32 len, u32 seed) @@ -61,7 +62,7 @@ static inline int nft_rhash_cmp(struct rhashtable_compare_arg *arg, return 1; if (nft_set_elem_is_dead(&he->ext)) return 1; - if (nft_set_elem_expired(&he->ext)) + if (__nft_set_elem_expired(&he->ext, x->tstamp)) return 1; if (!nft_set_elem_active(&he->ext, x->genmask)) return 1; @@ -86,6 +87,7 @@ bool nft_rhash_lookup(const struct net *net, const struct nft_set *set, .genmask = nft_genmask_cur(net), .set = set, .key = key, + .tstamp = get_jiffies_64(), }; he = rhashtable_lookup(&priv->ht, &arg, nft_rhash_params); @@ -104,6 +106,7 @@ static void *nft_rhash_get(const struct net *net, const struct nft_set *set, .genmask = nft_genmask_cur(net), .set = set, .key = elem->key.val.data, + .tstamp = get_jiffies_64(), }; he = rhashtable_lookup(&priv->ht, &arg, nft_rhash_params); @@ -127,6 +130,7 @@ static bool nft_rhash_update(struct nft_set *set, const u32 *key, .genmask = NFT_GENMASK_ANY, .set = set, .key = key, + .tstamp = get_jiffies_64(), }; he = rhashtable_lookup(&priv->ht, &arg, nft_rhash_params); @@ -170,6 +174,7 @@ static int nft_rhash_insert(const struct net *net, const struct nft_set *set, .genmask = nft_genmask_next(net), .set = set, .key = elem->key.val.data, + .tstamp = nft_net_tstamp(net), }; struct nft_rhash_elem *prev; @@ -212,6 +217,7 @@ static void *nft_rhash_deactivate(const struct net *net, .genmask = nft_genmask_next(net), .set = set, .key = elem->key.val.data, + .tstamp = nft_net_tstamp(net), }; rcu_read_lock(); diff --git a/net/netfilter/nft_set_pipapo.c b/net/netfilter/nft_set_pipapo.c index e4dd73093048..b72d9f7ff78f 100644 --- a/net/netfilter/nft_set_pipapo.c +++ b/net/netfilter/nft_set_pipapo.c @@ -504,6 +504,7 @@ bool nft_pipapo_lookup(const struct net *net, const struct nft_set *set, * @set: nftables API set representation * @data: Key data to be matched against existing elements * @genmask: If set, check that element is active in given genmask + * @tstamp: timestamp to check for expired elements * * This is essentially the same as the lookup function, except that it matches * key data against the uncommitted copy and doesn't use preallocated maps for @@ -513,7 +514,8 @@ bool nft_pipapo_lookup(const struct net *net, const struct nft_set *set, */ static struct nft_pipapo_elem *pipapo_get(const struct net *net, const struct nft_set *set, - const u8 *data, u8 genmask) + const u8 *data, u8 genmask, + u64 tstamp) { struct nft_pipapo_elem *ret = ERR_PTR(-ENOENT); struct nft_pipapo *priv = nft_set_priv(set); @@ -568,7 +570,7 @@ static struct nft_pipapo_elem *pipapo_get(const struct net *net, goto out; if (last) { - if (nft_set_elem_expired(&f->mt[b].e->ext)) + if (__nft_set_elem_expired(&f->mt[b].e->ext, tstamp)) goto next_match; if ((genmask && !nft_set_elem_active(&f->mt[b].e->ext, genmask))) @@ -605,7 +607,7 @@ static void *nft_pipapo_get(const struct net *net, const struct nft_set *set, const struct nft_set_elem *elem, unsigned int flags) { return pipapo_get(net, set, (const u8 *)elem->key.val.data, - nft_genmask_cur(net)); + nft_genmask_cur(net), get_jiffies_64()); } /** @@ -1199,6 +1201,7 @@ static int nft_pipapo_insert(const struct net *net, const struct nft_set *set, struct nft_pipapo *priv = nft_set_priv(set); struct nft_pipapo_match *m = priv->clone; u8 genmask = nft_genmask_next(net); + u64 tstamp = nft_net_tstamp(net); struct nft_pipapo_field *f; const u8 *start_p, *end_p; int i, bsize_max, err = 0; @@ -1208,7 +1211,7 @@ static int nft_pipapo_insert(const struct net *net, const struct nft_set *set, else end = start; - dup = pipapo_get(net, set, start, genmask); + dup = pipapo_get(net, set, start, genmask, tstamp); if (!IS_ERR(dup)) { /* Check if we already have the same exact entry */ const struct nft_data *dup_key, *dup_end; @@ -1230,7 +1233,7 @@ static int nft_pipapo_insert(const struct net *net, const struct nft_set *set, if (PTR_ERR(dup) == -ENOENT) { /* Look for partially overlapping entries */ - dup = pipapo_get(net, set, end, nft_genmask_next(net)); + dup = pipapo_get(net, set, end, nft_genmask_next(net), tstamp); } if (PTR_ERR(dup) != -ENOENT) { @@ -1582,6 +1585,7 @@ static void pipapo_gc(const struct nft_set *_set, struct nft_pipapo_match *m) struct nft_set *set = (struct nft_set *) _set; struct nft_pipapo *priv = nft_set_priv(set); struct net *net = read_pnet(&set->net); + u64 tstamp = nft_net_tstamp(net); int rules_f0, first_rule = 0; struct nft_pipapo_elem *e; struct nft_trans_gc *gc; @@ -1616,7 +1620,7 @@ static void pipapo_gc(const struct nft_set *_set, struct nft_pipapo_match *m) /* synchronous gc never fails, there is no need to set on * NFT_SET_ELEM_DEAD_BIT. */ - if (nft_set_elem_expired(&e->ext)) { + if (__nft_set_elem_expired(&e->ext, tstamp)) { priv->dirty = true; gc = nft_trans_gc_queue_sync(gc, GFP_ATOMIC); @@ -1787,7 +1791,7 @@ static void *pipapo_deactivate(const struct net *net, const struct nft_set *set, { struct nft_pipapo_elem *e; - e = pipapo_get(net, set, data, nft_genmask_next(net)); + e = pipapo_get(net, set, data, nft_genmask_next(net), nft_net_tstamp(net)); if (IS_ERR(e)) return NULL; diff --git a/net/netfilter/nft_set_rbtree.c b/net/netfilter/nft_set_rbtree.c index 5bf5572e945c..021d9e76129a 100644 --- a/net/netfilter/nft_set_rbtree.c +++ b/net/netfilter/nft_set_rbtree.c @@ -314,6 +314,7 @@ static int __nft_rbtree_insert(const struct net *net, const struct nft_set *set, struct nft_rbtree *priv = nft_set_priv(set); u8 cur_genmask = nft_genmask_cur(net); u8 genmask = nft_genmask_next(net); + u64 tstamp = nft_net_tstamp(net); int d; /* Descend the tree to search for an existing element greater than the @@ -361,7 +362,7 @@ static int __nft_rbtree_insert(const struct net *net, const struct nft_set *set, /* perform garbage collection to avoid bogus overlap reports * but skip new elements in this transaction. */ - if (nft_set_elem_expired(&rbe->ext) && + if (__nft_set_elem_expired(&rbe->ext, tstamp) && nft_set_elem_active(&rbe->ext, cur_genmask)) { const struct nft_rbtree_elem *removed_end; @@ -548,6 +549,7 @@ static void *nft_rbtree_deactivate(const struct net *net, const struct rb_node *parent = priv->root.rb_node; struct nft_rbtree_elem *rbe, *this = elem->priv; u8 genmask = nft_genmask_next(net); + u64 tstamp = nft_net_tstamp(net); int d; while (parent != NULL) { @@ -568,7 +570,7 @@ static void *nft_rbtree_deactivate(const struct net *net, nft_rbtree_interval_end(this)) { parent = parent->rb_right; continue; - } else if (nft_set_elem_expired(&rbe->ext)) { + } else if (__nft_set_elem_expired(&rbe->ext, tstamp)) { break; } else if (!nft_set_elem_active(&rbe->ext, genmask)) { parent = parent->rb_left; -- 2.34.1

2 1

2025

2024

2023

2022

2021

2020

2019

Kernel