April 2025 - Linux-stable-mirror

[PATCH 6.1.y 1/4] net: dsa: mv88e6xxx: fix atu_move_port_mask for 6341 family

by Marek Behún

commit 4ae01ec007716986e1a20f1285eb013cbf188830 upstream. The atu_move_port_mask for 6341 family (Topaz) is 0xf, not 0x1f. The PortVec field is 8 bits wide, not 11 as in 6390 family. Fix this. Fixes: e606ca36bbf2 ("net: dsa: mv88e6xxx: rework ATU Remove") Signed-off-by: Marek Behún <kabel(a)kernel.org> --- drivers/net/dsa/mv88e6xxx/chip.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/dsa/mv88e6xxx/chip.c b/drivers/net/dsa/mv88e6xxx/chip.c index 0541c9a7fc49..b5e06f66cd38 100644 --- a/drivers/net/dsa/mv88e6xxx/chip.c +++ b/drivers/net/dsa/mv88e6xxx/chip.c @@ -5813,7 +5813,7 @@ static const struct mv88e6xxx_info mv88e6xxx_table[] = { .global1_addr = 0x1b, .global2_addr = 0x1c, .age_time_coeff = 3750, - .atu_move_port_mask = 0x1f, + .atu_move_port_mask = 0xf, .g1_irqs = 9, .g2_irqs = 10, .pvt = true, @@ -6272,7 +6272,7 @@ static const struct mv88e6xxx_info mv88e6xxx_table[] = { .global1_addr = 0x1b, .global2_addr = 0x1c, .age_time_coeff = 3750, - .atu_move_port_mask = 0x1f, + .atu_move_port_mask = 0xf, .g1_irqs = 9, .g2_irqs = 10, .pvt = true, -- 2.49.0

2 months

2
7
0 0

[PATCHSET 6.12] xfs: backported fixes from upstream

by Darrick J. Wong

Hi all, Here's a bunch of hand-ported bug fixes for 6.12 LTS. If you're going to start using this code, I strongly recommend pulling from my git trees, which are linked below. With a bit of luck, this should all go splendidly. Comments and questions are, as always, welcome. --D kernel git tree: https://git.kernel.org/cgit/linux/kernel/git/djwong/xfs-linux.git/log/?h=ne… --- Commits in this patchset: * xfs: do not check NEEDSREPAIR if ro,norecovery mount. * xfs: Do not allow norecovery mount with quotacheck * xfs: rename xfs_iomap_swapfile_activate to xfs_vm_swap_activate * xfs: flush inodegc before swapon --- fs/xfs/xfs_aops.c | 41 +++++++++++++++++++++++++++++++++++++---- fs/xfs/xfs_qm_bhv.c | 49 ++++++++++++++++++++++++++++++++++--------------- fs/xfs/xfs_super.c | 8 ++++++-- 3 files changed, 77 insertions(+), 21 deletions(-)

2 months

2
8
0 0

[PATCH 6.1.y v2] net/sched: act_mirred: don't override retval if we already lost the skb

by jianqi.ren.cn＠windriver.com

From: Jakub Kicinski <kuba(a)kernel.org> [ Upstream commit 166c2c8a6a4dc2e4ceba9e10cfe81c3e469e3210 ] If we're redirecting the skb, and haven't called tcf_mirred_forward(), yet, we need to tell the core to drop the skb by setting the retcode to SHOT. If we have called tcf_mirred_forward(), however, the skb is out of our hands and returning SHOT will lead to UaF. Move the retval override to the error path which actually need it. Reviewed-by: Michal Swiatkowski <michal.swiatkowski(a)linux.intel.com> Fixes: e5cf1baf92cb ("act_mirred: use TC_ACT_REINSERT when possible") Signed-off-by: Jakub Kicinski <kuba(a)kernel.org> Acked-by: Jamal Hadi Salim <jhs(a)mojatatu.com> Signed-off-by: David S. Miller <davem(a)davemloft.net> [Minor conflict resolved due to code context change.] Signed-off-by: Jianqi Ren <jianqi.ren.cn(a)windriver.com> Signed-off-by: He Zhe <zhe.he(a)windriver.com> --- v2: Fix the following issue net/sched/act_mirred.c:265:6: error: variable 'is_redirect' is used uninitialized whenever 'if' condition is true found by the following tuxmake(https://lore.kernel.org/stable/CA+G9fYu+FEZ-3ye30Hk2sk1+LFsw7iO5AHu… tuxmake --runtime podman --target-arch arm --toolchain clang-20 --kconfig allmodconfig LLVM=1 LLVM_IAS=1 Verified the build test Signed-off-by: Jianqi Ren <jianqi.ren.cn(a)windriver.com> --- net/sched/act_mirred.c | 22 +++++++++++++--------- 1 file changed, 13 insertions(+), 9 deletions(-) diff --git a/net/sched/act_mirred.c b/net/sched/act_mirred.c index 36395e5db3b4..bbc34987bd09 100644 --- a/net/sched/act_mirred.c +++ b/net/sched/act_mirred.c @@ -255,31 +255,31 @@ static int tcf_mirred_act(struct sk_buff *skb, const struct tc_action *a, m_mac_header_xmit = READ_ONCE(m->tcfm_mac_header_xmit); m_eaction = READ_ONCE(m->tcfm_eaction); + is_redirect = tcf_mirred_is_act_redirect(m_eaction); retval = READ_ONCE(m->tcf_action); dev = rcu_dereference_bh(m->tcfm_dev); if (unlikely(!dev)) { pr_notice_once("tc mirred: target device is gone\n"); - goto out; + goto err_cant_do; } if (unlikely(!(dev->flags & IFF_UP)) || !netif_carrier_ok(dev)) { net_notice_ratelimited("tc mirred to Houston: device %s is down\n", dev->name); - goto out; + goto err_cant_do; } /* we could easily avoid the clone only if called by ingress and clsact; * since we can't easily detect the clsact caller, skip clone only for * ingress - that covers the TC S/W datapath. */ - is_redirect = tcf_mirred_is_act_redirect(m_eaction); at_ingress = skb_at_tc_ingress(skb); use_reinsert = at_ingress && is_redirect && tcf_mirred_can_reinsert(retval); if (!use_reinsert) { skb2 = skb_clone(skb, GFP_ATOMIC); if (!skb2) - goto out; + goto err_cant_do; } want_ingress = tcf_mirred_act_wants_ingress(m_eaction); @@ -321,12 +321,16 @@ static int tcf_mirred_act(struct sk_buff *skb, const struct tc_action *a, } err = tcf_mirred_forward(want_ingress, skb2); - if (err) { -out: + if (err) tcf_action_inc_overlimit_qstats(&m->common); - if (tcf_mirred_is_act_redirect(m_eaction)) - retval = TC_ACT_SHOT; - } + __this_cpu_dec(mirred_nest_level); + + return retval; + +err_cant_do: + if (is_redirect) + retval = TC_ACT_SHOT; + tcf_action_inc_overlimit_qstats(&m->common); __this_cpu_dec(mirred_nest_level); return retval; -- 2.34.1

2 months

3
3
0 0

[PATCH 5.15.y] jfs: define xtree root and page independently

by Aditya Dutt

From: Dave Kleikamp <dave.kleikamp(a)oracle.com> [ Upstream commit a779ed754e52d582b8c0e17959df063108bd0656 ] In order to make array bounds checking sane, provide a separate definition of the in-inode xtree root and the external xtree page. Signed-off-by: Dave Kleikamp <dave.kleikamp(a)oracle.com> Tested-by: Manas Ghandat <ghandatmanas(a)gmail.com> (cherry picked from commit a779ed754e52d582b8c0e17959df063108bd0656) Closes: https://syzkaller.appspot.com/bug?extid=ccb458b6679845ee0bae Reported-by: syzbot+ccb458b6679845ee0bae(a)syzkaller.appspotmail.com Signed-off-by: Aditya Dutt <duttaditya18(a)gmail.com> --- I am sending this as per the suggestion by Greg to submit backports for all the relevant stable trees: https://lore.kernel.org/stable/2025042210-stylized-nearest-ea59@gregkh/ This patch has been applied in >= 6.12 and has been backported to 6.6: 2ff51719ec615e1b373c1811443efe93594c41a9 I have already sent a mail for 6.1: https://lore.kernel.org/stable/20250427153045.90396-1-duttaditya18@gmail.co… syzbot checked the patch against 5.15.y and confirmed that the reproducer did not trigger any issues. check here: https://lore.kernel.org/lkml/67fea0bf.050a0220.186b78.0006.GAE@google.com/ fs/jfs/jfs_dinode.h | 2 +- fs/jfs/jfs_imap.c | 6 +++--- fs/jfs/jfs_incore.h | 2 +- fs/jfs/jfs_txnmgr.c | 4 ++-- fs/jfs/jfs_xtree.c | 4 ++-- fs/jfs/jfs_xtree.h | 37 +++++++++++++++++++++++-------------- 6 files changed, 32 insertions(+), 23 deletions(-) diff --git a/fs/jfs/jfs_dinode.h b/fs/jfs/jfs_dinode.h index 6b231d0d0071..603aae17a693 100644 --- a/fs/jfs/jfs_dinode.h +++ b/fs/jfs/jfs_dinode.h @@ -96,7 +96,7 @@ struct dinode { #define di_gengen u._file._u1._imap._gengen union { - xtpage_t _xtroot; + xtroot_t _xtroot; struct { u8 unused[16]; /* 16: */ dxd_t _dxd; /* 16: */ diff --git a/fs/jfs/jfs_imap.c b/fs/jfs/jfs_imap.c index c72e97f06579..0e7d2662f202 100644 --- a/fs/jfs/jfs_imap.c +++ b/fs/jfs/jfs_imap.c @@ -673,7 +673,7 @@ int diWrite(tid_t tid, struct inode *ip) * This is the special xtree inside the directory for storing * the directory table */ - xtpage_t *p, *xp; + xtroot_t *p, *xp; xad_t *xad; jfs_ip->xtlid = 0; @@ -687,7 +687,7 @@ int diWrite(tid_t tid, struct inode *ip) * copy xtree root from inode to dinode: */ p = &jfs_ip->i_xtroot; - xp = (xtpage_t *) &dp->di_dirtable; + xp = (xtroot_t *) &dp->di_dirtable; lv = ilinelock->lv; for (n = 0; n < ilinelock->index; n++, lv++) { memcpy(&xp->xad[lv->offset], &p->xad[lv->offset], @@ -716,7 +716,7 @@ int diWrite(tid_t tid, struct inode *ip) * regular file: 16 byte (XAD slot) granularity */ if (type & tlckXTREE) { - xtpage_t *p, *xp; + xtroot_t *p, *xp; xad_t *xad; /* diff --git a/fs/jfs/jfs_incore.h b/fs/jfs/jfs_incore.h index 721def69e732..dd4264aa9bed 100644 --- a/fs/jfs/jfs_incore.h +++ b/fs/jfs/jfs_incore.h @@ -66,7 +66,7 @@ struct jfs_inode_info { lid_t xtlid; /* lid of xtree lock on directory */ union { struct { - xtpage_t _xtroot; /* 288: xtree root */ + xtroot_t _xtroot; /* 288: xtree root */ struct inomap *_imap; /* 4: inode map header */ } file; struct { diff --git a/fs/jfs/jfs_txnmgr.c b/fs/jfs/jfs_txnmgr.c index 6c8680d3907a..3a547e0b934f 100644 --- a/fs/jfs/jfs_txnmgr.c +++ b/fs/jfs/jfs_txnmgr.c @@ -783,7 +783,7 @@ struct tlock *txLock(tid_t tid, struct inode *ip, struct metapage * mp, if (mp->xflag & COMMIT_PAGE) p = (xtpage_t *) mp->data; else - p = &jfs_ip->i_xtroot; + p = (xtpage_t *) &jfs_ip->i_xtroot; xtlck->lwm.offset = le16_to_cpu(p->header.nextindex); } @@ -1710,7 +1710,7 @@ static void xtLog(struct jfs_log * log, struct tblock * tblk, struct lrd * lrd, if (tlck->type & tlckBTROOT) { lrd->log.redopage.type |= cpu_to_le16(LOG_BTROOT); - p = &JFS_IP(ip)->i_xtroot; + p = (xtpage_t *) &JFS_IP(ip)->i_xtroot; if (S_ISDIR(ip->i_mode)) lrd->log.redopage.type |= cpu_to_le16(LOG_DIR_XTREE); diff --git a/fs/jfs/jfs_xtree.c b/fs/jfs/jfs_xtree.c index 3148e9b35f3b..34db519933b4 100644 --- a/fs/jfs/jfs_xtree.c +++ b/fs/jfs/jfs_xtree.c @@ -1224,7 +1224,7 @@ xtSplitRoot(tid_t tid, struct xtlock *xtlck; int rc; - sp = &JFS_IP(ip)->i_xtroot; + sp = (xtpage_t *) &JFS_IP(ip)->i_xtroot; INCREMENT(xtStat.split); @@ -3059,7 +3059,7 @@ static int xtRelink(tid_t tid, struct inode *ip, xtpage_t * p) */ void xtInitRoot(tid_t tid, struct inode *ip) { - xtpage_t *p; + xtroot_t *p; /* * acquire a transaction lock on the root diff --git a/fs/jfs/jfs_xtree.h b/fs/jfs/jfs_xtree.h index 5f51be8596b3..dc9b5f8d6385 100644 --- a/fs/jfs/jfs_xtree.h +++ b/fs/jfs/jfs_xtree.h @@ -65,24 +65,33 @@ struct xadlist { #define XTPAGEMAXSLOT 256 #define XTENTRYSTART 2 -/* - * xtree page: - */ -typedef union { - struct xtheader { - __le64 next; /* 8: */ - __le64 prev; /* 8: */ +struct xtheader { + __le64 next; /* 8: */ + __le64 prev; /* 8: */ - u8 flag; /* 1: */ - u8 rsrvd1; /* 1: */ - __le16 nextindex; /* 2: next index = number of entries */ - __le16 maxentry; /* 2: max number of entries */ - __le16 rsrvd2; /* 2: */ + u8 flag; /* 1: */ + u8 rsrvd1; /* 1: */ + __le16 nextindex; /* 2: next index = number of entries */ + __le16 maxentry; /* 2: max number of entries */ + __le16 rsrvd2; /* 2: */ - pxd_t self; /* 8: self */ - } header; /* (32) */ + pxd_t self; /* 8: self */ +}; +/* + * xtree root (in inode): + */ +typedef union { + struct xtheader header; xad_t xad[XTROOTMAXSLOT]; /* 16 * maxentry: xad array */ +} xtroot_t; + +/* + * xtree page: + */ +typedef union { + struct xtheader header; + xad_t xad[XTPAGEMAXSLOT]; /* 16 * maxentry: xad array */ } xtpage_t; /* -- 2.34.1

2 months

2
1
0 0

[PATCH 6.1.y v2] net/sched: act_mirred: don't override retval if we already lost the skb

by jianqi.ren.cn＠windriver.com

From: Jakub Kicinski <kuba(a)kernel.org> [ Upstream commit 166c2c8a6a4dc2e4ceba9e10cfe81c3e469e3210 ] If we're redirecting the skb, and haven't called tcf_mirred_forward(), yet, we need to tell the core to drop the skb by setting the retcode to SHOT. If we have called tcf_mirred_forward(), however, the skb is out of our hands and returning SHOT will lead to UaF. Move the retval override to the error path which actually need it. Reviewed-by: Michal Swiatkowski <michal.swiatkowski(a)linux.intel.com> Fixes: e5cf1baf92cb ("act_mirred: use TC_ACT_REINSERT when possible") Signed-off-by: Jakub Kicinski <kuba(a)kernel.org> Acked-by: Jamal Hadi Salim <jhs(a)mojatatu.com> Signed-off-by: David S. Miller <davem(a)davemloft.net> [Minor conflict resolved due to code context change.] Signed-off-by: Jianqi Ren <jianqi.ren.cn(a)windriver.com> Signed-off-by: He Zhe <zhe.he(a)windriver.com> --- v2: Fix the following issue net/sched/act_mirred.c:265:6: error: variable 'is_redirect' is used uninitialized whenever 'if' condition is true found by the following tuxmake (https://lore.kernel.org/stable/CA+G9fYu+FEZ-3ye30Hk2sk1+LFsw7iO5AHueUa9H1Ub…) Verified the build test by cmd(tuxmake --runtime podman --target-arch arm --toolchain clang-20 --kconfig allmodconfig LLVM=1 LLVM_IAS=1) --- net/sched/act_mirred.c | 22 +++++++++++++--------- 1 file changed, 13 insertions(+), 9 deletions(-) diff --git a/net/sched/act_mirred.c b/net/sched/act_mirred.c index 36395e5db3b4..bbc34987bd09 100644 --- a/net/sched/act_mirred.c +++ b/net/sched/act_mirred.c @@ -255,31 +255,31 @@ static int tcf_mirred_act(struct sk_buff *skb, const struct tc_action *a, m_mac_header_xmit = READ_ONCE(m->tcfm_mac_header_xmit); m_eaction = READ_ONCE(m->tcfm_eaction); + is_redirect = tcf_mirred_is_act_redirect(m_eaction); retval = READ_ONCE(m->tcf_action); dev = rcu_dereference_bh(m->tcfm_dev); if (unlikely(!dev)) { pr_notice_once("tc mirred: target device is gone\n"); - goto out; + goto err_cant_do; } if (unlikely(!(dev->flags & IFF_UP)) || !netif_carrier_ok(dev)) { net_notice_ratelimited("tc mirred to Houston: device %s is down\n", dev->name); - goto out; + goto err_cant_do; } /* we could easily avoid the clone only if called by ingress and clsact; * since we can't easily detect the clsact caller, skip clone only for * ingress - that covers the TC S/W datapath. */ - is_redirect = tcf_mirred_is_act_redirect(m_eaction); at_ingress = skb_at_tc_ingress(skb); use_reinsert = at_ingress && is_redirect && tcf_mirred_can_reinsert(retval); if (!use_reinsert) { skb2 = skb_clone(skb, GFP_ATOMIC); if (!skb2) - goto out; + goto err_cant_do; } want_ingress = tcf_mirred_act_wants_ingress(m_eaction); @@ -321,12 +321,16 @@ static int tcf_mirred_act(struct sk_buff *skb, const struct tc_action *a, } err = tcf_mirred_forward(want_ingress, skb2); - if (err) { -out: + if (err) tcf_action_inc_overlimit_qstats(&m->common); - if (tcf_mirred_is_act_redirect(m_eaction)) - retval = TC_ACT_SHOT; - } + __this_cpu_dec(mirred_nest_level); + + return retval; + +err_cant_do: + if (is_redirect) + retval = TC_ACT_SHOT; + tcf_action_inc_overlimit_qstats(&m->common); __this_cpu_dec(mirred_nest_level); return retval; -- 2.34.1

2 months

4
4
0 0

[PATCH 5.10.y v2] net/sched: act_mirred: don't override retval if we already lost the skb

by jianqi.ren.cn＠windriver.com

From: Jakub Kicinski <kuba(a)kernel.org> [ Upstream commit 166c2c8a6a4dc2e4ceba9e10cfe81c3e469e3210 ] If we're redirecting the skb, and haven't called tcf_mirred_forward(), yet, we need to tell the core to drop the skb by setting the retcode to SHOT. If we have called tcf_mirred_forward(), however, the skb is out of our hands and returning SHOT will lead to UaF. Move the retval override to the error path which actually need it. Reviewed-by: Michal Swiatkowski <michal.swiatkowski(a)linux.intel.com> Fixes: e5cf1baf92cb ("act_mirred: use TC_ACT_REINSERT when possible") Signed-off-by: Jakub Kicinski <kuba(a)kernel.org> Acked-by: Jamal Hadi Salim <jhs(a)mojatatu.com> Signed-off-by: David S. Miller <davem(a)davemloft.net> [Minor conflict resolved due to code context change.] Signed-off-by: Jianqi Ren <jianqi.ren.cn(a)windriver.com> Signed-off-by: He Zhe <zhe.he(a)windriver.com> --- v2: Fix the following issue net/sched/act_mirred.c:265:6: error: variable 'is_redirect' is used uninitialized whenever 'if' condition is true found by the following tuxmake (https://lore.kernel.org/stable/CA+G9fYu+FEZ-3ye30Hk2sk1+LFsw7iO5AHueUa9H1Ub…) tuxmake --runtime podman --target-arch arm --toolchain clang-20 --kconfig allmodconfig LLVM=1 LLVM_IAS=1 The above command line couldn't pass for some other issue unrelated to this patch, but we're sure that the above "is_redirect" issue has been fixed. --- net/sched/act_mirred.c | 22 +++++++++++++--------- 1 file changed, 13 insertions(+), 9 deletions(-) diff --git a/net/sched/act_mirred.c b/net/sched/act_mirred.c index 91a19460cb57..07b9f8335f05 100644 --- a/net/sched/act_mirred.c +++ b/net/sched/act_mirred.c @@ -256,31 +256,31 @@ static int tcf_mirred_act(struct sk_buff *skb, const struct tc_action *a, m_mac_header_xmit = READ_ONCE(m->tcfm_mac_header_xmit); m_eaction = READ_ONCE(m->tcfm_eaction); + is_redirect = tcf_mirred_is_act_redirect(m_eaction); retval = READ_ONCE(m->tcf_action); dev = rcu_dereference_bh(m->tcfm_dev); if (unlikely(!dev)) { pr_notice_once("tc mirred: target device is gone\n"); - goto out; + goto err_cant_do; } if (unlikely(!(dev->flags & IFF_UP)) || !netif_carrier_ok(dev)) { net_notice_ratelimited("tc mirred to Houston: device %s is down\n", dev->name); - goto out; + goto err_cant_do; } /* we could easily avoid the clone only if called by ingress and clsact; * since we can't easily detect the clsact caller, skip clone only for * ingress - that covers the TC S/W datapath. */ - is_redirect = tcf_mirred_is_act_redirect(m_eaction); at_ingress = skb_at_tc_ingress(skb); use_reinsert = at_ingress && is_redirect && tcf_mirred_can_reinsert(retval); if (!use_reinsert) { skb2 = skb_clone(skb, GFP_ATOMIC); if (!skb2) - goto out; + goto err_cant_do; } want_ingress = tcf_mirred_act_wants_ingress(m_eaction); @@ -323,12 +323,16 @@ static int tcf_mirred_act(struct sk_buff *skb, const struct tc_action *a, } err = tcf_mirred_forward(want_ingress, skb2); - if (err) { -out: + if (err) tcf_action_inc_overlimit_qstats(&m->common); - if (tcf_mirred_is_act_redirect(m_eaction)) - retval = TC_ACT_SHOT; - } + __this_cpu_dec(mirred_nest_level); + + return retval; + +err_cant_do: + if (is_redirect) + retval = TC_ACT_SHOT; + tcf_action_inc_overlimit_qstats(&m->common); __this_cpu_dec(mirred_nest_level); return retval; -- 2.34.1

2 months

2
1
0 0

[PATCH 6.12.y 1/5] Revert "net: dsa: mv88e6xxx: fix internal PHYs for 6320 family"

by Marek Behún

This reverts commit 2b27df6852444b76724f5d425826a38581d63407. For stable 6.12 it was misapplied to wrong entries of the `mv88e6xxx_table` array: instead of the MV88E6320 and MV88E6321 entries it was applied to the MV88E6240 and MV88E6352 entries. Signed-off-by: Marek Behún <kabel(a)kernel.org> --- drivers/net/dsa/mv88e6xxx/chip.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/drivers/net/dsa/mv88e6xxx/chip.c b/drivers/net/dsa/mv88e6xxx/chip.c index 28967a338aa9..720ff57c854d 100644 --- a/drivers/net/dsa/mv88e6xxx/chip.c +++ b/drivers/net/dsa/mv88e6xxx/chip.c @@ -6182,8 +6182,7 @@ static const struct mv88e6xxx_info mv88e6xxx_table[] = { .num_databases = 4096, .num_macs = 8192, .num_ports = 7, - .num_internal_phys = 2, - .internal_phys_offset = 3, + .num_internal_phys = 5, .num_gpio = 15, .max_vid = 4095, .max_sid = 63, @@ -6377,8 +6376,7 @@ static const struct mv88e6xxx_info mv88e6xxx_table[] = { .num_databases = 4096, .num_macs = 8192, .num_ports = 7, - .num_internal_phys = 2, - .internal_phys_offset = 3, + .num_internal_phys = 5, .num_gpio = 15, .max_vid = 4095, .max_sid = 63, -- 2.49.0

2 months

3
12
0 0

[PATCH 5.15.y v2] net/sched: act_mirred: don't override retval if we already lost the skb

by jianqi.ren.cn＠windriver.com

From: Jakub Kicinski <kuba(a)kernel.org> [ Upstream commit 166c2c8a6a4dc2e4ceba9e10cfe81c3e469e3210 ] If we're redirecting the skb, and haven't called tcf_mirred_forward(), yet, we need to tell the core to drop the skb by setting the retcode to SHOT. If we have called tcf_mirred_forward(), however, the skb is out of our hands and returning SHOT will lead to UaF. Move the retval override to the error path which actually need it. Reviewed-by: Michal Swiatkowski <michal.swiatkowski(a)linux.intel.com> Fixes: e5cf1baf92cb ("act_mirred: use TC_ACT_REINSERT when possible") Signed-off-by: Jakub Kicinski <kuba(a)kernel.org> Acked-by: Jamal Hadi Salim <jhs(a)mojatatu.com> Signed-off-by: David S. Miller <davem(a)davemloft.net> [Minor conflict resolved due to code context change.] Signed-off-by: Jianqi Ren <jianqi.ren.cn(a)windriver.com> Signed-off-by: He Zhe <zhe.he(a)windriver.com> --- v2: Fix the following issue net/sched/act_mirred.c:265:6: error: variable 'is_redirect' is used uninitialized whenever 'if' condition is true found by the following tuxmake (https://lore.kernel.org/stable/CA+G9fYu+FEZ-3ye30Hk2sk1+LFsw7iO5AHueUa9H1Ub…) Verified the build test by cmd(tuxmake --runtime podman --target-arch arm --toolchain clang-20 --kconfig allmodconfig LLVM=1 LLVM_IAS=1) --- net/sched/act_mirred.c | 22 +++++++++++++--------- 1 file changed, 13 insertions(+), 9 deletions(-) diff --git a/net/sched/act_mirred.c b/net/sched/act_mirred.c index 97cd4b2377d6..1aa1d10de30e 100644 --- a/net/sched/act_mirred.c +++ b/net/sched/act_mirred.c @@ -254,31 +254,31 @@ static int tcf_mirred_act(struct sk_buff *skb, const struct tc_action *a, m_mac_header_xmit = READ_ONCE(m->tcfm_mac_header_xmit); m_eaction = READ_ONCE(m->tcfm_eaction); + is_redirect = tcf_mirred_is_act_redirect(m_eaction); retval = READ_ONCE(m->tcf_action); dev = rcu_dereference_bh(m->tcfm_dev); if (unlikely(!dev)) { pr_notice_once("tc mirred: target device is gone\n"); - goto out; + goto err_cant_do; } if (unlikely(!(dev->flags & IFF_UP)) || !netif_carrier_ok(dev)) { net_notice_ratelimited("tc mirred to Houston: device %s is down\n", dev->name); - goto out; + goto err_cant_do; } /* we could easily avoid the clone only if called by ingress and clsact; * since we can't easily detect the clsact caller, skip clone only for * ingress - that covers the TC S/W datapath. */ - is_redirect = tcf_mirred_is_act_redirect(m_eaction); at_ingress = skb_at_tc_ingress(skb); use_reinsert = at_ingress && is_redirect && tcf_mirred_can_reinsert(retval); if (!use_reinsert) { skb2 = skb_clone(skb, GFP_ATOMIC); if (!skb2) - goto out; + goto err_cant_do; } want_ingress = tcf_mirred_act_wants_ingress(m_eaction); @@ -321,12 +321,16 @@ static int tcf_mirred_act(struct sk_buff *skb, const struct tc_action *a, } err = tcf_mirred_forward(want_ingress, skb2); - if (err) { -out: + if (err) tcf_action_inc_overlimit_qstats(&m->common); - if (tcf_mirred_is_act_redirect(m_eaction)) - retval = TC_ACT_SHOT; - } + __this_cpu_dec(mirred_nest_level); + + return retval; + +err_cant_do: + if (is_redirect) + retval = TC_ACT_SHOT; + tcf_action_inc_overlimit_qstats(&m->common); __this_cpu_dec(mirred_nest_level); return retval; -- 2.34.1

2 months

2
1
0 0

platform_shutdown crasher reaches LTS v5.10

by Chuck Lever

Howdy - My queue/5.10 CI test runners have hit this crash on shutdown: [ 42.547410] RIP: 0010:platform_shutdown+0x9/0x20 [ 42.547955] Code: b7 46 08 c3 cc cc cc cc 31 c0 83 bf a8 02 00 00 ff 75 ec c3 cc cc cc cc 66 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 8b 47 68 <48> 8b 40 e8 48 85 c0 74 09 48 83 ef 10 e9 35 2c 80 00 c3 cc cc cc [ 42.550016] RSP: 0018:ffffc2ff80013de0 EFLAGS: 00010246 [ 42.550630] RAX: 0000000000000000 RBX: ffff9e6805d2ec18 RCX: 0000000000000000 [ 42.551444] RDX: 0000000000000001 RSI: ffff9e6805d2ec18 RDI: ffff9e6805d2ec10 [ 42.552263] RBP: ffffffff9be79420 R08: ffff9e6805d2f408 R09: ffffffff9bc5c698 [ 42.553081] R10: 0000000000000000 R11: 0000000000000000 R12: ffff9e6805d2ec10 [ 42.553900] R13: ffff9e6805d2ec90 R14: 00000000fee1dead R15: 0000000000000000 [ 42.554717] FS: 00007f05a6960b80(0000) GS:ffff9e6b6fc00000(0000) knlGS:0000000000000000 [ 42.555633] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 42.556291] CR2: ffffffffffffffe8 CR3: 000000010960a006 CR4: 0000000000370ef0 [ 42.557105] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 [ 42.557919] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 [ 42.558733] Call Trace: [ 42.559044] device_shutdown+0x15b/0x1c0 [ 42.559512] __do_sys_reboot.cold+0x2f/0x5b [ 42.560015] ? vfs_writev+0x9b/0x110 [ 42.560444] ? do_writev+0x57/0xf0 [ 42.560859] do_syscall_64+0x33/0x40 [ 42.561289] entry_SYSCALL_64_after_hwframe+0x67/0xd1 It's likely due to this recently applied commit: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git/… Maybe commit 46e85af0cc53 ("driver core: platform: don't oops in platform_shutdown() on unbound devices") fixes this issue, but I didn't test it. -- Chuck Lever

2 months

2
1
0 0

[lvc-project] [PATCH 5.10/5.15] drm/amd/pm: vega10_hwmgr: fix potential off-by-one overflow in 'performance_levels'

by Igor Artemiev

From: Alexey Kodanev <aleksei.kodanev(a)bell-sw.com> commit 2cc4a5914ce952d6fc83b0f8089a23095ad4f677 upstream. Since 'hardwareActivityPerformanceLevels' is set to the size of the 'performance_levels' array in vega10_hwmgr_backend_init(), using the '<=' assertion to check for the next index value is incorrect. Replace it with '<'. Detected using the static analysis tool - Svace. Fixes: f83a9991648b ("drm/amd/powerplay: add Vega10 powerplay support (v5)") Reviewed-by: Evan Quan <evan.quan(a)amd.com> Signed-off-by: Alexey Kodanev <aleksei.kodanev(a)bell-sw.com> Signed-off-by: Alex Deucher <alexander.deucher(a)amd.com> [Igor: In order to adapt this patch to branch 5.10/5.15 the variable name 'vega10_ps' has been changed to 'vega10_power_state' as it is used in branch 5.10/5.15.] Signed-off-by: Igor Artemiev <Igor.A.Artemiev(a)mcst.ru> --- drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.c index 79a41180adf1..30eab5002077 100644 --- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.c +++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/vega10_hwmgr.c @@ -3171,7 +3171,7 @@ static int vega10_get_pp_table_entry_callback_func(struct pp_hwmgr *hwmgr, return -1); PP_ASSERT_WITH_CODE( - (vega10_power_state->performance_level_count <= + (vega10_power_state->performance_level_count < hwmgr->platform_descriptor. hardwareActivityPerformanceLevels), "Performance levels exceeds Driver limit!", -- 2.39.2

2 months

1
0
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror April 2025