From a786ef152cdcfebc923a67f63c7815806eefcf81 Mon Sep 17 00:00:00 2001 From: Daniel Vacek Date: Mon, 5 Nov 2018 18:10:40 +0100 Subject: [PATCH 01/79] x86/tsc: Make calibration refinement more robust The threshold in tsc_read_refs() is constant which may favor slower CPUs but may not be optimal for simple reading of reference on faster ones. Hence make it proportional to tsc_khz when available to compensate for this. The threshold guards against any disturbance like IRQs, NMIs, SMIs or CPU stealing by host on guest systems so rename it accordingly and fix comments as well. Also on some systems there is noticeable DMI bus contention at some point during boot keeping the readout failing (observed with about one in ~300 boots when testing). In that case retry also the second readout instead of simply bailing out unrefined. Usually the next second the readout returns fast just fine without any issues. Signed-off-by: Daniel Vacek Signed-off-by: Thomas Gleixner Cc: Borislav Petkov Cc: "H. Peter Anvin" Link: https://lkml.kernel.org/r/1541437840-29293-1-git-send-email-neelx@redhat.com --- arch/x86/kernel/tsc.c | 30 ++++++++++++++++-------------- 1 file changed, 16 insertions(+), 14 deletions(-) diff --git a/arch/x86/kernel/tsc.c b/arch/x86/kernel/tsc.c index e9f777bfed40..3fae23834069 100644 --- a/arch/x86/kernel/tsc.c +++ b/arch/x86/kernel/tsc.c @@ -297,15 +297,16 @@ static int __init tsc_setup(char *str) __setup("tsc=", tsc_setup); -#define MAX_RETRIES 5 -#define SMI_TRESHOLD 50000 +#define MAX_RETRIES 5 +#define TSC_DEFAULT_THRESHOLD 0x20000 /* - * Read TSC and the reference counters. Take care of SMI disturbance + * Read TSC and the reference counters. Take care of any disturbances */ static u64 tsc_read_refs(u64 *p, int hpet) { u64 t1, t2; + u64 thresh = tsc_khz ? tsc_khz >> 5 : TSC_DEFAULT_THRESHOLD; int i; for (i = 0; i < MAX_RETRIES; i++) { @@ -315,7 +316,7 @@ static u64 tsc_read_refs(u64 *p, int hpet) else *p = acpi_pm_read_early(); t2 = get_cycles(); - if ((t2 - t1) < SMI_TRESHOLD) + if ((t2 - t1) < thresh) return t2; } return ULLONG_MAX; @@ -703,15 +704,15 @@ static unsigned long pit_hpet_ptimer_calibrate_cpu(void) * zero. In each wait loop iteration we read the TSC and check * the delta to the previous read. We keep track of the min * and max values of that delta. The delta is mostly defined - * by the IO time of the PIT access, so we can detect when a - * SMI/SMM disturbance happened between the two reads. If the + * by the IO time of the PIT access, so we can detect when + * any disturbance happened between the two reads. If the * maximum time is significantly larger than the minimum time, * then we discard the result and have another try. * * 2) Reference counter. If available we use the HPET or the * PMTIMER as a reference to check the sanity of that value. * We use separate TSC readouts and check inside of the - * reference read for a SMI/SMM disturbance. We dicard + * reference read for any possible disturbance. We dicard * disturbed values here as well. We do that around the PIT * calibration delay loop as we have to wait for a certain * amount of time anyway. @@ -744,7 +745,7 @@ static unsigned long pit_hpet_ptimer_calibrate_cpu(void) if (ref1 == ref2) continue; - /* Check, whether the sampling was disturbed by an SMI */ + /* Check, whether the sampling was disturbed */ if (tsc1 == ULLONG_MAX || tsc2 == ULLONG_MAX) continue; @@ -1268,7 +1269,7 @@ static DECLARE_DELAYED_WORK(tsc_irqwork, tsc_refine_calibration_work); */ static void tsc_refine_calibration_work(struct work_struct *work) { - static u64 tsc_start = -1, ref_start; + static u64 tsc_start = ULLONG_MAX, ref_start; static int hpet; u64 tsc_stop, ref_stop, delta; unsigned long freq; @@ -1283,14 +1284,15 @@ static void tsc_refine_calibration_work(struct work_struct *work) * delayed the first time we expire. So set the workqueue * again once we know timers are working. */ - if (tsc_start == -1) { + if (tsc_start == ULLONG_MAX) { +restart: /* * Only set hpet once, to avoid mixing hardware * if the hpet becomes enabled later. */ hpet = is_hpet_enabled(); - schedule_delayed_work(&tsc_irqwork, HZ); tsc_start = tsc_read_refs(&ref_start, hpet); + schedule_delayed_work(&tsc_irqwork, HZ); return; } @@ -1300,9 +1302,9 @@ static void tsc_refine_calibration_work(struct work_struct *work) if (ref_start == ref_stop) goto out; - /* Check, whether the sampling was disturbed by an SMI */ - if (tsc_start == ULLONG_MAX || tsc_stop == ULLONG_MAX) - goto out; + /* Check, whether the sampling was disturbed */ + if (tsc_stop == ULLONG_MAX) + goto restart; delta = tsc_stop - tsc_start; delta *= 1000000LL; From d999c0ec2498e54b9328db6b2c1037710025add1 Mon Sep 17 00:00:00 2001 From: Roland Dreier Date: Fri, 30 Nov 2018 13:14:50 -0800 Subject: [PATCH 02/79] x86/hpet: Remove unused FSEC_PER_NSEC define The FSEC_PER_NSEC macro has had zero users since commit ab0e08f15d23 ("x86: hpet: Cleanup the clockevents init and register code"). Remove it. Signed-off-by: Roland Dreier Signed-off-by: Borislav Petkov Acked-by: Thomas Gleixner Cc: "H. Peter Anvin" Cc: Ingo Molnar Cc: x86-ml Link: https://lkml.kernel.org/r/20181130211450.5200-1-roland@purestorage.com --- arch/x86/kernel/hpet.c | 4 ---- 1 file changed, 4 deletions(-) diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c index b0acb22e5a46..dfd3aca82c61 100644 --- a/arch/x86/kernel/hpet.c +++ b/arch/x86/kernel/hpet.c @@ -21,10 +21,6 @@ #define HPET_MASK CLOCKSOURCE_MASK(32) -/* FSEC = 10^-15 - NSEC = 10^-9 */ -#define FSEC_PER_NSEC 1000000L - #define HPET_DEV_USED_BIT 2 #define HPET_DEV_USED (1 << HPET_DEV_USED_BIT) #define HPET_DEV_VALID 0x8 From 0977b2383de69dc48e9fa61c5c77878ed08d87fe Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Fri, 4 Jan 2019 14:16:59 +0100 Subject: [PATCH 03/79] selftests: xfrm: add block rules with adjacent/overlapping subnets The existing script lacks a policy pattern that triggers 'tree node merges' in the kernel. Consider adding policy affecting following subnet: pol1: dst 10.0.0.0/22 pol2: dst 10.0.0.0/23 # adds to existing 10.0.0.0/22 node -> no problems here. But now, lets consider reverse order: pol1: dst 10.0.0.0/24 pol2: dst 10.0.0.0/23 # CANNOT add to existing node When second policy gets added, the kernel must check that the new node ("10.0.0.0/23") doesn't overlap with any existing subnet. Example: dst 10.0.0.0/24 dst 10.0.0.1/24 dst 10.0.0.0/23 When the third policy gets added, the kernel must replace the nodes for the 10.0.0.0/24 and 10.0.0.1/24 policies with a single one and must merge all the subtrees/lists stored in those nodes into the new node. The existing test cases only have overlaps with a single node, so no merging takes place (we can always remove the 'old' node and replace it with the new subnet prefix). Add a few 'block policies' in a pattern that triggers this, with a priority that will make kernel prefer the 'esp' rules. Make sure the 'tunnel ping' tests still pass after they have been added. Signed-off-by: Florian Westphal Signed-off-by: Steffen Klassert --- tools/testing/selftests/net/xfrm_policy.sh | 109 +++++++++++++++++---- 1 file changed, 91 insertions(+), 18 deletions(-) diff --git a/tools/testing/selftests/net/xfrm_policy.sh b/tools/testing/selftests/net/xfrm_policy.sh index 8db35b99457c..b5a1b565a7e6 100755 --- a/tools/testing/selftests/net/xfrm_policy.sh +++ b/tools/testing/selftests/net/xfrm_policy.sh @@ -46,6 +46,58 @@ do_esp() { ip -net $ns xfrm policy add src $rnet dst $lnet dir fwd tmpl src $remote dst $me proto esp mode tunnel priority 100 action allow } +# add policies with different netmasks, to make sure kernel carries +# the policies contained within new netmask over when search tree is +# re-built. +# peer netns that are supposed to be encapsulated via esp have addresses +# in the 10.0.1.0/24 and 10.0.2.0/24 subnets, respectively. +# +# Adding a policy for '10.0.1.0/23' will make it necessary to +# alter the prefix of 10.0.1.0 subnet. +# In case new prefix overlaps with existing node, the node and all +# policies it carries need to be merged with the existing one(s). +# +# Do that here. +do_overlap() +{ + local ns=$1 + + # adds new nodes to tree (neither network exists yet in policy database). + ip -net $ns xfrm policy add src 10.1.0.0/24 dst 10.0.0.0/24 dir fwd priority 200 action block + + # adds a new node in the 10.0.0.0/24 tree (dst node exists). + ip -net $ns xfrm policy add src 10.2.0.0/24 dst 10.0.0.0/24 dir fwd priority 200 action block + + # adds a 10.2.0.0/24 node, but for different dst. + ip -net $ns xfrm policy add src 10.2.0.0/24 dst 10.0.1.0/24 dir fwd priority 200 action block + + # dst now overlaps with the 10.0.1.0/24 ESP policy in fwd. + # kernel must 'promote' existing one (10.0.0.0/24) to 10.0.0.0/23. + # But 10.0.0.0/23 also includes existing 10.0.1.0/24, so that node + # also has to be merged too, including source-sorted subtrees. + # old: + # 10.0.0.0/24 (node 1 in dst tree of the bin) + # 10.1.0.0/24 (node in src tree of dst node 1) + # 10.2.0.0/24 (node in src tree of dst node 1) + # 10.0.1.0/24 (node 2 in dst tree of the bin) + # 10.0.2.0/24 (node in src tree of dst node 2) + # 10.2.0.0/24 (node in src tree of dst node 2) + # + # The next 'policy add' adds dst '10.0.0.0/23', which means + # that dst node 1 and dst node 2 have to be merged including + # the sub-tree. As no duplicates are allowed, policies in + # the two '10.0.2.0/24' are also merged. + # + # after the 'add', internal search tree should look like this: + # 10.0.0.0/23 (node in dst tree of bin) + # 10.0.2.0/24 (node in src tree of dst node) + # 10.1.0.0/24 (node in src tree of dst node) + # 10.2.0.0/24 (node in src tree of dst node) + # + # 10.0.0.0/24 and 10.0.1.0/24 nodes have been merged as 10.0.0.0/23. + ip -net $ns xfrm policy add src 10.1.0.0/24 dst 10.0.0.0/23 dir fwd priority 200 action block +} + do_esp_policy_get_check() { local ns=$1 local lnet=$2 @@ -160,6 +212,41 @@ check_xfrm() { return $lret } +check_exceptions() +{ + logpostfix="$1" + local lret=0 + + # ping to .254 should be excluded from the tunnel (exception is in place). + check_xfrm 0 254 + if [ $? -ne 0 ]; then + echo "FAIL: expected ping to .254 to fail ($logpostfix)" + lret=1 + else + echo "PASS: ping to .254 bypassed ipsec tunnel ($logpostfix)" + fi + + # ping to .253 should use use ipsec due to direct policy exception. + check_xfrm 1 253 + if [ $? -ne 0 ]; then + echo "FAIL: expected ping to .253 to use ipsec tunnel ($logpostfix)" + lret=1 + else + echo "PASS: direct policy matches ($logpostfix)" + fi + + # ping to .2 should use ipsec. + check_xfrm 1 2 + if [ $? -ne 0 ]; then + echo "FAIL: expected ping to .2 to use ipsec tunnel ($logpostfix)" + lret=1 + else + echo "PASS: policy matches ($logpostfix)" + fi + + return $lret +} + #check for needed privileges if [ "$(id -u)" -ne 0 ];then echo "SKIP: Need root privileges" @@ -270,31 +357,17 @@ do_exception ns4 10.0.3.10 10.0.3.1 10.0.1.253 10.0.1.240/28 do_exception ns3 dead:3::1 dead:3::10 dead:2::fd dead:2:f0::/96 do_exception ns4 dead:3::10 dead:3::1 dead:1::fd dead:1:f0::/96 -# ping to .254 should now be excluded from the tunnel -check_xfrm 0 254 +check_exceptions "exceptions" if [ $? -ne 0 ]; then - echo "FAIL: expected ping to .254 to fail" ret=1 -else - echo "PASS: ping to .254 bypassed ipsec tunnel" fi -# ping to .253 should use use ipsec due to direct policy exception. -check_xfrm 1 253 -if [ $? -ne 0 ]; then - echo "FAIL: expected ping to .253 to use ipsec tunnel" - ret=1 -else - echo "PASS: direct policy matches" -fi +# insert block policies with adjacent/overlapping netmasks +do_overlap ns3 -# ping to .2 should use ipsec. -check_xfrm 1 2 +check_exceptions "exceptions and block policies" if [ $? -ne 0 ]; then - echo "FAIL: expected ping to .2 to use ipsec tunnel" ret=1 -else - echo "PASS: policy matches" fi for i in 1 2 3 4;do ip netns del ns$i;done From 355b00d1e14051c13aea48c1c5430c486fed2d7a Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Fri, 4 Jan 2019 14:17:00 +0100 Subject: [PATCH 04/79] xfrm: policy: use hlist rcu variants on inexact insert, part 2 This function was modeled on the 'exact' insert one, which did not use the rcu variant either. When I fixed the 'exact' insert I forgot to propagate this to my development tree, so the inexact variant retained the bug. Fixes: 9cf545ebd591d ("xfrm: policy: store inexact policies in a tree ordered by destination address") Signed-off-by: Florian Westphal Signed-off-by: Steffen Klassert --- net/xfrm/xfrm_policy.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c index 934492bad8e0..628b389af2ba 100644 --- a/net/xfrm/xfrm_policy.c +++ b/net/xfrm/xfrm_policy.c @@ -856,9 +856,9 @@ static void xfrm_policy_inexact_list_reinsert(struct net *net, } if (newpos) - hlist_add_behind(&policy->bydst, newpos); + hlist_add_behind_rcu(&policy->bydst, newpos); else - hlist_add_head(&policy->bydst, &n->hhead); + hlist_add_head_rcu(&policy->bydst, &n->hhead); /* paranoia checks follow. * Check that the reinserted policy matches at least From 7a474c36586f4277f930ab7e6865c97e44dfc3bc Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Fri, 4 Jan 2019 14:17:01 +0100 Subject: [PATCH 05/79] xfrm: policy: increment xfrm_hash_generation on hash rebuild Hash rebuild will re-set all the inexact entries, then re-insert them. Lookups that can occur in parallel will therefore not find any policies. This was safe when lookups were still guarded by rwlock. After rcu-ification, lookups check the hash_generation seqcount to detect when a hash resize takes place. Hash rebuild missed the needed increment. Hash resizes and hash rebuilds cannot occur in parallel (both acquire hash_resize_mutex), so just increment xfrm_hash_generation, like resize. Fixes: a7c44247f704e3 ("xfrm: policy: make xfrm_policy_lookup_bytype lockless") Signed-off-by: Florian Westphal Signed-off-by: Steffen Klassert --- net/xfrm/xfrm_policy.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c index 628b389af2ba..d8fba27a4bfb 100644 --- a/net/xfrm/xfrm_policy.c +++ b/net/xfrm/xfrm_policy.c @@ -1235,6 +1235,7 @@ static void xfrm_hash_rebuild(struct work_struct *work) } while (read_seqretry(&net->xfrm.policy_hthresh.lock, seq)); spin_lock_bh(&net->xfrm.xfrm_policy_lock); + write_seqcount_begin(&xfrm_policy_hash_generation); /* make sure that we can insert the indirect policies again before * we start with destructive action. @@ -1334,6 +1335,7 @@ static void xfrm_hash_rebuild(struct work_struct *work) out_unlock: __xfrm_policy_inexact_flush(net); + write_seqcount_end(&xfrm_policy_hash_generation); spin_unlock_bh(&net->xfrm.xfrm_policy_lock); mutex_unlock(&hash_resize_mutex); From 1548bc4e0512700cf757192c106b3a20ab639223 Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Fri, 4 Jan 2019 14:17:02 +0100 Subject: [PATCH 06/79] xfrm: policy: delete inexact policies from inexact list on hash rebuild An xfrm hash rebuild has to reset the inexact policy list before the policies get re-inserted: A change of hash thresholds will result in policies to get moved from inexact tree to the policy hash table. If the thresholds are increased again later, they get moved from hash table to inexact tree. We must unlink all policies from the inexact tree before re-insertion. Otherwise 'migrate' may find policies that are in main hash table a second time, when it searches the inexact lists. Furthermore, re-insertion without deletion can cause elements ->next to point back to itself, causing soft lockups or double-frees. Reported-by: syzbot+9d971dd21eb26567036b@syzkaller.appspotmail.com Fixes: 9cf545ebd591da ("xfrm: policy: store inexact policies in a tree ordered by destination address") Signed-off-by: Florian Westphal Signed-off-by: Steffen Klassert --- net/xfrm/xfrm_policy.c | 23 ++++++++++------------- 1 file changed, 10 insertions(+), 13 deletions(-) diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c index d8fba27a4bfb..24dfd1e47cf0 100644 --- a/net/xfrm/xfrm_policy.c +++ b/net/xfrm/xfrm_policy.c @@ -680,16 +680,6 @@ static void xfrm_hash_resize(struct work_struct *work) mutex_unlock(&hash_resize_mutex); } -static void xfrm_hash_reset_inexact_table(struct net *net) -{ - struct xfrm_pol_inexact_bin *b; - - lockdep_assert_held(&net->xfrm.xfrm_policy_lock); - - list_for_each_entry(b, &net->xfrm.inexact_bins, inexact_bins) - INIT_HLIST_HEAD(&b->hhead); -} - /* Make sure *pol can be inserted into fastbin. * Useful to check that later insert requests will be sucessful * (provided xfrm_policy_lock is held throughout). @@ -1279,10 +1269,14 @@ static void xfrm_hash_rebuild(struct work_struct *work) } /* reset the bydst and inexact table in all directions */ - xfrm_hash_reset_inexact_table(net); - for (dir = 0; dir < XFRM_POLICY_MAX; dir++) { - INIT_HLIST_HEAD(&net->xfrm.policy_inexact[dir]); + struct hlist_node *n; + + hlist_for_each_entry_safe(policy, n, + &net->xfrm.policy_inexact[dir], + bydst_inexact_list) + hlist_del_init(&policy->bydst_inexact_list); + hmask = net->xfrm.policy_bydst[dir].hmask; odst = net->xfrm.policy_bydst[dir].table; for (i = hmask; i >= 0; i--) @@ -1314,6 +1308,9 @@ static void xfrm_hash_rebuild(struct work_struct *work) newpos = NULL; chain = policy_hash_bysel(net, &policy->selector, policy->family, dir); + + hlist_del_rcu(&policy->bydst); + if (!chain) { void *p = xfrm_policy_inexact_insert(policy, dir, 0); From 1d38900cb85d5d311dbd23c2c93294527b82cd2b Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Fri, 4 Jan 2019 14:17:03 +0100 Subject: [PATCH 07/79] xfrm: policy: fix reinsertion on node merge "newpos" has wrong scope. It must be NULL on each iteration of the loop. Otherwise, when policy is to be inserted at the start, we would instead insert at point found by the previous loop-iteration instead. Also, we need to unlink the policy before we reinsert it to the new node, else we can get next-points-to-self loops. Because policies are only ordered by priority it is irrelevant which policy is "more recent" except when two policies have same priority. (the more recent one is placed after the older one). In these cases, we can use the ->pos id number to know which one is the 'older': the higher the id, the more recent the policy. So we only need to unlink all policies from the node that is about to be removed, and insert them to the replacement node. Fixes: 9cf545ebd591da ("xfrm: policy: store inexact policies in a tree ordered by destination address") Signed-off-by: Florian Westphal Signed-off-by: Steffen Klassert --- net/xfrm/xfrm_policy.c | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-) diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c index 24dfd1e47cf0..e691683223ee 100644 --- a/net/xfrm/xfrm_policy.c +++ b/net/xfrm/xfrm_policy.c @@ -823,13 +823,13 @@ static void xfrm_policy_inexact_list_reinsert(struct net *net, u16 family) { unsigned int matched_s, matched_d; - struct hlist_node *newpos = NULL; struct xfrm_policy *policy, *p; matched_s = 0; matched_d = 0; list_for_each_entry_reverse(policy, &net->xfrm.policy_all, walk.all) { + struct hlist_node *newpos = NULL; bool matches_s, matches_d; if (!policy->bydst_reinsert) @@ -839,7 +839,10 @@ static void xfrm_policy_inexact_list_reinsert(struct net *net, policy->bydst_reinsert = false; hlist_for_each_entry(p, &n->hhead, bydst) { - if (policy->priority >= p->priority) + if (policy->priority > p->priority) + newpos = &p->bydst; + else if (policy->priority == p->priority && + policy->pos > p->pos) newpos = &p->bydst; else break; @@ -955,12 +958,11 @@ static void xfrm_policy_inexact_node_merge(struct net *net, family); } - hlist_for_each_entry(tmp, &v->hhead, bydst) - tmp->bydst_reinsert = true; - hlist_for_each_entry(tmp, &n->hhead, bydst) + hlist_for_each_entry(tmp, &v->hhead, bydst) { tmp->bydst_reinsert = true; + hlist_del_rcu(&tmp->bydst); + } - INIT_HLIST_HEAD(&n->hhead); xfrm_policy_inexact_list_reinsert(net, n, family); } From fcf86f55f2d4a6b58da5feccd45d9584edc17c5a Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Fri, 4 Jan 2019 14:17:04 +0100 Subject: [PATCH 08/79] selftests: xfrm: alter htresh to trigger move of policies to hash table ... and back to inexact tree. Repeat ping test after each htresh change: lookup results must not change. Signed-off-by: Florian Westphal Signed-off-by: Steffen Klassert --- tools/testing/selftests/net/xfrm_policy.sh | 44 ++++++++++++++++++++-- 1 file changed, 40 insertions(+), 4 deletions(-) diff --git a/tools/testing/selftests/net/xfrm_policy.sh b/tools/testing/selftests/net/xfrm_policy.sh index b5a1b565a7e6..8ce54600d4d1 100755 --- a/tools/testing/selftests/net/xfrm_policy.sh +++ b/tools/testing/selftests/net/xfrm_policy.sh @@ -28,6 +28,19 @@ KEY_AES=0x0123456789abcdef0123456789012345 SPI1=0x1 SPI2=0x2 +do_esp_policy() { + local ns=$1 + local me=$2 + local remote=$3 + local lnet=$4 + local rnet=$5 + + # to encrypt packets as they go out (includes forwarded packets that need encapsulation) + ip -net $ns xfrm policy add src $lnet dst $rnet dir out tmpl src $me dst $remote proto esp mode tunnel priority 100 action allow + # to fwd decrypted packets after esp processing: + ip -net $ns xfrm policy add src $rnet dst $lnet dir fwd tmpl src $remote dst $me proto esp mode tunnel priority 100 action allow +} + do_esp() { local ns=$1 local me=$2 @@ -40,10 +53,7 @@ do_esp() { ip -net $ns xfrm state add src $remote dst $me proto esp spi $spi_in enc aes $KEY_AES auth sha1 $KEY_SHA mode tunnel sel src $rnet dst $lnet ip -net $ns xfrm state add src $me dst $remote proto esp spi $spi_out enc aes $KEY_AES auth sha1 $KEY_SHA mode tunnel sel src $lnet dst $rnet - # to encrypt packets as they go out (includes forwarded packets that need encapsulation) - ip -net $ns xfrm policy add src $lnet dst $rnet dir out tmpl src $me dst $remote proto esp mode tunnel priority 100 action allow - # to fwd decrypted packets after esp processing: - ip -net $ns xfrm policy add src $rnet dst $lnet dir fwd tmpl src $remote dst $me proto esp mode tunnel priority 100 action allow + do_esp_policy $ns $me $remote $lnet $rnet } # add policies with different netmasks, to make sure kernel carries @@ -370,6 +380,32 @@ if [ $? -ne 0 ]; then ret=1 fi +for n in ns3 ns4;do + ip -net $n xfrm policy set hthresh4 28 24 hthresh6 126 125 + sleep $((RANDOM%5)) +done + +check_exceptions "exceptions and block policies after hresh changes" + +# full flush of policy db, check everything gets freed incl. internal meta data +ip -net ns3 xfrm policy flush + +do_esp_policy ns3 10.0.3.1 10.0.3.10 10.0.1.0/24 10.0.2.0/24 +do_exception ns3 10.0.3.1 10.0.3.10 10.0.2.253 10.0.2.240/28 + +# move inexact policies to hash table +ip -net ns3 xfrm policy set hthresh4 16 16 + +sleep $((RANDOM%5)) +check_exceptions "exceptions and block policies after hthresh change in ns3" + +# restore original hthresh settings -- move policies back to tables +for n in ns3 ns4;do + ip -net $n xfrm policy set hthresh4 32 32 hthresh6 128 128 + sleep $((RANDOM%5)) +done +check_exceptions "exceptions and block policies after hresh change to normal" + for i in 1 2 3 4;do ip netns del ns$i;done exit $ret From 12750abad517a991c4568969bc748db302ab52cd Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Fri, 4 Jan 2019 14:17:05 +0100 Subject: [PATCH 09/79] xfrm: policy: fix infinite loop when merging src-nodes With very small change to test script we can trigger softlockup due to bogus assignment of 'p' (policy to be examined) on restart. Previously the two to-be-merged nodes had same address/prefixlength pair, so no erase/reinsert was necessary, we only had to append the list from node a to b. If prefix lengths are different, the node has to be deleted and re-inserted into the tree, with the updated prefix length. This was broken; due to bogus update to 'p' this loops forever. Add a 'restart' label and use that instead. While at it, don't perform the unneeded reinserts of the policies that are already sorted into the 'new' node. A previous patch in this series made xfrm_policy_inexact_list_reinsert() use the relative position indicator to sort policies according to age in case priorities are identical. Fixes: 6ac098b2a9d30 ("xfrm: policy: add 2nd-level saddr trees for inexact policies") Signed-off-by: Florian Westphal Signed-off-by: Steffen Klassert --- net/xfrm/xfrm_policy.c | 15 +++++++-------- tools/testing/selftests/net/xfrm_policy.sh | 4 ++-- 2 files changed, 9 insertions(+), 10 deletions(-) diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c index e691683223ee..8cfd75b62396 100644 --- a/net/xfrm/xfrm_policy.c +++ b/net/xfrm/xfrm_policy.c @@ -886,12 +886,13 @@ static void xfrm_policy_inexact_node_reinsert(struct net *net, struct rb_root *new, u16 family) { - struct rb_node **p, *parent = NULL; struct xfrm_pol_inexact_node *node; + struct rb_node **p, *parent; /* we should not have another subtree here */ WARN_ON_ONCE(!RB_EMPTY_ROOT(&n->root)); - +restart: + parent = NULL; p = &new->rb_node; while (*p) { u8 prefixlen; @@ -911,12 +912,11 @@ static void xfrm_policy_inexact_node_reinsert(struct net *net, } else { struct xfrm_policy *tmp; - hlist_for_each_entry(tmp, &node->hhead, bydst) - tmp->bydst_reinsert = true; - hlist_for_each_entry(tmp, &n->hhead, bydst) + hlist_for_each_entry(tmp, &n->hhead, bydst) { tmp->bydst_reinsert = true; + hlist_del_rcu(&tmp->bydst); + } - INIT_HLIST_HEAD(&node->hhead); xfrm_policy_inexact_list_reinsert(net, node, family); if (node->prefixlen == n->prefixlen) { @@ -928,8 +928,7 @@ static void xfrm_policy_inexact_node_reinsert(struct net *net, kfree_rcu(n, rcu); n = node; n->prefixlen = prefixlen; - *p = new->rb_node; - parent = NULL; + goto restart; } } diff --git a/tools/testing/selftests/net/xfrm_policy.sh b/tools/testing/selftests/net/xfrm_policy.sh index 8ce54600d4d1..71d7fdc513c1 100755 --- a/tools/testing/selftests/net/xfrm_policy.sh +++ b/tools/testing/selftests/net/xfrm_policy.sh @@ -78,8 +78,8 @@ do_overlap() # adds a new node in the 10.0.0.0/24 tree (dst node exists). ip -net $ns xfrm policy add src 10.2.0.0/24 dst 10.0.0.0/24 dir fwd priority 200 action block - # adds a 10.2.0.0/24 node, but for different dst. - ip -net $ns xfrm policy add src 10.2.0.0/24 dst 10.0.1.0/24 dir fwd priority 200 action block + # adds a 10.2.0.0/23 node, but for different dst. + ip -net $ns xfrm policy add src 10.2.0.0/23 dst 10.0.1.0/24 dir fwd priority 200 action block # dst now overlaps with the 10.0.1.0/24 ESP policy in fwd. # kernel must 'promote' existing one (10.0.0.0/24) to 10.0.0.0/23. From dd9ee3444014e8f28c0eefc9fffc9ac9c5248c12 Mon Sep 17 00:00:00 2001 From: Su Yanjun Date: Sun, 6 Jan 2019 21:31:20 -0500 Subject: [PATCH 10/79] vti4: Fix a ipip packet processing bug in 'IPCOMP' virtual tunnel Recently we run a network test over ipcomp virtual tunnel.We find that if a ipv4 packet needs fragment, then the peer can't receive it. We deep into the code and find that when packet need fragment the smaller fragment will be encapsulated by ipip not ipcomp. So when the ipip packet goes into xfrm, it's skb->dev is not properly set. The ipv4 reassembly code always set skb'dev to the last fragment's dev. After ipv4 defrag processing, when the kernel rp_filter parameter is set, the skb will be drop by -EXDEV error. This patch adds compatible support for the ipip process in ipcomp virtual tunnel. Signed-off-by: Su Yanjun Signed-off-by: Steffen Klassert --- net/ipv4/ip_vti.c | 50 +++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 50 insertions(+) diff --git a/net/ipv4/ip_vti.c b/net/ipv4/ip_vti.c index d7b43e700023..68a21bf75dd0 100644 --- a/net/ipv4/ip_vti.c +++ b/net/ipv4/ip_vti.c @@ -74,6 +74,33 @@ static int vti_input(struct sk_buff *skb, int nexthdr, __be32 spi, return 0; } +static int vti_input_ipip(struct sk_buff *skb, int nexthdr, __be32 spi, + int encap_type) +{ + struct ip_tunnel *tunnel; + const struct iphdr *iph = ip_hdr(skb); + struct net *net = dev_net(skb->dev); + struct ip_tunnel_net *itn = net_generic(net, vti_net_id); + + tunnel = ip_tunnel_lookup(itn, skb->dev->ifindex, TUNNEL_NO_KEY, + iph->saddr, iph->daddr, 0); + if (tunnel) { + if (!xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb)) + goto drop; + + XFRM_TUNNEL_SKB_CB(skb)->tunnel.ip4 = tunnel; + + skb->dev = tunnel->dev; + + return xfrm_input(skb, nexthdr, spi, encap_type); + } + + return -EINVAL; +drop: + kfree_skb(skb); + return 0; +} + static int vti_rcv(struct sk_buff *skb) { XFRM_SPI_SKB_CB(skb)->family = AF_INET; @@ -82,6 +109,14 @@ static int vti_rcv(struct sk_buff *skb) return vti_input(skb, ip_hdr(skb)->protocol, 0, 0); } +static int vti_rcv_ipip(struct sk_buff *skb) +{ + XFRM_SPI_SKB_CB(skb)->family = AF_INET; + XFRM_SPI_SKB_CB(skb)->daddroff = offsetof(struct iphdr, daddr); + + return vti_input_ipip(skb, ip_hdr(skb)->protocol, ip_hdr(skb)->saddr, 0); +} + static int vti_rcv_cb(struct sk_buff *skb, int err) { unsigned short family; @@ -435,6 +470,12 @@ static struct xfrm4_protocol vti_ipcomp4_protocol __read_mostly = { .priority = 100, }; +static struct xfrm_tunnel ipip_handler __read_mostly = { + .handler = vti_rcv_ipip, + .err_handler = vti4_err, + .priority = 0, +}; + static int __net_init vti_init_net(struct net *net) { int err; @@ -603,6 +644,13 @@ static int __init vti_init(void) if (err < 0) goto xfrm_proto_comp_failed; + msg = "ipip tunnel"; + err = xfrm4_tunnel_register(&ipip_handler, AF_INET); + if (err < 0) { + pr_info("%s: cant't register tunnel\n",__func__); + goto xfrm_tunnel_failed; + } + msg = "netlink interface"; err = rtnl_link_register(&vti_link_ops); if (err < 0) @@ -612,6 +660,8 @@ static int __init vti_init(void) rtnl_link_failed: xfrm4_protocol_deregister(&vti_ipcomp4_protocol, IPPROTO_COMP); +xfrm_tunnel_failed: + xfrm4_tunnel_deregister(&ipip_handler, AF_INET); xfrm_proto_comp_failed: xfrm4_protocol_deregister(&vti_ah4_protocol, IPPROTO_AH); xfrm_proto_ah_failed: From 35e6103861a3a970de6c84688c6e7a1f65b164ca Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Wed, 9 Jan 2019 14:37:34 +0100 Subject: [PATCH 11/79] xfrm: refine validation of template and selector families The check assumes that in transport mode, the first templates family must match the address family of the policy selector. Syzkaller managed to build a template using MODE_ROUTEOPTIMIZATION, with ipv4-in-ipv6 chain, leading to following splat: BUG: KASAN: stack-out-of-bounds in xfrm_state_find+0x1db/0x1854 Read of size 4 at addr ffff888063e57aa0 by task a.out/2050 xfrm_state_find+0x1db/0x1854 xfrm_tmpl_resolve+0x100/0x1d0 xfrm_resolve_and_create_bundle+0x108/0x1000 [..] Problem is that addresses point into flowi4 struct, but xfrm_state_find treats them as being ipv6 because it uses templ->encap_family is used (AF_INET6 in case of reproducer) rather than family (AF_INET). This patch inverts the logic: Enforce 'template family must match selector' EXCEPT for tunnel and BEET mode. In BEET and Tunnel mode, xfrm_tmpl_resolve_one will have remote/local address pointers changed to point at the addresses found in the template, rather than the flowi ones, so no oob read will occur. Reported-by: 3ntr0py1337@gmail.com Reported-by: Daniel Borkmann Signed-off-by: Florian Westphal Signed-off-by: Steffen Klassert --- net/xfrm/xfrm_user.c | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-) diff --git a/net/xfrm/xfrm_user.c b/net/xfrm/xfrm_user.c index 277c1c46fe94..c6d26afcf89d 100644 --- a/net/xfrm/xfrm_user.c +++ b/net/xfrm/xfrm_user.c @@ -1488,10 +1488,15 @@ static int validate_tmpl(int nr, struct xfrm_user_tmpl *ut, u16 family) if (!ut[i].family) ut[i].family = family; - if ((ut[i].mode == XFRM_MODE_TRANSPORT) && - (ut[i].family != prev_family)) - return -EINVAL; - + switch (ut[i].mode) { + case XFRM_MODE_TUNNEL: + case XFRM_MODE_BEET: + break; + default: + if (ut[i].family != prev_family) + return -EINVAL; + break; + } if (ut[i].mode >= XFRM_MODE_MAX) return -EINVAL; From 5962dd22f0ff6f7d72fff974b3c637d52586643e Mon Sep 17 00:00:00 2001 From: Sinan Kaya Date: Wed, 2 Jan 2019 18:10:37 +0000 Subject: [PATCH 12/79] x86/intel/lpss: Make PCI dependency explicit After commit 5d32a66541c4 ("PCI/ACPI: Allow ACPI to be built without CONFIG_PCI set") dependencies on CONFIG_PCI that previously were satisfied implicitly through dependencies on CONFIG_ACPI have to be specified directly. LPSS code relies on PCI infrastructure but this dependency has not been explicitly called out so do that. Fixes: 5d32a66541c46 ("PCI/ACPI: Allow ACPI to be built without CONFIG_PCI set") Signed-off-by: Sinan Kaya Signed-off-by: Borislav Petkov Cc: "H. Peter Anvin" Cc: linux-acpi@vger.kernel.org Cc: x86-ml Link: https://lkml.kernel.org/r/20190102181038.4418-11-okaya@kernel.org --- arch/x86/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 15af091611e2..4b4a7f32b68e 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -617,7 +617,7 @@ config X86_INTEL_QUARK config X86_INTEL_LPSS bool "Intel Low Power Subsystem Support" - depends on X86 && ACPI + depends on X86 && ACPI && PCI select COMMON_CLK select PINCTRL select IOSF_MBI From 7e6fc2f50a3197d0e82d1c0e86282976c9e6c8a4 Mon Sep 17 00:00:00 2001 From: Daniel Drake Date: Mon, 7 Jan 2019 11:40:24 +0800 Subject: [PATCH 13/79] x86/kaslr: Fix incorrect i8254 outb() parameters The outb() function takes parameters value and port, in that order. Fix the parameters used in the kalsr i8254 fallback code. Fixes: 5bfce5ef55cb ("x86, kaslr: Provide randomness functions") Signed-off-by: Daniel Drake Signed-off-by: Thomas Gleixner Cc: bp@alien8.de Cc: hpa@zytor.com Cc: linux@endlessm.com Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/20190107034024.15005-1-drake@endlessm.com --- arch/x86/lib/kaslr.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/arch/x86/lib/kaslr.c b/arch/x86/lib/kaslr.c index 79778ab200e4..a53665116458 100644 --- a/arch/x86/lib/kaslr.c +++ b/arch/x86/lib/kaslr.c @@ -36,8 +36,8 @@ static inline u16 i8254(void) u16 status, timer; do { - outb(I8254_PORT_CONTROL, - I8254_CMD_READBACK | I8254_SELECT_COUNTER0); + outb(I8254_CMD_READBACK | I8254_SELECT_COUNTER0, + I8254_PORT_CONTROL); status = inb(I8254_PORT_COUNTER0); timer = inb(I8254_PORT_COUNTER0); timer |= inb(I8254_PORT_COUNTER0) << 8; From a31e184e4f69965c99c04cc5eb8a4920e0c63737 Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Wed, 2 Jan 2019 13:56:55 -0800 Subject: [PATCH 14/79] x86/pkeys: Properly copy pkey state at fork() Memory protection key behavior should be the same in a child as it was in the parent before a fork. But, there is a bug that resets the state in the child at fork instead of preserving it. The creation of new mm's is a bit convoluted. At fork(), the code does: 1. memcpy() the parent mm to initialize child 2. mm_init() to initalize some select stuff stuff 3. dup_mmap() to create true copies that memcpy() did not do right For pkeys two bits of state need to be preserved across a fork: 'execute_only_pkey' and 'pkey_allocation_map'. Those are preserved by the memcpy(), but mm_init() invokes init_new_context() which overwrites 'execute_only_pkey' and 'pkey_allocation_map' with "new" values. The author of the code erroneously believed that init_new_context is *only* called at execve()-time. But, alas, init_new_context() is used at execve() and fork(). The result is that, after a fork(), the child's pkey state ends up looking like it does after an execve(), which is totally wrong. pkeys that are already allocated can be allocated again, for instance. To fix this, add code called by dup_mmap() to copy the pkey state from parent to child explicitly. Also add a comment above init_new_context() to make it more clear to the next poor sod what this code is used for. Fixes: e8c24d3a23a ("x86/pkeys: Allocation/free syscalls") Signed-off-by: Dave Hansen Signed-off-by: Thomas Gleixner Reviewed-by: Thomas Gleixner Cc: bp@alien8.de Cc: hpa@zytor.com Cc: peterz@infradead.org Cc: mpe@ellerman.id.au Cc: will.deacon@arm.com Cc: luto@kernel.org Cc: jroedel@suse.de Cc: stable@vger.kernel.org Cc: Borislav Petkov Cc: "H. Peter Anvin" Cc: Peter Zijlstra Cc: Michael Ellerman Cc: Will Deacon Cc: Andy Lutomirski Cc: Joerg Roedel Link: https://lkml.kernel.org/r/20190102215655.7A69518C@viggo.jf.intel.com --- arch/x86/include/asm/mmu_context.h | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) diff --git a/arch/x86/include/asm/mmu_context.h b/arch/x86/include/asm/mmu_context.h index 0ca50611e8ce..19d18fae6ec6 100644 --- a/arch/x86/include/asm/mmu_context.h +++ b/arch/x86/include/asm/mmu_context.h @@ -178,6 +178,10 @@ static inline void switch_ldt(struct mm_struct *prev, struct mm_struct *next) void enter_lazy_tlb(struct mm_struct *mm, struct task_struct *tsk); +/* + * Init a new mm. Used on mm copies, like at fork() + * and on mm's that are brand-new, like at execve(). + */ static inline int init_new_context(struct task_struct *tsk, struct mm_struct *mm) { @@ -228,8 +232,22 @@ do { \ } while (0) #endif +static inline void arch_dup_pkeys(struct mm_struct *oldmm, + struct mm_struct *mm) +{ +#ifdef CONFIG_X86_INTEL_MEMORY_PROTECTION_KEYS + if (!cpu_feature_enabled(X86_FEATURE_OSPKE)) + return; + + /* Duplicate the oldmm pkey state in mm: */ + mm->context.pkey_allocation_map = oldmm->context.pkey_allocation_map; + mm->context.execute_only_pkey = oldmm->context.execute_only_pkey; +#endif +} + static inline int arch_dup_mmap(struct mm_struct *oldmm, struct mm_struct *mm) { + arch_dup_pkeys(oldmm, mm); paravirt_arch_dup_mmap(oldmm, mm); return ldt_dup_context(oldmm, mm); } From e1812933b17be7814f51b6c310c5d1ced7a9a5f5 Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Wed, 2 Jan 2019 13:56:57 -0800 Subject: [PATCH 15/79] x86/selftests/pkeys: Fork() to check for state being preserved There was a bug where the per-mm pkey state was not being preserved across fork() in the child. fork() is performed in the pkey selftests, but all of the pkey activity is performed in the parent. The child does not perform any actions sensitive to pkey state. To make the test more sensitive to these kinds of bugs, add a fork() where the parent exits, and execution continues in the child. To achieve this let the key exhaustion test not terminate at the first allocation failure and fork after 2*NR_PKEYS loops and continue in the child. Signed-off-by: Dave Hansen Signed-off-by: Thomas Gleixner Cc: bp@alien8.de Cc: hpa@zytor.com Cc: peterz@infradead.org Cc: mpe@ellerman.id.au Cc: will.deacon@arm.com Cc: luto@kernel.org Cc: jroedel@suse.de Cc: stable@vger.kernel.org Cc: Borislav Petkov Cc: "H. Peter Anvin" Cc: Peter Zijlstra Cc: Michael Ellerman Cc: Will Deacon Cc: Andy Lutomirski Cc: Joerg Roedel Link: https://lkml.kernel.org/r/20190102215657.585704B7@viggo.jf.intel.com --- tools/testing/selftests/x86/protection_keys.c | 41 ++++++++++++++----- 1 file changed, 31 insertions(+), 10 deletions(-) diff --git a/tools/testing/selftests/x86/protection_keys.c b/tools/testing/selftests/x86/protection_keys.c index 460b4bdf4c1e..5d546dcdbc80 100644 --- a/tools/testing/selftests/x86/protection_keys.c +++ b/tools/testing/selftests/x86/protection_keys.c @@ -1133,6 +1133,21 @@ void test_pkey_syscalls_bad_args(int *ptr, u16 pkey) pkey_assert(err); } +void become_child(void) +{ + pid_t forkret; + + forkret = fork(); + pkey_assert(forkret >= 0); + dprintf3("[%d] fork() ret: %d\n", getpid(), forkret); + + if (!forkret) { + /* in the child */ + return; + } + exit(0); +} + /* Assumes that all pkeys other than 'pkey' are unallocated */ void test_pkey_alloc_exhaust(int *ptr, u16 pkey) { @@ -1141,7 +1156,7 @@ void test_pkey_alloc_exhaust(int *ptr, u16 pkey) int nr_allocated_pkeys = 0; int i; - for (i = 0; i < NR_PKEYS*2; i++) { + for (i = 0; i < NR_PKEYS*3; i++) { int new_pkey; dprintf1("%s() alloc loop: %d\n", __func__, i); new_pkey = alloc_pkey(); @@ -1152,20 +1167,26 @@ void test_pkey_alloc_exhaust(int *ptr, u16 pkey) if ((new_pkey == -1) && (errno == ENOSPC)) { dprintf2("%s() failed to allocate pkey after %d tries\n", __func__, nr_allocated_pkeys); - break; + } else { + /* + * Ensure the number of successes never + * exceeds the number of keys supported + * in the hardware. + */ + pkey_assert(nr_allocated_pkeys < NR_PKEYS); + allocated_pkeys[nr_allocated_pkeys++] = new_pkey; } - pkey_assert(nr_allocated_pkeys < NR_PKEYS); - allocated_pkeys[nr_allocated_pkeys++] = new_pkey; + + /* + * Make sure that allocation state is properly + * preserved across fork(). + */ + if (i == NR_PKEYS*2) + become_child(); } dprintf3("%s()::%d\n", __func__, __LINE__); - /* - * ensure it did not reach the end of the loop without - * failure: - */ - pkey_assert(i < NR_PKEYS*2); - /* * There are 16 pkeys supported in hardware. Three are * allocated by the time we get here: From bddda606ec76550dd63592e32a6e87e7d32583f7 Mon Sep 17 00:00:00 2001 From: Srinivas Ramana Date: Thu, 20 Dec 2018 19:05:57 +0530 Subject: [PATCH 16/79] genirq: Make sure the initial affinity is not empty If all CPUs in the irq_default_affinity mask are offline when an interrupt is initialized then irq_setup_affinity() can set an empty affinity mask for a newly allocated interrupt. Fix this by falling back to cpu_online_mask in case the resulting affinity mask is zero. Signed-off-by: Srinivas Ramana Signed-off-by: Thomas Gleixner Cc: linux-arm-msm@vger.kernel.org Link: https://lkml.kernel.org/r/1545312957-8504-1-git-send-email-sramana@codeaurora.org --- kernel/irq/manage.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index a4888ce4667a..84b54a17b95d 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -393,6 +393,9 @@ int irq_setup_affinity(struct irq_desc *desc) } cpumask_and(&mask, cpu_online_mask, set); + if (cpumask_empty(&mask)) + cpumask_copy(&mask, cpu_online_mask); + if (node != NUMA_NO_NODE) { const struct cpumask *nodemask = cpumask_of_node(node); From bf7d28c53453ea904584960de55e33e03b9d93b1 Mon Sep 17 00:00:00 2001 From: Peng Hao Date: Sat, 29 Dec 2018 14:34:12 +0800 Subject: [PATCH 17/79] x86/mm/mem_encrypt: Fix erroneous sizeof() Using sizeof(pointer) for determining the size of a memset() only works when the size of the pointer and the size of type to which it points are the same. For pte_t this is only true for 64bit and 32bit-NONPAE. On 32bit PAE systems this is wrong as the pointer size is 4 byte but the PTE entry is 8 bytes. It's actually not a real world issue as this code depends on 64bit, but it's wrong nevertheless. Use sizeof(*p) for correctness sake. Fixes: aad983913d77 ("x86/mm/encrypt: Simplify sme_populate_pgd() and sme_populate_pgd_large()") Signed-off-by: Peng Hao Signed-off-by: Thomas Gleixner Cc: Kirill A. Shutemov Cc: Tom Lendacky Cc: dave.hansen@linux.intel.com Cc: peterz@infradead.org Cc: luto@kernel.org Link: https://lkml.kernel.org/r/1546065252-97996-1-git-send-email-peng.hao2@zte.com.cn --- arch/x86/mm/mem_encrypt_identity.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/arch/x86/mm/mem_encrypt_identity.c b/arch/x86/mm/mem_encrypt_identity.c index a19ef1a416ff..4aa9b1480866 100644 --- a/arch/x86/mm/mem_encrypt_identity.c +++ b/arch/x86/mm/mem_encrypt_identity.c @@ -158,8 +158,8 @@ static void __init sme_populate_pgd(struct sme_populate_pgd_data *ppd) pmd = pmd_offset(pud, ppd->vaddr); if (pmd_none(*pmd)) { pte = ppd->pgtable_area; - memset(pte, 0, sizeof(pte) * PTRS_PER_PTE); - ppd->pgtable_area += sizeof(pte) * PTRS_PER_PTE; + memset(pte, 0, sizeof(*pte) * PTRS_PER_PTE); + ppd->pgtable_area += sizeof(*pte) * PTRS_PER_PTE; set_pmd(pmd, __pmd(PMD_FLAGS | __pa(pte))); } From 993a110319a4a60aadbd02f6defdebe048f7773b Mon Sep 17 00:00:00 2001 From: Dave Young Date: Fri, 28 Dec 2018 09:12:47 +0800 Subject: [PATCH 18/79] x86/kexec: Fix a kexec_file_load() failure Commit b6664ba42f14 ("s390, kexec_file: drop arch_kexec_mem_walk()") changed the behavior of kexec_locate_mem_hole(): it will try to allocate free memory only when kbuf.mem is initialized to zero. However, x86's kexec_file_load() implementation reuses a struct kexec_buf allocated on the stack and its kbuf.mem member gets set by each kexec_add_buffer() invocation. The second kexec_add_buffer() will reuse the same kbuf but not reinitialize kbuf.mem. Therefore, explictily reset kbuf.mem each time in order for kexec_locate_mem_hole() to locate a free memory region each time. [ bp: massage commit message. ] Fixes: b6664ba42f14 ("s390, kexec_file: drop arch_kexec_mem_walk()") Signed-off-by: Dave Young Signed-off-by: Borislav Petkov Acked-by: Baoquan He Cc: "Eric W. Biederman" Cc: "H. Peter Anvin" Cc: AKASHI Takahiro Cc: Andrew Morton Cc: Ingo Molnar Cc: Martin Schwidefsky Cc: Philipp Rudo Cc: Thomas Gleixner Cc: Vivek Goyal Cc: Yannik Sembritzki Cc: Yi Wang Cc: kexec@lists.infradead.org Cc: x86-ml Link: https://lkml.kernel.org/r/20181228011247.GA9999@dhcp-128-65.nay.redhat.com --- arch/x86/kernel/crash.c | 1 + arch/x86/kernel/kexec-bzimage64.c | 2 ++ 2 files changed, 3 insertions(+) diff --git a/arch/x86/kernel/crash.c b/arch/x86/kernel/crash.c index c8b07d8ea5a2..17ffc869cab8 100644 --- a/arch/x86/kernel/crash.c +++ b/arch/x86/kernel/crash.c @@ -470,6 +470,7 @@ int crash_load_segments(struct kimage *image) kbuf.memsz = kbuf.bufsz; kbuf.buf_align = ELF_CORE_HEADER_ALIGN; + kbuf.mem = KEXEC_BUF_MEM_UNKNOWN; ret = kexec_add_buffer(&kbuf); if (ret) { vfree((void *)image->arch.elf_headers); diff --git a/arch/x86/kernel/kexec-bzimage64.c b/arch/x86/kernel/kexec-bzimage64.c index 278cd07228dd..0d5efa34f359 100644 --- a/arch/x86/kernel/kexec-bzimage64.c +++ b/arch/x86/kernel/kexec-bzimage64.c @@ -434,6 +434,7 @@ static void *bzImage64_load(struct kimage *image, char *kernel, kbuf.memsz = PAGE_ALIGN(header->init_size); kbuf.buf_align = header->kernel_alignment; kbuf.buf_min = MIN_KERNEL_LOAD_ADDR; + kbuf.mem = KEXEC_BUF_MEM_UNKNOWN; ret = kexec_add_buffer(&kbuf); if (ret) goto out_free_params; @@ -448,6 +449,7 @@ static void *bzImage64_load(struct kimage *image, char *kernel, kbuf.bufsz = kbuf.memsz = initrd_len; kbuf.buf_align = PAGE_SIZE; kbuf.buf_min = MIN_INITRD_LOAD_ADDR; + kbuf.mem = KEXEC_BUF_MEM_UNKNOWN; ret = kexec_add_buffer(&kbuf); if (ret) goto out_free_params; From 93ad0fc088c5b4631f796c995bdd27a082ef33a6 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Fri, 11 Jan 2019 14:33:16 +0100 Subject: [PATCH 19/79] posix-cpu-timers: Unbreak timer rearming The recent commit which prevented a division by 0 issue in the alarm timer code broke posix CPU timers as an unwanted side effect. The reason is that the common rearm code checks for timer->it_interval being 0 now. What went unnoticed is that the posix cpu timer setup does not initialize timer->it_interval as it stores the interval in CPU timer specific storage. The reason for the separate storage is historical as the posix CPU timers always had a 64bit nanoseconds representation internally while timer->it_interval is type ktime_t which used to be a modified timespec representation on 32bit machines. Instead of reverting the offending commit and fixing the alarmtimer issue in the alarmtimer code, store the interval in timer->it_interval at CPU timer setup time so the common code check works. This also repairs the existing inconistency of the posix CPU timer code which kept a single shot timer armed despite of the interval being 0. The separate storage can be removed in mainline, but that needs to be a separate commit as the current one has to be backported to stable kernels. Fixes: 0e334db6bb4b ("posix-timers: Fix division by zero bug") Reported-by: H.J. Lu Signed-off-by: Thomas Gleixner Cc: John Stultz Cc: Peter Zijlstra Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/20190111133500.840117406@linutronix.de --- kernel/time/posix-cpu-timers.c | 1 + 1 file changed, 1 insertion(+) diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c index 8f0644af40be..80f955210861 100644 --- a/kernel/time/posix-cpu-timers.c +++ b/kernel/time/posix-cpu-timers.c @@ -685,6 +685,7 @@ static int posix_cpu_timer_set(struct k_itimer *timer, int timer_flags, * set up the signal and overrun bookkeeping. */ timer->it.cpu.incr = timespec64_to_ns(&new->it_interval); + timer->it_interval = ns_to_ktime(timer->it.cpu.incr); /* * This acts as a modification timestamp for the timer, From e2612cd496e7b465711d219ea6118893d7253f52 Mon Sep 17 00:00:00 2001 From: Benedict Wong Date: Mon, 14 Jan 2019 11:24:38 -0800 Subject: [PATCH 20/79] xfrm: Make set-mark default behavior backward compatible MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Fixes 9b42c1f179a6, which changed the default route lookup behavior for tunnel mode SAs in the outbound direction to use the skb mark, whereas previously mark=0 was used if the output mark was unspecified. In mark-based routing schemes such as Android’s, this change in default behavior causes routing loops or lookup failures. This patch restores the default behavior of using a 0 mark while still incorporating the skb mark if the SET_MARK (and SET_MARK_MASK) is specified. Tested with additions to Android's kernel unit test suite: https://android-review.googlesource.com/c/kernel/tests/+/860150 Fixes: 9b42c1f179a6 ("xfrm: Extend the output_mark to support input direction and masking") Signed-off-by: Benedict Wong Signed-off-by: Steffen Klassert --- net/xfrm/xfrm_policy.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c index 8cfd75b62396..ba0a4048c846 100644 --- a/net/xfrm/xfrm_policy.c +++ b/net/xfrm/xfrm_policy.c @@ -2600,7 +2600,10 @@ static struct dst_entry *xfrm_bundle_create(struct xfrm_policy *policy, dst_copy_metrics(dst1, dst); if (xfrm[i]->props.mode != XFRM_MODE_TRANSPORT) { - __u32 mark = xfrm_smark_get(fl->flowi_mark, xfrm[i]); + __u32 mark = 0; + + if (xfrm[i]->props.smark.v || xfrm[i]->props.smark.m) + mark = xfrm_smark_get(fl->flowi_mark, xfrm[i]); family = xfrm[i]->props.family; dst = xfrm_dst_lookup(xfrm[i], tos, fl->flowi_oif, From 1d47f48bf2d1608c2d6eb76b3ec7a5ec0c3f9e95 Mon Sep 17 00:00:00 2001 From: Loic Pallardy Date: Fri, 11 Jan 2019 18:54:31 +0100 Subject: [PATCH 21/79] irqchip/stm32-exti: Add domain translate function Domain translate function is needed to recover irq configuration parameters from DT node Fixes: 927abfc4461e ("irqchip/stm32: Add stm32mp1 support with hierarchy domain") Signed-off-by: Loic Pallardy Signed-off-by: Marc Zyngier --- drivers/irqchip/irq-stm32-exti.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/irqchip/irq-stm32-exti.c b/drivers/irqchip/irq-stm32-exti.c index 6edfd4bfa169..a93296b9b45d 100644 --- a/drivers/irqchip/irq-stm32-exti.c +++ b/drivers/irqchip/irq-stm32-exti.c @@ -822,6 +822,7 @@ static int __init stm32_exti_init(const struct stm32_exti_drv_data *drv_data, static const struct irq_domain_ops stm32_exti_h_domain_ops = { .alloc = stm32_exti_h_domain_alloc, .free = irq_domain_free_irqs_common, + .xlate = irq_domain_xlate_twocell, }; static int From c530bb8a726a37811e9fb5d68cd6b5408173b545 Mon Sep 17 00:00:00 2001 From: Yang Yingliang Date: Fri, 26 Oct 2018 15:51:17 +0800 Subject: [PATCH 22/79] irqchip/gic-v3-mbi: Fix uninitialized mbi_lock The mbi_lock mutex is left uninitialized, so let's use DEFINE_MUTEX to initialize it statically. Fixes: 505287525c24d ("irqchip/gic-v3: Add support for Message Based Interrupts as an MSI controller") Signed-off-by: Yang Yingliang Signed-off-by: Marc Zyngier --- drivers/irqchip/irq-gic-v3-mbi.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/irqchip/irq-gic-v3-mbi.c b/drivers/irqchip/irq-gic-v3-mbi.c index ad70e7c416e3..fbfa7ff6deb1 100644 --- a/drivers/irqchip/irq-gic-v3-mbi.c +++ b/drivers/irqchip/irq-gic-v3-mbi.c @@ -24,7 +24,7 @@ struct mbi_range { unsigned long *bm; }; -static struct mutex mbi_lock; +static DEFINE_MUTEX(mbi_lock); static phys_addr_t mbi_phys_base; static struct mbi_range *mbi_ranges; static unsigned int mbi_range_nr; From 8fa4e55bbfbfca3ad996517f679ab3648c03d8bb Mon Sep 17 00:00:00 2001 From: Linus Walleij Date: Thu, 3 Jan 2019 11:26:23 +0100 Subject: [PATCH 23/79] irqchip/madera: Drop GPIO includes This irqchip does not use anything GPIO-related so drop the GPIO includes. Signed-off-by: Linus Walleij Acked-by: Richard Fitzgerald Signed-off-by: Marc Zyngier --- drivers/irqchip/irq-madera.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/drivers/irqchip/irq-madera.c b/drivers/irqchip/irq-madera.c index e9256dee1a45..8b81271c823c 100644 --- a/drivers/irqchip/irq-madera.c +++ b/drivers/irqchip/irq-madera.c @@ -7,7 +7,6 @@ */ #include -#include #include #include #include @@ -16,7 +15,6 @@ #include #include #include -#include #include #include #include From 70921ae25f944423f0abf096f73455c586da0652 Mon Sep 17 00:00:00 2001 From: Jonathan Corbet Date: Tue, 8 Jan 2019 17:04:32 -0700 Subject: [PATCH 24/79] genirq: Fix the kerneldoc comment for struct irq_affinity_desc A recent commit added a new field but did not update the kerneldoc comment, leading to this build warning: ./include/linux/interrupt.h:268: warning: Function parameter or member 'is_managed' not described in 'irq_affinity_desc' Add the missing information, making the docs build 0.001% quieter. Fixes: c410abbbacb9 ("genirq/affinity: Add is_managed to struct irq_affinity_desc") Signed-off-by: Jonathan Corbet Signed-off-by: Thomas Gleixner Cc: Dou Liyang Link: https://lkml.kernel.org/r/20190108170432.59bae8a6@lwn.net --- include/linux/interrupt.h | 1 + 1 file changed, 1 insertion(+) diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h index c672f34235e7..4a728dba02e2 100644 --- a/include/linux/interrupt.h +++ b/include/linux/interrupt.h @@ -260,6 +260,7 @@ struct irq_affinity { /** * struct irq_affinity_desc - Interrupt affinity descriptor * @mask: cpumask to hold the affinity assignment + * @is_managed: 1 if the interrupt is managed internally */ struct irq_affinity_desc { struct cpumask mask; From fc24d75a7f91837d7918e40719575951820b2b8f Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Tue, 15 Jan 2019 09:58:16 -0700 Subject: [PATCH 25/79] x86/entry/64/compat: Fix stack switching for XEN PV While in the native case entry into the kernel happens on the trampoline stack, PV Xen kernels get entered with the current thread stack right away. Hence source and destination stacks are identical in that case, and special care is needed. Other than in sync_regs() the copying done on the INT80 path isn't NMI / #MC safe, as either of these events occurring in the middle of the stack copying would clobber data on the (source) stack. There is similar code in interrupt_entry() and nmi(), but there is no fixup required because those code paths are unreachable in XEN PV guests. [ tglx: Sanitized subject, changelog, Fixes tag and stable mail address. Sigh ] Fixes: 7f2590a110b8 ("x86/entry/64: Use a per-CPU trampoline stack for IDT entries") Signed-off-by: Jan Beulich Signed-off-by: Thomas Gleixner Reviewed-by: Juergen Gross Acked-by: Andy Lutomirski Cc: Peter Anvin Cc: xen-devel@lists.xenproject.org> Cc: Boris Ostrovsky Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/5C3E1128020000780020DFAD@prv1-mh.provo.novell.com --- arch/x86/entry/entry_64_compat.S | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/arch/x86/entry/entry_64_compat.S b/arch/x86/entry/entry_64_compat.S index 8eaf8952c408..39913770a44d 100644 --- a/arch/x86/entry/entry_64_compat.S +++ b/arch/x86/entry/entry_64_compat.S @@ -361,7 +361,8 @@ ENTRY(entry_INT80_compat) /* Need to switch before accessing the thread stack. */ SWITCH_TO_KERNEL_CR3 scratch_reg=%rdi - movq %rsp, %rdi + /* In the Xen PV case we already run on the thread stack. */ + ALTERNATIVE "movq %rsp, %rdi", "jmp .Lint80_keep_stack", X86_FEATURE_XENPV movq PER_CPU_VAR(cpu_current_top_of_stack), %rsp pushq 6*8(%rdi) /* regs->ss */ @@ -370,8 +371,9 @@ ENTRY(entry_INT80_compat) pushq 3*8(%rdi) /* regs->cs */ pushq 2*8(%rdi) /* regs->ip */ pushq 1*8(%rdi) /* regs->orig_ax */ - pushq (%rdi) /* pt_regs->di */ +.Lint80_keep_stack: + pushq %rsi /* pt_regs->si */ xorl %esi, %esi /* nospec si */ pushq %rdx /* pt_regs->dx */ From 12fee4cd5be2c4a73cc13d7ad76eb2d2feda8a71 Mon Sep 17 00:00:00 2001 From: Huacai Chen Date: Thu, 17 Jan 2019 11:00:09 +0800 Subject: [PATCH 26/79] genirq/irqdesc: Fix double increment in alloc_descs() The recent rework of alloc_descs() introduced a double increment of the loop counter. As a consequence only every second affinity mask is validated. Remove it. [ tglx: Massaged changelog ] Fixes: c410abbbacb9 ("genirq/affinity: Add is_managed to struct irq_affinity_desc") Signed-off-by: Huacai Chen Signed-off-by: Thomas Gleixner Cc: Fuxin Zhang Cc: Zhangjin Wu Cc: Huacai Chen Cc: Dou Liyang Link: https://lkml.kernel.org/r/1547694009-16261-1-git-send-email-chenhc@lemote.com --- kernel/irq/irqdesc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/irq/irqdesc.c b/kernel/irq/irqdesc.c index ee062b7939d3..ef8ad36cadcf 100644 --- a/kernel/irq/irqdesc.c +++ b/kernel/irq/irqdesc.c @@ -457,7 +457,7 @@ static int alloc_descs(unsigned int start, unsigned int cnt, int node, /* Validate affinity mask(s) */ if (affinity) { - for (i = 0; i < cnt; i++, i++) { + for (i = 0; i < cnt; i++) { if (cpumask_empty(&affinity[i].mask)) return -EINVAL; } From 12c44aba6618b7f6c437076e5722237190f6cd5f Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Mon, 14 Jan 2019 14:28:48 +0100 Subject: [PATCH 27/79] netfilter: nft_compat: use refcnt_t type for nft_xt reference count Using standard integer type was fine while all operations on it were guarded by the nftnl subsys mutex. This isn't true anymore: 1. transactions are guarded only by a pernet mutex, so concurrent rule manipulation in different netns is racy 2. the ->destroy hook runs from a work queue after the transaction mutex has been released already. cpu0 cpu1 (net 1) cpu2 (net 2) kworker nft_compat->destroy nft_compat->init nft_compat->init if (--nft_xt->ref == 0) nft_xt->ref++ nft_xt->ref++ Switch to refcount_t. Doing this however only fixes a minor aspect, nft_compat also performs linked-list operations in an unsafe way. This is addressed in the next two patches. Fixes: f102d66b335a ("netfilter: nf_tables: use dedicated mutex to guard transactions") Fixes: 0935d5588400 ("netfilter: nf_tables: asynchronous release") Reported-by: Taehee Yoo Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso --- net/netfilter/nft_compat.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) diff --git a/net/netfilter/nft_compat.c b/net/netfilter/nft_compat.c index 7334e0b80a5e..acc85acad31b 100644 --- a/net/netfilter/nft_compat.c +++ b/net/netfilter/nft_compat.c @@ -26,7 +26,7 @@ struct nft_xt { struct list_head head; struct nft_expr_ops ops; - unsigned int refcnt; + refcount_t refcnt; /* Unlike other expressions, ops doesn't have static storage duration. * nft core assumes they do. We use kfree_rcu so that nft core can @@ -45,7 +45,7 @@ struct nft_xt_match_priv { static bool nft_xt_put(struct nft_xt *xt) { - if (--xt->refcnt == 0) { + if (refcount_dec_and_test(&xt->refcnt)) { list_del(&xt->head); kfree_rcu(xt, rcu_head); return true; @@ -273,7 +273,7 @@ nft_target_init(const struct nft_ctx *ctx, const struct nft_expr *expr, return -EINVAL; nft_xt = container_of(expr->ops, struct nft_xt, ops); - nft_xt->refcnt++; + refcount_inc(&nft_xt->refcnt); return 0; } @@ -486,7 +486,7 @@ __nft_match_init(const struct nft_ctx *ctx, const struct nft_expr *expr, return ret; nft_xt = container_of(expr->ops, struct nft_xt, ops); - nft_xt->refcnt++; + refcount_inc(&nft_xt->refcnt); return 0; } @@ -789,7 +789,7 @@ nft_match_select_ops(const struct nft_ctx *ctx, goto err; } - nft_match->refcnt = 0; + refcount_set(&nft_match->refcnt, 0); nft_match->ops.type = &nft_match_type; nft_match->ops.eval = nft_match_eval; nft_match->ops.init = nft_match_init; @@ -893,7 +893,7 @@ nft_target_select_ops(const struct nft_ctx *ctx, goto err; } - nft_target->refcnt = 0; + refcount_set(&nft_target->refcnt, 0); nft_target->ops.type = &nft_target_type; nft_target->ops.size = NFT_EXPR_SIZE(XT_ALIGN(target->targetsize)); nft_target->ops.init = nft_target_init; @@ -964,7 +964,7 @@ static void __exit nft_compat_module_exit(void) list_for_each_entry_safe(xt, next, &nft_target_list, head) { struct xt_target *target = xt->ops.data; - if (WARN_ON_ONCE(xt->refcnt)) + if (WARN_ON_ONCE(refcount_read(&xt->refcnt))) continue; module_put(target->me); kfree(xt); @@ -973,7 +973,7 @@ static void __exit nft_compat_module_exit(void) list_for_each_entry_safe(xt, next, &nft_match_list, head) { struct xt_match *match = xt->ops.data; - if (WARN_ON_ONCE(xt->refcnt)) + if (WARN_ON_ONCE(refcount_read(&xt->refcnt))) continue; module_put(match->me); kfree(xt); From cf52572ebbd7189a1966c2b5fc34b97078cd1dce Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Mon, 14 Jan 2019 14:28:49 +0100 Subject: [PATCH 28/79] netfilter: nft_compat: make lists per netns There are two problems with nft_compat since the netlink config plane uses a per-netns mutex: 1. Concurrent add/del accesses to the same list 2. accesses to a list element after it has been free'd already. This patch fixes the first problem. Freeing occurs from a work queue, after transaction mutexes have been released, i.e., it still possible for a new transaction (even from same net ns) to find the to-be-deleted expression in the list. The ->destroy functions are not allowed to have any such side effects, i.e. the list_del() in the destroy function is not allowed. This part of the problem is solved in the next patch. I tried to make this work by serializing list access via mutex and by moving list_del() to a deactivate callback, but Taehee spotted following race on this approach: NET #0 NET #1 >select_ops() ->init() ->select_ops() ->deactivate() ->destroy() nft_xt_put() kfree_rcu(xt, rcu_head); ->init() <-- use-after-free occurred. Unfortunately, we can't increment reference count in select_ops(), because we can't undo the refcount increase in case a different expression fails in the same batch. (The destroy hook will only be called in case the expression was initialized successfully). Fixes: f102d66b335a ("netfilter: nf_tables: use dedicated mutex to guard transactions") Reported-by: Taehee Yoo Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso --- net/netfilter/nft_compat.c | 129 +++++++++++++++++++++++++------------ 1 file changed, 89 insertions(+), 40 deletions(-) diff --git a/net/netfilter/nft_compat.c b/net/netfilter/nft_compat.c index acc85acad31b..abed3490a8f8 100644 --- a/net/netfilter/nft_compat.c +++ b/net/netfilter/nft_compat.c @@ -22,6 +22,7 @@ #include #include #include +#include struct nft_xt { struct list_head head; @@ -43,6 +44,20 @@ struct nft_xt_match_priv { void *info; }; +struct nft_compat_net { + struct list_head nft_target_list; + struct list_head nft_match_list; +}; + +static unsigned int nft_compat_net_id __read_mostly; +static struct nft_expr_type nft_match_type; +static struct nft_expr_type nft_target_type; + +static struct nft_compat_net *nft_compat_pernet(struct net *net) +{ + return net_generic(net, nft_compat_net_id); +} + static bool nft_xt_put(struct nft_xt *xt) { if (refcount_dec_and_test(&xt->refcnt)) { @@ -734,10 +749,6 @@ static const struct nfnetlink_subsystem nfnl_compat_subsys = { .cb = nfnl_nft_compat_cb, }; -static LIST_HEAD(nft_match_list); - -static struct nft_expr_type nft_match_type; - static bool nft_match_cmp(const struct xt_match *match, const char *name, u32 rev, u32 family) { @@ -749,6 +760,7 @@ static const struct nft_expr_ops * nft_match_select_ops(const struct nft_ctx *ctx, const struct nlattr * const tb[]) { + struct nft_compat_net *cn; struct nft_xt *nft_match; struct xt_match *match; unsigned int matchsize; @@ -765,8 +777,10 @@ nft_match_select_ops(const struct nft_ctx *ctx, rev = ntohl(nla_get_be32(tb[NFTA_MATCH_REV])); family = ctx->family; + cn = nft_compat_pernet(ctx->net); + /* Re-use the existing match if it's already loaded. */ - list_for_each_entry(nft_match, &nft_match_list, head) { + list_for_each_entry(nft_match, &cn->nft_match_list, head) { struct xt_match *match = nft_match->ops.data; if (nft_match_cmp(match, mt_name, rev, family)) @@ -810,7 +824,7 @@ nft_match_select_ops(const struct nft_ctx *ctx, nft_match->ops.size = matchsize; - list_add(&nft_match->head, &nft_match_list); + list_add(&nft_match->head, &cn->nft_match_list); return &nft_match->ops; err: @@ -826,10 +840,6 @@ static struct nft_expr_type nft_match_type __read_mostly = { .owner = THIS_MODULE, }; -static LIST_HEAD(nft_target_list); - -static struct nft_expr_type nft_target_type; - static bool nft_target_cmp(const struct xt_target *tg, const char *name, u32 rev, u32 family) { @@ -841,6 +851,7 @@ static const struct nft_expr_ops * nft_target_select_ops(const struct nft_ctx *ctx, const struct nlattr * const tb[]) { + struct nft_compat_net *cn; struct nft_xt *nft_target; struct xt_target *target; char *tg_name; @@ -861,8 +872,9 @@ nft_target_select_ops(const struct nft_ctx *ctx, strcmp(tg_name, "standard") == 0) return ERR_PTR(-EINVAL); + cn = nft_compat_pernet(ctx->net); /* Re-use the existing target if it's already loaded. */ - list_for_each_entry(nft_target, &nft_target_list, head) { + list_for_each_entry(nft_target, &cn->nft_target_list, head) { struct xt_target *target = nft_target->ops.data; if (!target->target) @@ -907,7 +919,7 @@ nft_target_select_ops(const struct nft_ctx *ctx, else nft_target->ops.eval = nft_target_eval_xt; - list_add(&nft_target->head, &nft_target_list); + list_add(&nft_target->head, &cn->nft_target_list); return &nft_target->ops; err: @@ -923,13 +935,74 @@ static struct nft_expr_type nft_target_type __read_mostly = { .owner = THIS_MODULE, }; +static int __net_init nft_compat_init_net(struct net *net) +{ + struct nft_compat_net *cn = nft_compat_pernet(net); + + INIT_LIST_HEAD(&cn->nft_target_list); + INIT_LIST_HEAD(&cn->nft_match_list); + + return 0; +} + +static void __net_exit nft_compat_exit_net(struct net *net) +{ + struct nft_compat_net *cn = nft_compat_pernet(net); + struct nft_xt *xt, *next; + + if (list_empty(&cn->nft_match_list) && + list_empty(&cn->nft_target_list)) + return; + + /* If there was an error that caused nft_xt expr to not be initialized + * fully and noone else requested the same expression later, the lists + * contain 0-refcount entries that still hold module reference. + * + * Clean them here. + */ + mutex_lock(&net->nft.commit_mutex); + list_for_each_entry_safe(xt, next, &cn->nft_target_list, head) { + struct xt_target *target = xt->ops.data; + + list_del_init(&xt->head); + + if (refcount_read(&xt->refcnt)) + continue; + module_put(target->me); + kfree(xt); + } + + list_for_each_entry_safe(xt, next, &cn->nft_match_list, head) { + struct xt_match *match = xt->ops.data; + + list_del_init(&xt->head); + + if (refcount_read(&xt->refcnt)) + continue; + module_put(match->me); + kfree(xt); + } + mutex_unlock(&net->nft.commit_mutex); +} + +static struct pernet_operations nft_compat_net_ops = { + .init = nft_compat_init_net, + .exit = nft_compat_exit_net, + .id = &nft_compat_net_id, + .size = sizeof(struct nft_compat_net), +}; + static int __init nft_compat_module_init(void) { int ret; + ret = register_pernet_subsys(&nft_compat_net_ops); + if (ret < 0) + goto err_target; + ret = nft_register_expr(&nft_match_type); if (ret < 0) - return ret; + goto err_pernet; ret = nft_register_expr(&nft_target_type); if (ret < 0) @@ -942,45 +1015,21 @@ static int __init nft_compat_module_init(void) } return ret; - err_target: nft_unregister_expr(&nft_target_type); err_match: nft_unregister_expr(&nft_match_type); +err_pernet: + unregister_pernet_subsys(&nft_compat_net_ops); return ret; } static void __exit nft_compat_module_exit(void) { - struct nft_xt *xt, *next; - - /* list should be empty here, it can be non-empty only in case there - * was an error that caused nft_xt expr to not be initialized fully - * and noone else requested the same expression later. - * - * In this case, the lists contain 0-refcount entries that still - * hold module reference. - */ - list_for_each_entry_safe(xt, next, &nft_target_list, head) { - struct xt_target *target = xt->ops.data; - - if (WARN_ON_ONCE(refcount_read(&xt->refcnt))) - continue; - module_put(target->me); - kfree(xt); - } - - list_for_each_entry_safe(xt, next, &nft_match_list, head) { - struct xt_match *match = xt->ops.data; - - if (WARN_ON_ONCE(refcount_read(&xt->refcnt))) - continue; - module_put(match->me); - kfree(xt); - } nfnetlink_subsys_unregister(&nfnl_compat_subsys); nft_unregister_expr(&nft_target_type); nft_unregister_expr(&nft_match_type); + unregister_pernet_subsys(&nft_compat_net_ops); } MODULE_ALIAS_NFNL_SUBSYS(NFNL_SUBSYS_NFT_COMPAT); From b2e3d68d1251a051a620f9086e18f7ffa6833b5b Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Mon, 14 Jan 2019 14:28:50 +0100 Subject: [PATCH 29/79] netfilter: nft_compat: destroy function must not have side effects The nft_compat destroy function deletes the nft_xt object from a list. This isn't allowed anymore. Destroy functions are called asynchronously, i.e. next batch can find the object that has a pending ->destroy() invocation: cpu0 cpu1 worker ->destroy for_each_entry() if (x == ... return x->ops; list_del(x) kfree_rcu(x) expr->ops->... // ops was free'd To resolve this, the list_del needs to occur before the transaction mutex gets released. nf_tables has a 'deactivate' hook for this purpose, so use that to unlink the object from the list. Fixes: 0935d5588400 ("netfilter: nf_tables: asynchronous release") Reported-by: Taehee Yoo Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso --- net/netfilter/nft_compat.c | 48 +++++++++++++++++++++++++++++++++++++- 1 file changed, 47 insertions(+), 1 deletion(-) diff --git a/net/netfilter/nft_compat.c b/net/netfilter/nft_compat.c index abed3490a8f8..5eb269428832 100644 --- a/net/netfilter/nft_compat.c +++ b/net/netfilter/nft_compat.c @@ -29,6 +29,9 @@ struct nft_xt { struct nft_expr_ops ops; refcount_t refcnt; + /* used only when transaction mutex is locked */ + unsigned int listcnt; + /* Unlike other expressions, ops doesn't have static storage duration. * nft core assumes they do. We use kfree_rcu so that nft core can * can check expr->ops->size even after nft_compat->destroy() frees @@ -61,7 +64,7 @@ static struct nft_compat_net *nft_compat_pernet(struct net *net) static bool nft_xt_put(struct nft_xt *xt) { if (refcount_dec_and_test(&xt->refcnt)) { - list_del(&xt->head); + WARN_ON_ONCE(!list_empty(&xt->head)); kfree_rcu(xt, rcu_head); return true; } @@ -555,6 +558,43 @@ nft_match_destroy(const struct nft_ctx *ctx, const struct nft_expr *expr) __nft_match_destroy(ctx, expr, nft_expr_priv(expr)); } +static void nft_compat_activate(const struct nft_ctx *ctx, + const struct nft_expr *expr, + struct list_head *h) +{ + struct nft_xt *xt = container_of(expr->ops, struct nft_xt, ops); + + if (xt->listcnt == 0) + list_add(&xt->head, h); + + xt->listcnt++; +} + +static void nft_compat_activate_mt(const struct nft_ctx *ctx, + const struct nft_expr *expr) +{ + struct nft_compat_net *cn = nft_compat_pernet(ctx->net); + + nft_compat_activate(ctx, expr, &cn->nft_match_list); +} + +static void nft_compat_activate_tg(const struct nft_ctx *ctx, + const struct nft_expr *expr) +{ + struct nft_compat_net *cn = nft_compat_pernet(ctx->net); + + nft_compat_activate(ctx, expr, &cn->nft_target_list); +} + +static void nft_compat_deactivate(const struct nft_ctx *ctx, + const struct nft_expr *expr) +{ + struct nft_xt *xt = container_of(expr->ops, struct nft_xt, ops); + + if (--xt->listcnt == 0) + list_del_init(&xt->head); +} + static void nft_match_large_destroy(const struct nft_ctx *ctx, const struct nft_expr *expr) { @@ -808,6 +848,8 @@ nft_match_select_ops(const struct nft_ctx *ctx, nft_match->ops.eval = nft_match_eval; nft_match->ops.init = nft_match_init; nft_match->ops.destroy = nft_match_destroy; + nft_match->ops.activate = nft_compat_activate_mt; + nft_match->ops.deactivate = nft_compat_deactivate; nft_match->ops.dump = nft_match_dump; nft_match->ops.validate = nft_match_validate; nft_match->ops.data = match; @@ -824,6 +866,7 @@ nft_match_select_ops(const struct nft_ctx *ctx, nft_match->ops.size = matchsize; + nft_match->listcnt = 1; list_add(&nft_match->head, &cn->nft_match_list); return &nft_match->ops; @@ -910,6 +953,8 @@ nft_target_select_ops(const struct nft_ctx *ctx, nft_target->ops.size = NFT_EXPR_SIZE(XT_ALIGN(target->targetsize)); nft_target->ops.init = nft_target_init; nft_target->ops.destroy = nft_target_destroy; + nft_target->ops.activate = nft_compat_activate_tg; + nft_target->ops.deactivate = nft_compat_deactivate; nft_target->ops.dump = nft_target_dump; nft_target->ops.validate = nft_target_validate; nft_target->ops.data = target; @@ -919,6 +964,7 @@ nft_target_select_ops(const struct nft_ctx *ctx, else nft_target->ops.eval = nft_target_eval_xt; + nft_target->listcnt = 1; list_add(&nft_target->head, &cn->nft_target_list); return &nft_target->ops; From 8208d1708b88b412ca97f50a6d951242c88cbbac Mon Sep 17 00:00:00 2001 From: Marc Zyngier Date: Fri, 18 Jan 2019 14:08:59 +0000 Subject: [PATCH 30/79] irqchip/gic-v3-its: Align PCI Multi-MSI allocation on their size The way we allocate events works fine in most cases, except when multiple PCI devices share an ITS-visible DevID, and that one of them is trying to use MultiMSI allocation. In that case, our allocation is not guaranteed to be zero-based anymore, and we have to make sure we allocate it on a boundary that is compatible with the PCI Multi-MSI constraints. Fix this by allocating the full region upfront instead of iterating over the number of MSIs. MSI-X are always allocated one by one, so this shouldn't change anything on that front. Fixes: b48ac83d6bbc2 ("irqchip: GICv3: ITS: MSI support") Cc: stable@vger.kernel.org Reported-by: Ard Biesheuvel Tested-by: Ard Biesheuvel Signed-off-by: Marc Zyngier --- drivers/irqchip/irq-gic-v3-its.c | 25 +++++++++++++------------ 1 file changed, 13 insertions(+), 12 deletions(-) diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c index db20e992a40f..7f2a45445b00 100644 --- a/drivers/irqchip/irq-gic-v3-its.c +++ b/drivers/irqchip/irq-gic-v3-its.c @@ -2399,13 +2399,14 @@ static void its_free_device(struct its_device *its_dev) kfree(its_dev); } -static int its_alloc_device_irq(struct its_device *dev, irq_hw_number_t *hwirq) +static int its_alloc_device_irq(struct its_device *dev, int nvecs, irq_hw_number_t *hwirq) { int idx; - idx = find_first_zero_bit(dev->event_map.lpi_map, - dev->event_map.nr_lpis); - if (idx == dev->event_map.nr_lpis) + idx = bitmap_find_free_region(dev->event_map.lpi_map, + dev->event_map.nr_lpis, + get_count_order(nvecs)); + if (idx < 0) return -ENOSPC; *hwirq = dev->event_map.lpi_base + idx; @@ -2501,21 +2502,21 @@ static int its_irq_domain_alloc(struct irq_domain *domain, unsigned int virq, int err; int i; - for (i = 0; i < nr_irqs; i++) { - err = its_alloc_device_irq(its_dev, &hwirq); - if (err) - return err; + err = its_alloc_device_irq(its_dev, nr_irqs, &hwirq); + if (err) + return err; - err = its_irq_gic_domain_alloc(domain, virq + i, hwirq); + for (i = 0; i < nr_irqs; i++) { + err = its_irq_gic_domain_alloc(domain, virq + i, hwirq + i); if (err) return err; irq_domain_set_hwirq_and_chip(domain, virq + i, - hwirq, &its_irq_chip, its_dev); + hwirq + i, &its_irq_chip, its_dev); irqd_set_single_target(irq_desc_get_irq_data(irq_to_desc(virq + i))); pr_debug("ID:%d pID:%d vID:%d\n", - (int)(hwirq - its_dev->event_map.lpi_base), - (int) hwirq, virq + i); + (int)(hwirq + i - its_dev->event_map.lpi_base), + (int)(hwirq + i), virq + i); } return 0; From 6dc080eeb2ba01973bfff0d79844d7a59e12542e Mon Sep 17 00:00:00 2001 From: Prateek Sood Date: Fri, 30 Nov 2018 20:40:56 +0530 Subject: [PATCH 31/79] sched/wait: Fix rcuwait_wake_up() ordering For some peculiar reason rcuwait_wake_up() has the right barrier in the comment, but not in the code. This mistake has been observed to cause a deadlock in the following situation: P1 P2 percpu_up_read() percpu_down_write() rcu_sync_is_idle() // false rcu_sync_enter() ... __percpu_up_read() [S] ,- __this_cpu_dec(*sem->read_count) | smp_rmb(); [L] | task = rcu_dereference(w->task) // NULL | | [S] w->task = current | smp_mb(); | [L] readers_active_check() // fail `-> Where the smp_rmb() (obviously) fails to constrain the store. [ peterz: Added changelog. ] Signed-off-by: Prateek Sood Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Andrea Parri Acked-by: Davidlohr Bueso Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Fixes: 8f95c90ceb54 ("sched/wait, RCU: Introduce rcuwait machinery") Link: https://lkml.kernel.org/r/1543590656-7157-1-git-send-email-prsood@codeaurora.org Signed-off-by: Ingo Molnar --- kernel/exit.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/exit.c b/kernel/exit.c index 284f2fe9a293..3fb7be001964 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -307,7 +307,7 @@ void rcuwait_wake_up(struct rcuwait *w) * MB (A) MB (B) * [L] cond [L] tsk */ - smp_rmb(); /* (B) */ + smp_mb(); /* (B) */ /* * Avoid using task_rcu_dereference() magic as long as we are careful, From e6018c0f5c996e61639adce6a0697391a2861916 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 17 Dec 2018 10:14:53 +0100 Subject: [PATCH 32/79] sched/wake_q: Document wake_q_add() The only guarantee provided by wake_q_add() is that a wakeup will happen after it, it does _NOT_ guarantee the wakeup will be delayed until the matching wake_up_q(). If wake_q_add() fails the cmpxchg() a concurrent wakeup is pending and that can happen at any time after the cmpxchg(). This means we should not rely on the wakeup happening at wake_q_up(), but should be ready for wake_q_add() to issue the wakeup. The delay; if provided (most likely); should only result in more efficient behaviour. Reported-by: Yongji Xie Signed-off-by: Peter Zijlstra (Intel) Cc: Davidlohr Bueso Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: Waiman Long Cc: Will Deacon Signed-off-by: Ingo Molnar --- include/linux/sched/wake_q.h | 6 +++++- kernel/sched/core.c | 12 ++++++++++++ 2 files changed, 17 insertions(+), 1 deletion(-) diff --git a/include/linux/sched/wake_q.h b/include/linux/sched/wake_q.h index 10b19a192b2d..545f37138057 100644 --- a/include/linux/sched/wake_q.h +++ b/include/linux/sched/wake_q.h @@ -24,9 +24,13 @@ * called near the end of a function. Otherwise, the list can be * re-initialized for later re-use by wake_q_init(). * - * Note that this can cause spurious wakeups. schedule() callers + * NOTE that this can cause spurious wakeups. schedule() callers * must ensure the call is done inside a loop, confirming that the * wakeup condition has in fact occurred. + * + * NOTE that there is no guarantee the wakeup will happen any later than the + * wake_q_add() location. Therefore task must be ready to be woken at the + * location of the wake_q_add(). */ #include diff --git a/kernel/sched/core.c b/kernel/sched/core.c index a674c7db2f29..cc814933f7d6 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -396,6 +396,18 @@ static bool set_nr_if_polling(struct task_struct *p) #endif #endif +/** + * wake_q_add() - queue a wakeup for 'later' waking. + * @head: the wake_q_head to add @task to + * @task: the task to queue for 'later' wakeup + * + * Queue a task for later wakeup, most likely by the wake_up_q() call in the + * same context, _HOWEVER_ this is not guaranteed, the wakeup can come + * instantly. + * + * This function must be used as-if it were wake_up_process(); IOW the task + * must be ready to be woken at this location. + */ void wake_q_add(struct wake_q_head *head, struct task_struct *task) { struct wake_q_node *node = &task->wake_q; From 4c4e3731564c8945ac5ac90fc2a1e1f21cb79c92 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 17 Dec 2018 10:14:53 +0100 Subject: [PATCH 33/79] sched/wake_q: Fix wakeup ordering for wake_q Notable cmpxchg() does not provide ordering when it fails, however wake_q_add() requires ordering in this specific case too. Without this it would be possible for the concurrent wakeup to not observe our prior state. Andrea Parri provided: C wake_up_q-wake_q_add { int next = 0; int y = 0; } P0(int *next, int *y) { int r0; /* in wake_up_q() */ WRITE_ONCE(*next, 1); /* node->next = NULL */ smp_mb(); /* implied by wake_up_process() */ r0 = READ_ONCE(*y); } P1(int *next, int *y) { int r1; /* in wake_q_add() */ WRITE_ONCE(*y, 1); /* wake_cond = true */ smp_mb__before_atomic(); r1 = cmpxchg_relaxed(next, 1, 2); } exists (0:r0=0 /\ 1:r1=0) This "exists" clause cannot be satisfied according to the LKMM: Test wake_up_q-wake_q_add Allowed States 3 0:r0=0; 1:r1=1; 0:r0=1; 1:r1=0; 0:r0=1; 1:r1=1; No Witnesses Positive: 0 Negative: 3 Condition exists (0:r0=0 /\ 1:r1=0) Observation wake_up_q-wake_q_add Never 0 3 Reported-by: Yongji Xie Signed-off-by: Peter Zijlstra (Intel) Cc: Davidlohr Bueso Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: Waiman Long Cc: Will Deacon Signed-off-by: Ingo Molnar --- kernel/sched/core.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/kernel/sched/core.c b/kernel/sched/core.c index cc814933f7d6..d8d76a65cfdd 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -417,10 +417,11 @@ void wake_q_add(struct wake_q_head *head, struct task_struct *task) * its already queued (either by us or someone else) and will get the * wakeup due to that. * - * This cmpxchg() executes a full barrier, which pairs with the full - * barrier executed by the wakeup in wake_up_q(). + * In order to ensure that a pending wakeup will observe our pending + * state, even in the failed case, an explicit smp_mb() must be used. */ - if (cmpxchg(&node->next, NULL, WAKE_Q_TAIL)) + smp_mb__before_atomic(); + if (cmpxchg_relaxed(&node->next, NULL, WAKE_Q_TAIL)) return; get_task_struct(task); From b061c38bef43406df8e73c5be06cbfacad5ee6ad Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 29 Nov 2018 14:44:49 +0100 Subject: [PATCH 34/79] futex: Fix (possible) missed wakeup We must not rely on wake_q_add() to delay the wakeup; in particular commit: 1d0dcb3ad9d3 ("futex: Implement lockless wakeups") moved wake_q_add() before smp_store_release(&q->lock_ptr, NULL), which could result in futex_wait() waking before observing ->lock_ptr == NULL and going back to sleep again. Signed-off-by: Peter Zijlstra (Intel) Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Fixes: 1d0dcb3ad9d3 ("futex: Implement lockless wakeups") Signed-off-by: Ingo Molnar --- kernel/futex.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) diff --git a/kernel/futex.c b/kernel/futex.c index be3bff2315ff..fdd312da0992 100644 --- a/kernel/futex.c +++ b/kernel/futex.c @@ -1452,11 +1452,7 @@ static void mark_wake_futex(struct wake_q_head *wake_q, struct futex_q *q) if (WARN(q->pi_state || q->rt_waiter, "refusing to wake PI futex\n")) return; - /* - * Queue the task for later wakeup for after we've released - * the hb->lock. wake_q_add() grabs reference to p. - */ - wake_q_add(wake_q, p); + get_task_struct(p); __unqueue_futex(q); /* * The waiting task can free the futex_q as soon as q->lock_ptr = NULL @@ -1466,6 +1462,13 @@ static void mark_wake_futex(struct wake_q_head *wake_q, struct futex_q *q) * plist_del in __unqueue_futex(). */ smp_store_release(&q->lock_ptr, NULL); + + /* + * Queue the task for later wakeup for after we've released + * the hb->lock. wake_q_add() grabs reference to p. + */ + wake_q_add(wake_q, p); + put_task_struct(p); } /* From e158488be27b157802753a59b336142dc0eb0380 Mon Sep 17 00:00:00 2001 From: Xie Yongji Date: Thu, 29 Nov 2018 20:50:30 +0800 Subject: [PATCH 35/79] locking/rwsem: Fix (possible) missed wakeup Because wake_q_add() can imply an immediate wakeup (cmpxchg failure case), we must not rely on the wakeup being delayed. However, commit: e38513905eea ("locking/rwsem: Rework zeroing reader waiter->task") relies on exactly that behaviour in that the wakeup must not happen until after we clear waiter->task. [ peterz: Added changelog. ] Signed-off-by: Xie Yongji Signed-off-by: Zhang Yu Signed-off-by: Peter Zijlstra (Intel) Cc: Linus Torvalds Cc: Peter Zijlstra Cc: Thomas Gleixner Fixes: e38513905eea ("locking/rwsem: Rework zeroing reader waiter->task") Link: https://lkml.kernel.org/r/1543495830-2644-1-git-send-email-xieyongji@baidu.com Signed-off-by: Ingo Molnar --- kernel/locking/rwsem-xadd.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) diff --git a/kernel/locking/rwsem-xadd.c b/kernel/locking/rwsem-xadd.c index 09b180063ee1..50d9af615dc4 100644 --- a/kernel/locking/rwsem-xadd.c +++ b/kernel/locking/rwsem-xadd.c @@ -198,15 +198,22 @@ static void __rwsem_mark_wake(struct rw_semaphore *sem, woken++; tsk = waiter->task; - wake_q_add(wake_q, tsk); + get_task_struct(tsk); list_del(&waiter->list); /* - * Ensure that the last operation is setting the reader + * Ensure calling get_task_struct() before setting the reader * waiter to nil such that rwsem_down_read_failed() cannot * race with do_exit() by always holding a reference count * to the task to wakeup. */ smp_store_release(&waiter->task, NULL); + /* + * Ensure issuing the wakeup (either by us or someone else) + * after setting the reader waiter to nil. + */ + wake_q_add(wake_q, tsk); + /* wake_q_add() already take the task ref */ + put_task_struct(tsk); } adjustment = woken * RWSEM_ACTIVE_READ_BIAS - adjustment; From 625210cfa6c0c26ea422f655bf68288176f174e6 Mon Sep 17 00:00:00 2001 From: Sinan Kaya Date: Mon, 21 Jan 2019 23:19:58 +0000 Subject: [PATCH 36/79] x86/Kconfig: Select PCI_LOCKLESS_CONFIG if PCI is enabled After commit 5d32a66541c4 ("PCI/ACPI: Allow ACPI to be built without CONFIG_PCI set") dependencies on CONFIG_PCI that previously were satisfied implicitly through dependencies on CONFIG_ACPI have to be specified directly. PCI_LOCKLESS_CONFIG depends on PCI but this dependency has not been mentioned in the Kconfig so add an explicit dependency here and fix WARNING: unmet direct dependencies detected for PCI_LOCKLESS_CONFIG Depends on [n]: PCI [=n] Selected by [y]: - X86 [=y] Fixes: 5d32a66541c46 ("PCI/ACPI: Allow ACPI to be built without CONFIG_PCI set") Signed-off-by: Sinan Kaya Signed-off-by: Borislav Petkov Cc: "H. Peter Anvin" Cc: Ingo Molnar Cc: Thomas Gleixner Cc: linux-acpi@vger.kernel.org Cc: x86-ml Link: https://lkml.kernel.org/r/20190121231958.28255-2-okaya@kernel.org --- arch/x86/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 4b4a7f32b68e..26387c7bf305 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -198,7 +198,7 @@ config X86 select IRQ_FORCED_THREADING select NEED_SG_DMA_LENGTH select PCI_DOMAINS if PCI - select PCI_LOCKLESS_CONFIG + select PCI_LOCKLESS_CONFIG if PCI select PERF_EVENTS select RTC_LIB select RTC_MC146818_LIB From 53ab60baa1ac4f20b080a22c13b77b6373922fd7 Mon Sep 17 00:00:00 2001 From: ZhangXiaoxu Date: Thu, 10 Jan 2019 16:39:06 +0800 Subject: [PATCH 37/79] ipvs: Fix signed integer overflow when setsockopt timeout There is a UBSAN bug report as below: UBSAN: Undefined behaviour in net/netfilter/ipvs/ip_vs_ctl.c:2227:21 signed integer overflow: -2147483647 * 1000 cannot be represented in type 'int' Reproduce program: #include #include #include #define IPPROTO_IP 0 #define IPPROTO_RAW 255 #define IP_VS_BASE_CTL (64+1024+64) #define IP_VS_SO_SET_TIMEOUT (IP_VS_BASE_CTL+10) /* The argument to IP_VS_SO_GET_TIMEOUT */ struct ipvs_timeout_t { int tcp_timeout; int tcp_fin_timeout; int udp_timeout; }; int main() { int ret = -1; int sockfd = -1; struct ipvs_timeout_t to; sockfd = socket(AF_INET, SOCK_RAW, IPPROTO_RAW); if (sockfd == -1) { printf("socket init error\n"); return -1; } to.tcp_timeout = -2147483647; to.tcp_fin_timeout = -2147483647; to.udp_timeout = -2147483647; ret = setsockopt(sockfd, IPPROTO_IP, IP_VS_SO_SET_TIMEOUT, (char *)(&to), sizeof(to)); printf("setsockopt return %d\n", ret); return ret; } Return -EINVAL if the timeout value is negative or max than 'INT_MAX / HZ'. Signed-off-by: ZhangXiaoxu Acked-by: Simon Horman Signed-off-by: Pablo Neira Ayuso --- net/netfilter/ipvs/ip_vs_ctl.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) diff --git a/net/netfilter/ipvs/ip_vs_ctl.c b/net/netfilter/ipvs/ip_vs_ctl.c index 432141f04af3..7d6318664eb2 100644 --- a/net/netfilter/ipvs/ip_vs_ctl.c +++ b/net/netfilter/ipvs/ip_vs_ctl.c @@ -2220,6 +2220,18 @@ static int ip_vs_set_timeout(struct netns_ipvs *ipvs, struct ip_vs_timeout_user u->tcp_fin_timeout, u->udp_timeout); +#ifdef CONFIG_IP_VS_PROTO_TCP + if (u->tcp_timeout < 0 || u->tcp_timeout > (INT_MAX / HZ) || + u->tcp_fin_timeout < 0 || u->tcp_fin_timeout > (INT_MAX / HZ)) { + return -EINVAL; + } +#endif + +#ifdef CONFIG_IP_VS_PROTO_UDP + if (u->udp_timeout < 0 || u->udp_timeout > (INT_MAX / HZ)) + return -EINVAL; +#endif + #ifdef CONFIG_IP_VS_PROTO_TCP if (u->tcp_timeout) { pd = ip_vs_proto_data_get(ipvs, IPPROTO_TCP); From 245b6c6558128327d330549b23d09594c46f58df Mon Sep 17 00:00:00 2001 From: Thor Thayer Date: Tue, 22 Jan 2019 11:48:04 -0600 Subject: [PATCH 38/79] EDAC, altera: Fix S10 persistent register offset Correct the persistent register offset where address and status are stored. Fixes: 08f08bfb7b4c ("EDAC, altera: Merge Stratix10 into the Arria10 SDRAM probe routine") Signed-off-by: Thor Thayer Signed-off-by: Borislav Petkov Cc: James Morse Cc: Mauro Carvalho Chehab Cc: devicetree@vger.kernel.org Cc: dinguyen@kernel.org Cc: linux-edac Cc: mark.rutland@arm.com Cc: robh+dt@kernel.org Cc: stable Link: https://lkml.kernel.org/r/1548179287-21760-2-git-send-email-thor.thayer@linux.intel.com --- drivers/edac/altera_edac.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/edac/altera_edac.h b/drivers/edac/altera_edac.h index 4213cb0bb2a7..f8664bac9fa8 100644 --- a/drivers/edac/altera_edac.h +++ b/drivers/edac/altera_edac.h @@ -295,8 +295,8 @@ struct altr_sdram_mc_data { #define S10_SYSMGR_ECC_INTSTAT_DERR_OFST 0xA0 /* Sticky registers for Uncorrected Errors */ -#define S10_SYSMGR_UE_VAL_OFST 0x120 -#define S10_SYSMGR_UE_ADDR_OFST 0x124 +#define S10_SYSMGR_UE_VAL_OFST 0x220 +#define S10_SYSMGR_UE_ADDR_OFST 0x224 #define S10_DDR0_IRQ_MASK BIT(16) From 4e046de0f50e04acd48eb373d6a9061ddf014e0c Mon Sep 17 00:00:00 2001 From: Bodong Wang Date: Sun, 13 Jan 2019 22:47:26 -0600 Subject: [PATCH 39/79] Revert "net/mlx5e: E-Switch, Initialize eswitch only if eswitch manager" This reverts commit 5f5991f36dce1e69dd8bd7495763eec2e28f08e7. With the original commit, eswitch instance will not be initialized for a function which is vport group manager but not eswitch manager such as host PF on SmartNIC (BlueField) card. This will result in a kernel crash when such a vport group manager is trying to access vports in its group. E.g, PF vport manager (not eswitch manager) tries to configure the MAC of its VF vport, a kernel trace will happen similar as bellow: BUG: unable to handle kernel NULL pointer dereference at 0000000000000000 ... RIP: 0010:mlx5_eswitch_get_vport_config+0xc/0x180 [mlx5_core] ... Fixes: 5f5991f36dce ("net/mlx5e: E-Switch, Initialize eswitch only if eswitch manager") Signed-off-by: Bodong Wang Reported-by: Yuval Avnery Reviewed-by: Daniel Jurgens Reviewed-by: Or Gerlitz Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c index a44ea7b85614..ab7a038c207c 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c @@ -1728,7 +1728,7 @@ int mlx5_eswitch_init(struct mlx5_core_dev *dev) int vport_num; int err; - if (!MLX5_ESWITCH_MANAGER(dev)) + if (!MLX5_VPORT_MANAGER(dev)) return 0; esw_info(dev, @@ -1797,7 +1797,7 @@ int mlx5_eswitch_init(struct mlx5_core_dev *dev) void mlx5_eswitch_cleanup(struct mlx5_eswitch *esw) { - if (!esw || !MLX5_ESWITCH_MANAGER(esw->dev)) + if (!esw || !MLX5_VPORT_MANAGER(esw->dev)) return; esw_info(esw->dev, "cleanup\n"); From 92b32772940b981240f6471a70c0fdbff23983ea Mon Sep 17 00:00:00 2001 From: Shay Agroskin Date: Sun, 6 Jan 2019 17:55:59 +0200 Subject: [PATCH 40/79] net/mlx5e: Fix wrong private flag usage causing checksum disable MLX5E_PFLAG_* definitions were changed from bitmask to enumerated values. However, in mlx5e_open_rq(), the proper API (MLX5E_GET_PFLAG macro) was not used to read the flag value of MLX5E_PFLAG_RX_NO_CSUM_COMPLETE. Fixed it. Fixes: 8ff57c18e9f6 ("net/mlx5e: Improve ethtool private-flags code structure") Signed-off-by: Shay Agroskin Reviewed-by: Tariq Toukan Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index 8cfd2ec7c0a2..01819e5c9975 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -950,7 +950,7 @@ static int mlx5e_open_rq(struct mlx5e_channel *c, if (params->rx_dim_enabled) __set_bit(MLX5E_RQ_STATE_AM, &c->rq.state); - if (params->pflags & MLX5E_PFLAG_RX_NO_CSUM_COMPLETE) + if (MLX5E_GET_PFLAG(params, MLX5E_PFLAG_RX_NO_CSUM_COMPLETE)) __set_bit(MLX5E_RQ_STATE_NO_CSUM_COMPLETE, &c->rq.state); return 0; From 33814e5d127e21f53b52e17b0722c1b57d4f4d29 Mon Sep 17 00:00:00 2001 From: Moni Shoua Date: Wed, 26 Dec 2018 19:21:21 +0200 Subject: [PATCH 41/79] net/mlx5: Take lock with IRQs disabled to avoid deadlock The lock in qp_table might be taken from process context or from interrupt context. This may lead to a deadlock unless it is taken with IRQs disabled. Discovered by lockdep ================================ WARNING: inconsistent lock state 4.20.0-rc6 -------------------------------- inconsistent {HARDIRQ-ON-W} -> {IN-HARDIRQ-W} python/12572 [HC1[1]:SC0[0]:HE0:SE1] takes: 00000000052a4df4 (&(&table->lock)->rlock#2){?.+.}, /0x50 [mlx5_core] {HARDIRQ-ON-W} state was registered at: _raw_spin_lock+0x33/0x70 mlx5_get_rsc+0x1a/0x50 [mlx5_core] mlx5_ib_eqe_pf_action+0x493/0x1be0 [mlx5_ib] process_one_work+0x90c/0x1820 worker_thread+0x87/0xbb0 kthread+0x320/0x3e0 ret_from_fork+0x24/0x30 irq event stamp: 103928 hardirqs last enabled at (103927): [] nk+0x1a/0x1c hardirqs last disabled at (103928): [] unk+0x1a/0x1c softirqs last enabled at (103924): [] tcp_sendmsg+0x31/0x40 softirqs last disabled at (103922): [] 80 other info that might help us debug this: Possible unsafe locking scenario: CPU0 ---- lock(&(&table->lock)->rlock#2); lock(&(&table->lock)->rlock#2); *** DEADLOCK *** Fixes: 032080ab43ac ("IB/mlx5: Lock QP during page fault handling") Signed-off-by: Moni Shoua Reviewed-by: Leon Romanovsky Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/qp.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/qp.c b/drivers/net/ethernet/mellanox/mlx5/core/qp.c index 388f205a497f..370ca94b6775 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/qp.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/qp.c @@ -44,14 +44,15 @@ static struct mlx5_core_rsc_common * mlx5_get_rsc(struct mlx5_qp_table *table, u32 rsn) { struct mlx5_core_rsc_common *common; + unsigned long flags; - spin_lock(&table->lock); + spin_lock_irqsave(&table->lock, flags); common = radix_tree_lookup(&table->tree, rsn); if (common) atomic_inc(&common->refcount); - spin_unlock(&table->lock); + spin_unlock_irqrestore(&table->lock, flags); return common; } From 9d2cbdc5d334967c35b5f58c7bf3208e17325647 Mon Sep 17 00:00:00 2001 From: Aya Levin Date: Mon, 24 Dec 2018 09:48:42 +0200 Subject: [PATCH 42/79] net/mlx5e: Allow MAC invalidation while spoofchk is ON Prior to this patch the driver prohibited spoof checking on invalid MAC. Now the user can set this configuration if it wishes to. This is required since libvirt might invalidate the VF Mac by setting it to zero, while spoofcheck is ON. Fixes: 1ab2068a4c66 ("net/mlx5: Implement vports admin state backup/restore") Signed-off-by: Aya Levin Reviewed-by: Eran Ben Elisha Signed-off-by: Saeed Mahameed --- .../net/ethernet/mellanox/mlx5/core/eswitch.c | 18 ++++++------------ 1 file changed, 6 insertions(+), 12 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c index ab7a038c207c..5b492b67f4e1 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c @@ -1134,13 +1134,6 @@ static int esw_vport_ingress_config(struct mlx5_eswitch *esw, int err = 0; u8 *smac_v; - if (vport->info.spoofchk && !is_valid_ether_addr(vport->info.mac)) { - mlx5_core_warn(esw->dev, - "vport[%d] configure ingress rules failed, illegal mac with spoofchk\n", - vport->vport); - return -EPERM; - } - esw_vport_cleanup_ingress_rules(esw, vport); if (!vport->info.vlan && !vport->info.qos && !vport->info.spoofchk) { @@ -1827,13 +1820,10 @@ int mlx5_eswitch_set_vport_mac(struct mlx5_eswitch *esw, mutex_lock(&esw->state_lock); evport = &esw->vports[vport]; - if (evport->info.spoofchk && !is_valid_ether_addr(mac)) { + if (evport->info.spoofchk && !is_valid_ether_addr(mac)) mlx5_core_warn(esw->dev, - "MAC invalidation is not allowed when spoofchk is on, vport(%d)\n", + "Set invalid MAC while spoofchk is on, vport(%d)\n", vport); - err = -EPERM; - goto unlock; - } err = mlx5_modify_nic_vport_mac_address(esw->dev, vport, mac); if (err) { @@ -1979,6 +1969,10 @@ int mlx5_eswitch_set_vport_spoofchk(struct mlx5_eswitch *esw, evport = &esw->vports[vport]; pschk = evport->info.spoofchk; evport->info.spoofchk = spoofchk; + if (pschk && !is_valid_ether_addr(evport->info.mac)) + mlx5_core_warn(esw->dev, + "Spoofchk in set while MAC is invalid, vport(%d)\n", + evport->vport); if (evport->enabled && esw->mode == SRIOV_LEGACY) err = esw_vport_ingress_config(esw, evport); if (err) From c12ecc2305648822970002871230979359edb2c0 Mon Sep 17 00:00:00 2001 From: Or Gerlitz Date: Wed, 25 Apr 2018 17:32:04 +0300 Subject: [PATCH 43/79] net/mlx5e: Move to use common phys port names for vport representors With VF LAG commit 491c37e49b48 "net/mlx5e: In case of LAG, one switch parent id is used for all representors", both uplinks and all the VFs (on both of them) get the same switchdev id. This cause the provisioning system method to identify the rep of a given VF from the parent PF PCI device using switchev id and physical port name to break, since VFm of PF0 will have the (id, name) as VFm of PF1. To fix that, we align to use the framework agreed upstream and set by nfp commit 168c478e107e "nfp: wire get_phys_port_name on representors": $ cat /sys/class/net/eth4_*/phys_port_name p0 pf0vf0 pf0vf1 Now, the names will be different, e.g. pf0vf0 vs. pf1vf0. Fixes: 491c37e49b48 ("net/mlx5e: In case of LAG, one switch parent id is used for all representors") Signed-off-by: Or Gerlitz Reported-by: Waleed Musa Reviewed-by: Roi Dayan Signed-off-by: Saeed Mahameed --- .../net/ethernet/mellanox/mlx5/core/en_rep.c | 12 +++++++++-- drivers/net/ethernet/mellanox/mlx5/core/lag.c | 21 +++++++++++++++++++ .../ethernet/mellanox/mlx5/core/mlx5_core.h | 2 ++ 3 files changed, 33 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index 04736212a21c..f75227222db3 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -1126,9 +1126,17 @@ static int mlx5e_rep_get_phys_port_name(struct net_device *dev, struct mlx5e_priv *priv = netdev_priv(dev); struct mlx5e_rep_priv *rpriv = priv->ppriv; struct mlx5_eswitch_rep *rep = rpriv->rep; - int ret; + int ret, pf_num; + + ret = mlx5_lag_get_pf_num(priv->mdev, &pf_num); + if (ret) + return ret; + + if (rep->vport == FDB_UPLINK_VPORT) + ret = snprintf(buf, len, "p%d", pf_num); + else + ret = snprintf(buf, len, "pf%dvf%d", pf_num, rep->vport - 1); - ret = snprintf(buf, len, "%d", rep->vport - 1); if (ret >= len) return -EOPNOTSUPP; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag.c index 3a6baed722d8..2d223385dc81 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lag.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lag.c @@ -616,6 +616,27 @@ void mlx5_lag_add(struct mlx5_core_dev *dev, struct net_device *netdev) } } +int mlx5_lag_get_pf_num(struct mlx5_core_dev *dev, int *pf_num) +{ + struct mlx5_lag *ldev; + int n; + + ldev = mlx5_lag_dev_get(dev); + if (!ldev) { + mlx5_core_warn(dev, "no lag device, can't get pf num\n"); + return -EINVAL; + } + + for (n = 0; n < MLX5_MAX_PORTS; n++) + if (ldev->pf[n].dev == dev) { + *pf_num = n; + return 0; + } + + mlx5_core_warn(dev, "wasn't able to locate pf in the lag device\n"); + return -EINVAL; +} + /* Must be called with intf_mutex held */ void mlx5_lag_remove(struct mlx5_core_dev *dev) { diff --git a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h index c68dcea5985b..5300b0b6d836 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/mlx5_core.h @@ -187,6 +187,8 @@ static inline int mlx5_lag_is_lacp_owner(struct mlx5_core_dev *dev) MLX5_CAP_GEN(dev, lag_master); } +int mlx5_lag_get_pf_num(struct mlx5_core_dev *dev, int *pf_num); + void mlx5_reload_interface(struct mlx5_core_dev *mdev, int protocol); void mlx5_lag_update(struct mlx5_core_dev *dev); From 6ce966fd2671899a48037abe7bf1df80a5adf029 Mon Sep 17 00:00:00 2001 From: Or Gerlitz Date: Tue, 1 Jan 2019 12:44:52 +0200 Subject: [PATCH 44/79] net/mlx5e: Unblock setting vid 0 for VFs through the uplink rep It turns out that libvirt uses 0-vid as a default if no vlan was set for the guest (which is the case for switchdev mode) and errs if we disallow that: error: Failed to start domain vm75 error: Cannot set interface MAC/vlanid to 6a:66:2d:48:92:c2/0 \ for ifname enp59s0f0 vf 0: Operation not supported So allow this in order not to break existing systems. Signed-off-by: Or Gerlitz Reported-by: Maor Dickman Reviewed-by: Gavi Teitz Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index f75227222db3..f2573c2d2b5c 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -1293,6 +1293,18 @@ static int mlx5e_uplink_rep_set_mac(struct net_device *netdev, void *addr) return 0; } +static int mlx5e_uplink_rep_set_vf_vlan(struct net_device *dev, int vf, u16 vlan, u8 qos, + __be16 vlan_proto) +{ + netdev_warn_once(dev, "legacy vf vlan setting isn't supported in switchdev mode\n"); + + if (vlan != 0) + return -EOPNOTSUPP; + + /* allow setting 0-vid for compatibility with libvirt */ + return 0; +} + static const struct switchdev_ops mlx5e_rep_switchdev_ops = { .switchdev_port_attr_get = mlx5e_attr_get, }; @@ -1327,6 +1339,7 @@ static const struct net_device_ops mlx5e_netdev_ops_uplink_rep = { .ndo_set_vf_rate = mlx5e_set_vf_rate, .ndo_get_vf_config = mlx5e_get_vf_config, .ndo_get_vf_stats = mlx5e_get_vf_stats, + .ndo_set_vf_vlan = mlx5e_uplink_rep_set_vf_vlan, }; bool mlx5e_eswitch_rep(struct net_device *netdev) From 947b7ac135b16aa60f9141ff72bd494eda0edb5e Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Sun, 27 Jan 2019 06:35:28 -0700 Subject: [PATCH 45/79] Revert "block: cover another queue enter recursion via BIO_QUEUE_ENTERED" We can't touch a bio after ->make_request_fn(), for all we know it could already have been completed by the time this function returns. This reverts commit 698cef173983b086977e633e46476e0f925ca01e. Reported-by: syzbot+4df6ca820108fd248943@syzkaller.appspotmail.com Signed-off-by: Jens Axboe --- block/blk-core.c | 11 ----------- block/blk-merge.c | 10 ++++++++++ 2 files changed, 10 insertions(+), 11 deletions(-) diff --git a/block/blk-core.c b/block/blk-core.c index 1ccec27d20c3..3c5f61ceeb67 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -1083,18 +1083,7 @@ blk_qc_t generic_make_request(struct bio *bio) /* Create a fresh bio_list for all subordinate requests */ bio_list_on_stack[1] = bio_list_on_stack[0]; bio_list_init(&bio_list_on_stack[0]); - - /* - * Since we're recursing into make_request here, ensure - * that we mark this bio as already having entered the queue. - * If not, and the queue is going away, we can get stuck - * forever on waiting for the queue reference to drop. But - * that will never happen, as we're already holding a - * reference to it. - */ - bio_set_flag(bio, BIO_QUEUE_ENTERED); ret = q->make_request_fn(q, bio); - bio_clear_flag(bio, BIO_QUEUE_ENTERED); /* sort new bios into those for a lower level * and those for the same level diff --git a/block/blk-merge.c b/block/blk-merge.c index d79a22f111d1..71e9ac03f621 100644 --- a/block/blk-merge.c +++ b/block/blk-merge.c @@ -272,6 +272,16 @@ void blk_queue_split(struct request_queue *q, struct bio **bio) /* there isn't chance to merge the splitted bio */ split->bi_opf |= REQ_NOMERGE; + /* + * Since we're recursing into make_request here, ensure + * that we mark this bio as already having entered the queue. + * If not, and the queue is going away, we can get stuck + * forever on waiting for the queue reference to drop. But + * that will never happen, as we're already holding a + * reference to it. + */ + bio_set_flag(*bio, BIO_QUEUE_ENTERED); + bio_chain(split, *bio); trace_block_split(q, split, (*bio)->bi_iter.bi_sector); generic_make_request(*bio); From 63346650c1a94a92be61a57416ac88c0a47c4327 Mon Sep 17 00:00:00 2001 From: Cong Wang Date: Thu, 24 Jan 2019 14:18:18 -0800 Subject: [PATCH 46/79] netrom: switch to sock timer API sk_reset_timer() and sk_stop_timer() properly handle sock refcnt for timer function. Switching to them could fix a refcounting bug reported by syzbot. Reported-and-tested-by: syzbot+defa700d16f1bd1b9a05@syzkaller.appspotmail.com Cc: Ralf Baechle Cc: linux-hams@vger.kernel.org Signed-off-by: Cong Wang Signed-off-by: David S. Miller --- net/netrom/nr_timer.c | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) diff --git a/net/netrom/nr_timer.c b/net/netrom/nr_timer.c index cbd51ed5a2d7..908e53ab47a4 100644 --- a/net/netrom/nr_timer.c +++ b/net/netrom/nr_timer.c @@ -52,21 +52,21 @@ void nr_start_t1timer(struct sock *sk) { struct nr_sock *nr = nr_sk(sk); - mod_timer(&nr->t1timer, jiffies + nr->t1); + sk_reset_timer(sk, &nr->t1timer, jiffies + nr->t1); } void nr_start_t2timer(struct sock *sk) { struct nr_sock *nr = nr_sk(sk); - mod_timer(&nr->t2timer, jiffies + nr->t2); + sk_reset_timer(sk, &nr->t2timer, jiffies + nr->t2); } void nr_start_t4timer(struct sock *sk) { struct nr_sock *nr = nr_sk(sk); - mod_timer(&nr->t4timer, jiffies + nr->t4); + sk_reset_timer(sk, &nr->t4timer, jiffies + nr->t4); } void nr_start_idletimer(struct sock *sk) @@ -74,37 +74,37 @@ void nr_start_idletimer(struct sock *sk) struct nr_sock *nr = nr_sk(sk); if (nr->idle > 0) - mod_timer(&nr->idletimer, jiffies + nr->idle); + sk_reset_timer(sk, &nr->idletimer, jiffies + nr->idle); } void nr_start_heartbeat(struct sock *sk) { - mod_timer(&sk->sk_timer, jiffies + 5 * HZ); + sk_reset_timer(sk, &sk->sk_timer, jiffies + 5 * HZ); } void nr_stop_t1timer(struct sock *sk) { - del_timer(&nr_sk(sk)->t1timer); + sk_stop_timer(sk, &nr_sk(sk)->t1timer); } void nr_stop_t2timer(struct sock *sk) { - del_timer(&nr_sk(sk)->t2timer); + sk_stop_timer(sk, &nr_sk(sk)->t2timer); } void nr_stop_t4timer(struct sock *sk) { - del_timer(&nr_sk(sk)->t4timer); + sk_stop_timer(sk, &nr_sk(sk)->t4timer); } void nr_stop_idletimer(struct sock *sk) { - del_timer(&nr_sk(sk)->idletimer); + sk_stop_timer(sk, &nr_sk(sk)->idletimer); } void nr_stop_heartbeat(struct sock *sk) { - del_timer(&sk->sk_timer); + sk_stop_timer(sk, &sk->sk_timer); } int nr_t1timer_running(struct sock *sk) From 6571ebce112a21ec9be68ef2f53b96fcd41fd81b Mon Sep 17 00:00:00 2001 From: Tomonori Sakita Date: Fri, 25 Jan 2019 11:02:22 +0900 Subject: [PATCH 47/79] net: altera_tse: fix msgdma_tx_completion on non-zero fill_level case If fill_level was not zero and status was not BUSY, result of "tx_prod - tx_cons - inuse" might be zero. Subtracting 1 unconditionally results invalid negative return value on this case. Make sure not to return an negative value. Signed-off-by: Tomonori Sakita Signed-off-by: Atsushi Nemoto Reviewed-by: Dalon L Westergreen Acked-by: Thor Thayer Signed-off-by: David S. Miller --- drivers/net/ethernet/altera/altera_msgdma.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/altera/altera_msgdma.c b/drivers/net/ethernet/altera/altera_msgdma.c index 0fb986ba3290..0ae723f75341 100644 --- a/drivers/net/ethernet/altera/altera_msgdma.c +++ b/drivers/net/ethernet/altera/altera_msgdma.c @@ -145,7 +145,8 @@ u32 msgdma_tx_completions(struct altera_tse_private *priv) & 0xffff; if (inuse) { /* Tx FIFO is not empty */ - ready = priv->tx_prod - priv->tx_cons - inuse - 1; + ready = max_t(int, + priv->tx_prod - priv->tx_cons - inuse - 1, 0); } else { /* Check for buffered last packet */ status = csrrd32(priv->tx_dma_csr, msgdma_csroffs(status)); From b0cf029234f9b18e10703ba5147f0389c382bccc Mon Sep 17 00:00:00 2001 From: Bernard Pidoux Date: Fri, 25 Jan 2019 11:46:40 +0100 Subject: [PATCH 48/79] net/rose: fix NULL ax25_cb kernel panic When an internally generated frame is handled by rose_xmit(), rose_route_frame() is called: if (!rose_route_frame(skb, NULL)) { dev_kfree_skb(skb); stats->tx_errors++; return NETDEV_TX_OK; } We have the same code sequence in Net/Rom where an internally generated frame is handled by nr_xmit() calling nr_route_frame(skb, NULL). However, in this function NULL argument is tested while it is not in rose_route_frame(). Then kernel panic occurs later on when calling ax25cmp() with a NULL ax25_cb argument as reported many times and recently with syzbot. We need to test if ax25 is NULL before using it. Testing: Built kernel with CONFIG_ROSE=y. Signed-off-by: Bernard Pidoux Acked-by: Dmitry Vyukov Reported-by: syzbot+1a2c456a1ea08fa5b5f7@syzkaller.appspotmail.com Cc: "David S. Miller" Cc: Ralf Baechle Cc: Bernard Pidoux Cc: linux-hams@vger.kernel.org Cc: netdev@vger.kernel.org Cc: linux-kernel@vger.kernel.org Signed-off-by: David S. Miller --- net/rose/rose_route.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/net/rose/rose_route.c b/net/rose/rose_route.c index 77e9f85a2c92..f2ff21d7df08 100644 --- a/net/rose/rose_route.c +++ b/net/rose/rose_route.c @@ -850,6 +850,7 @@ void rose_link_device_down(struct net_device *dev) /* * Route a frame to an appropriate AX.25 connection. + * A NULL ax25_cb indicates an internally generated frame. */ int rose_route_frame(struct sk_buff *skb, ax25_cb *ax25) { @@ -867,6 +868,10 @@ int rose_route_frame(struct sk_buff *skb, ax25_cb *ax25) if (skb->len < ROSE_MIN_LEN) return res; + + if (!ax25) + return rose_loopback_queue(skb, NULL); + frametype = skb->data[2]; lci = ((skb->data[0] << 8) & 0xF00) + ((skb->data[1] << 0) & 0x0FF); if (frametype == ROSE_CALL_REQUEST && From f17b5f06cb92ef2250513a1e154c47b78df07d40 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Sun, 27 Jan 2019 15:18:05 -0800 Subject: [PATCH 49/79] Linux 5.0-rc4 --- Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Makefile b/Makefile index f5b1d0d168e0..141653226f3c 100644 --- a/Makefile +++ b/Makefile @@ -2,7 +2,7 @@ VERSION = 5 PATCHLEVEL = 0 SUBLEVEL = 0 -EXTRAVERSION = -rc3 +EXTRAVERSION = -rc4 NAME = Shy Crocodile # *DOCUMENTATION* From 263c6d75f9a544a3c2f8f6a26de4f4808d8f59cf Mon Sep 17 00:00:00 2001 From: Yonglong Liu Date: Sat, 26 Jan 2019 17:18:25 +0800 Subject: [PATCH 50/79] net: hns: Fix for missing of_node_put() after of_parse_phandle() In hns enet driver, we use of_parse_handle() to get hold of the device node related to "ae-handle" but we have missed to put the node reference using of_node_put() after we are done using the node. This patch fixes it. Note: This problem is stated in Link: https://lkml.org/lkml/2018/12/22/217 Fixes: 48189d6aaf1e ("net: hns: enet specifies a reference to dsaf") Reported-by: Alexey Khoroshilov Signed-off-by: Yonglong Liu Signed-off-by: Peng Li Signed-off-by: David S. Miller --- drivers/net/ethernet/hisilicon/hns/hns_enet.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/drivers/net/ethernet/hisilicon/hns/hns_enet.c b/drivers/net/ethernet/hisilicon/hns/hns_enet.c index 5b33238c6680..60e7d7ae3787 100644 --- a/drivers/net/ethernet/hisilicon/hns/hns_enet.c +++ b/drivers/net/ethernet/hisilicon/hns/hns_enet.c @@ -2418,6 +2418,8 @@ static int hns_nic_dev_probe(struct platform_device *pdev) out_notify_fail: (void)cancel_work_sync(&priv->service_task); out_read_prop_fail: + /* safe for ACPI FW */ + of_node_put(to_of_node(priv->fwnode)); free_netdev(ndev); return ret; } @@ -2447,6 +2449,9 @@ static int hns_nic_dev_remove(struct platform_device *pdev) set_bit(NIC_STATE_REMOVING, &priv->state); (void)cancel_work_sync(&priv->service_task); + /* safe for ACPI FW */ + of_node_put(to_of_node(priv->fwnode)); + free_netdev(ndev); return 0; } From ed29ca8b9592562559c64d027fb5eb126e463e2c Mon Sep 17 00:00:00 2001 From: Yonglong Liu Date: Sat, 26 Jan 2019 17:18:26 +0800 Subject: [PATCH 51/79] net: hns: Restart autoneg need return failed when autoneg off The hns driver of earlier devices, when autoneg off, restart autoneg will return -EINVAL, so make the hns driver for the latest devices do the same. Signed-off-by: Yonglong Liu Signed-off-by: Peng Li Signed-off-by: David S. Miller --- drivers/net/ethernet/hisilicon/hns/hns_ethtool.c | 16 +++++++++------- 1 file changed, 9 insertions(+), 7 deletions(-) diff --git a/drivers/net/ethernet/hisilicon/hns/hns_ethtool.c b/drivers/net/ethernet/hisilicon/hns/hns_ethtool.c index 8e9b95871d30..ce15d2350db9 100644 --- a/drivers/net/ethernet/hisilicon/hns/hns_ethtool.c +++ b/drivers/net/ethernet/hisilicon/hns/hns_ethtool.c @@ -1157,16 +1157,18 @@ static int hns_get_regs_len(struct net_device *net_dev) */ static int hns_nic_nway_reset(struct net_device *netdev) { - int ret = 0; struct phy_device *phy = netdev->phydev; - if (netif_running(netdev)) { - /* if autoneg is disabled, don't restart auto-negotiation */ - if (phy && phy->autoneg == AUTONEG_ENABLE) - ret = genphy_restart_aneg(phy); - } + if (!netif_running(netdev)) + return 0; - return ret; + if (!phy) + return -EOPNOTSUPP; + + if (phy->autoneg != AUTONEG_ENABLE) + return -EINVAL; + + return genphy_restart_aneg(phy); } static u32 From cec8abba13e6a26729dfed41019720068eeeff2b Mon Sep 17 00:00:00 2001 From: Yonglong Liu Date: Sat, 26 Jan 2019 17:18:27 +0800 Subject: [PATCH 52/79] net: hns: Fix wrong read accesses via Clause 45 MDIO protocol When reading phy registers via Clause 45 MDIO protocol, after write address operation, the driver use another write address operation, so can not read the right value of any phy registers. This patch fixes it. Signed-off-by: Yonglong Liu Signed-off-by: Peng Li Signed-off-by: David S. Miller --- drivers/net/ethernet/hisilicon/hns_mdio.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/hisilicon/hns_mdio.c b/drivers/net/ethernet/hisilicon/hns_mdio.c index 017e08452d8c..baf5cc251f32 100644 --- a/drivers/net/ethernet/hisilicon/hns_mdio.c +++ b/drivers/net/ethernet/hisilicon/hns_mdio.c @@ -321,7 +321,7 @@ static int hns_mdio_read(struct mii_bus *bus, int phy_id, int regnum) } hns_mdio_cmd_write(mdio_dev, is_c45, - MDIO_C45_WRITE_ADDR, phy_id, devad); + MDIO_C45_READ, phy_id, devad); } /* Step 5: waitting for MDIO_COMMAND_REG 's mdio_start==0,*/ From c69c29a1a0a8f68cd87e98ba4a5a79fb8ef2a58c Mon Sep 17 00:00:00 2001 From: Alexey Khoroshilov Date: Sat, 26 Jan 2019 22:48:57 +0300 Subject: [PATCH 53/79] net: stmmac: dwmac-rk: fix error handling in rk_gmac_powerup() If phy_power_on() fails in rk_gmac_powerup(), clocks are left enabled. Found by Linux Driver Verification project (linuxtesting.org). Signed-off-by: Alexey Khoroshilov Signed-off-by: David S. Miller --- drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c index 7b923362ee55..3b174eae77c1 100644 --- a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c @@ -1342,8 +1342,10 @@ static int rk_gmac_powerup(struct rk_priv_data *bsp_priv) } ret = phy_power_on(bsp_priv, true); - if (ret) + if (ret) { + gmac_clk_enable(bsp_priv, false); return ret; + } pm_runtime_enable(dev); pm_runtime_get_sync(dev); From 50c2936634bcb1db78a8ca63249236810c11a80f Mon Sep 17 00:00:00 2001 From: Johannes Berg Date: Sat, 26 Jan 2019 21:12:19 +0100 Subject: [PATCH 54/79] decnet: fix DN_IFREQ_SIZE Digging through the ioctls with Al because of the previous patches, we found that on 64-bit decnet's dn_dev_ioctl() is wrong, because struct ifreq::ifr_ifru is actually 24 bytes (not 16 as expected from struct sockaddr) due to the ifru_map and ifru_settings members. Clearly, decnet expects the ioctl to be called with a struct like struct ifreq_dn { char ifr_name[IFNAMSIZ]; struct sockaddr_dn ifr_addr; }; since it does struct ifreq *ifr = ...; struct sockaddr_dn *sdn = (struct sockaddr_dn *)&ifr->ifr_addr; This means that DN_IFREQ_SIZE is too big for what it wants on 64-bit, as it is sizeof(struct ifreq) - sizeof(struct sockaddr) + sizeof(struct sockaddr_dn) This assumes that sizeof(struct sockaddr) is the size of ifr_ifru but that isn't true. Fix this to use offsetof(struct ifreq, ifr_ifru). This indeed doesn't really matter much - the result is that we copy in/out 8 bytes more than we should on 64-bit platforms. In case the "struct ifreq_dn" lands just on the end of a page though it might lead to faults. As far as I can tell, it has been like this forever, so it seems very likely that nobody cares. Signed-off-by: Johannes Berg Signed-off-by: David S. Miller --- net/decnet/dn_dev.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/decnet/dn_dev.c b/net/decnet/dn_dev.c index d0b3e69c6b39..0962f9201baa 100644 --- a/net/decnet/dn_dev.c +++ b/net/decnet/dn_dev.c @@ -56,7 +56,7 @@ #include #include -#define DN_IFREQ_SIZE (sizeof(struct ifreq) - sizeof(struct sockaddr) + sizeof(struct sockaddr_dn)) +#define DN_IFREQ_SIZE (offsetof(struct ifreq, ifr_ifru) + sizeof(struct sockaddr_dn)) static char dn_rt_all_end_mcast[ETH_ALEN] = {0xAB,0x00,0x00,0x04,0x00,0x00}; static char dn_rt_all_rt_mcast[ETH_ALEN] = {0xAB,0x00,0x00,0x03,0x00,0x00}; From 146820cc240f4389cf33481c058d9493aef95e25 Mon Sep 17 00:00:00 2001 From: Nir Dotan Date: Sun, 27 Jan 2019 09:26:22 +0200 Subject: [PATCH 55/79] ip6mr: Fix notifiers call on mroute_clean_tables() When the MC route socket is closed, mroute_clean_tables() is called to cleanup existing routes. Mistakenly notifiers call was put on the cleanup of the unresolved MC route entries cache. In a case where the MC socket closes before an unresolved route expires, the notifier call leads to a crash, caused by the driver trying to increment a non initialized refcount_t object [1] and then when handling is done, to decrement it [2]. This was detected by a test recently added in commit 6d4efada3b82 ("selftests: forwarding: Add multicast routing test"). Fix that by putting notifiers call on the resolved entries traversal, instead of on the unresolved entries traversal. [1] [ 245.748967] refcount_t: increment on 0; use-after-free. [ 245.754829] WARNING: CPU: 3 PID: 3223 at lib/refcount.c:153 refcount_inc_checked+0x2b/0x30 ... [ 245.802357] Hardware name: Mellanox Technologies Ltd. MSN2740/SA001237, BIOS 5.6.5 06/07/2016 [ 245.811873] RIP: 0010:refcount_inc_checked+0x2b/0x30 ... [ 245.907487] Call Trace: [ 245.910231] mlxsw_sp_router_fib_event.cold.181+0x42/0x47 [mlxsw_spectrum] [ 245.917913] notifier_call_chain+0x45/0x7 [ 245.922484] atomic_notifier_call_chain+0x15/0x20 [ 245.927729] call_fib_notifiers+0x15/0x30 [ 245.932205] mroute_clean_tables+0x372/0x3f [ 245.936971] ip6mr_sk_done+0xb1/0xc0 [ 245.940960] ip6_mroute_setsockopt+0x1da/0x5f0 ... [2] [ 246.128487] refcount_t: underflow; use-after-free. [ 246.133859] WARNING: CPU: 0 PID: 7 at lib/refcount.c:187 refcount_sub_and_test_checked+0x4c/0x60 [ 246.183521] Hardware name: Mellanox Technologies Ltd. MSN2740/SA001237, BIOS 5.6.5 06/07/2016 ... [ 246.193062] Workqueue: mlxsw_core_ordered mlxsw_sp_router_fibmr_event_work [mlxsw_spectrum] [ 246.202394] RIP: 0010:refcount_sub_and_test_checked+0x4c/0x60 ... [ 246.298889] Call Trace: [ 246.301617] refcount_dec_and_test_checked+0x11/0x20 [ 246.307170] mlxsw_sp_router_fibmr_event_work.cold.196+0x47/0x78 [mlxsw_spectrum] [ 246.315531] process_one_work+0x1fa/0x3f0 [ 246.320005] worker_thread+0x2f/0x3e0 [ 246.324083] kthread+0x118/0x130 [ 246.327683] ? wq_update_unbound_numa+0x1b0/0x1b0 [ 246.332926] ? kthread_park+0x80/0x80 [ 246.337013] ret_from_fork+0x1f/0x30 Fixes: 088aa3eec2ce ("ip6mr: Support fib notifications") Signed-off-by: Nir Dotan Reviewed-by: Ido Schimmel Signed-off-by: David S. Miller --- net/ipv6/ip6mr.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/net/ipv6/ip6mr.c b/net/ipv6/ip6mr.c index 30337b38274b..cc01aa3f2b5e 100644 --- a/net/ipv6/ip6mr.c +++ b/net/ipv6/ip6mr.c @@ -1516,6 +1516,9 @@ static void mroute_clean_tables(struct mr_table *mrt, bool all) continue; rhltable_remove(&mrt->mfc_hash, &c->mnode, ip6mr_rht_params); list_del_rcu(&c->list); + call_ip6mr_mfc_entry_notifiers(read_pnet(&mrt->net), + FIB_EVENT_ENTRY_DEL, + (struct mfc6_cache *)c, mrt->id); mr6_netlink_event(mrt, (struct mfc6_cache *)c, RTM_DELROUTE); mr_cache_put(c); } @@ -1524,10 +1527,6 @@ static void mroute_clean_tables(struct mr_table *mrt, bool all) spin_lock_bh(&mfc_unres_lock); list_for_each_entry_safe(c, tmp, &mrt->mfc_unres_queue, list) { list_del(&c->list); - call_ip6mr_mfc_entry_notifiers(read_pnet(&mrt->net), - FIB_EVENT_ENTRY_DEL, - (struct mfc6_cache *)c, - mrt->id); mr6_netlink_event(mrt, (struct mfc6_cache *)c, RTM_DELROUTE); ip6mr_destroy_unres(mrt, (struct mfc6_cache *)c); From 6fb6e6371f8c463020a41cc0ed1915e140219c3d Mon Sep 17 00:00:00 2001 From: Andrew Lunn Date: Sun, 27 Jan 2019 22:48:00 +0100 Subject: [PATCH 56/79] net: dsa: mv88e6xxx: Fix serdes irq setup going recursive Duec to a typo, mv88e6390_serdes_irq_setup() calls itself, rather than mv88e6390x_serdes_irq_setup(). It then blows the stack, and shortly after the machine blows up. Fixes: 2defda1f4b91 ("net: dsa: mv88e6xxx: Add support for SERDES on ports 2-8 for 6390X") Signed-off-by: Andrew Lunn Signed-off-by: David S. Miller --- drivers/net/dsa/mv88e6xxx/serdes.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/dsa/mv88e6xxx/serdes.c b/drivers/net/dsa/mv88e6xxx/serdes.c index 2caa8c8b4b55..1bfc5ff8d81d 100644 --- a/drivers/net/dsa/mv88e6xxx/serdes.c +++ b/drivers/net/dsa/mv88e6xxx/serdes.c @@ -664,7 +664,7 @@ int mv88e6390_serdes_irq_setup(struct mv88e6xxx_chip *chip, int port) if (port < 9) return 0; - return mv88e6390_serdes_irq_setup(chip, port); + return mv88e6390x_serdes_irq_setup(chip, port); } void mv88e6390x_serdes_irq_free(struct mv88e6xxx_chip *chip, int port) From 2035f3ff8eaa29cfb5c8e2160b0f6e85eeb21a95 Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Mon, 21 Jan 2019 21:54:36 +0100 Subject: [PATCH 57/79] netfilter: ebtables: compat: un-break 32bit setsockopt when no rules are present Unlike ip(6)tables ebtables only counts user-defined chains. The effect is that a 32bit ebtables binary on a 64bit kernel can do 'ebtables -N FOO' only after adding at least one rule, else the request fails with -EINVAL. This is a similar fix as done in 3f1e53abff84 ("netfilter: ebtables: don't attempt to allocate 0-sized compat array"). Fixes: 7d7d7e02111e9 ("netfilter: compat: reject huge allocation requests") Reported-by: Francesco Ruggeri Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso --- net/bridge/netfilter/ebtables.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) diff --git a/net/bridge/netfilter/ebtables.c b/net/bridge/netfilter/ebtables.c index 5e55cef0cec3..6693e209efe8 100644 --- a/net/bridge/netfilter/ebtables.c +++ b/net/bridge/netfilter/ebtables.c @@ -2293,9 +2293,12 @@ static int compat_do_replace(struct net *net, void __user *user, xt_compat_lock(NFPROTO_BRIDGE); - ret = xt_compat_init_offsets(NFPROTO_BRIDGE, tmp.nentries); - if (ret < 0) - goto out_unlock; + if (tmp.nentries) { + ret = xt_compat_init_offsets(NFPROTO_BRIDGE, tmp.nentries); + if (ret < 0) + goto out_unlock; + } + ret = compat_copy_entries(entries_tmp, tmp.entries_size, &state); if (ret < 0) goto out_unlock; From 1a6a0951fc009f6d9fe8ebea2d2417d80d54097b Mon Sep 17 00:00:00 2001 From: Fernando Fernandez Mancera Date: Mon, 21 Jan 2019 12:53:21 +0100 Subject: [PATCH 58/79] netfilter: nfnetlink_osf: add missing fmatch check When we check the tcp options of a packet and it doesn't match the current fingerprint, the tcp packet option pointer must be restored to its initial value in order to do the proper tcp options check for the next fingerprint. Here we can see an example. Assumming the following fingerprint base with two lines: S10:64:1:60:M*,S,T,N,W6: Linux:3.0::Linux 3.0 S20:64:1:60:M*,S,T,N,W7: Linux:4.19:arch:Linux 4.1 Where TCP options are the last field in the OS signature, all of them overlap except by the last one, ie. 'W6' versus 'W7'. In case a packet for Linux 4.19 kicks in, the osf finds no matching because the TCP options pointer is updated after checking for the TCP options in the first line. Therefore, reset pointer back to where it should be. Fixes: 11eeef41d5f6 ("netfilter: passive OS fingerprint xtables match") Signed-off-by: Fernando Fernandez Mancera Signed-off-by: Pablo Neira Ayuso --- net/netfilter/nfnetlink_osf.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/net/netfilter/nfnetlink_osf.c b/net/netfilter/nfnetlink_osf.c index 6f41dd74729d..1f1d90c1716b 100644 --- a/net/netfilter/nfnetlink_osf.c +++ b/net/netfilter/nfnetlink_osf.c @@ -66,6 +66,7 @@ static bool nf_osf_match_one(const struct sk_buff *skb, int ttl_check, struct nf_osf_hdr_ctx *ctx) { + const __u8 *optpinit = ctx->optp; unsigned int check_WSS = 0; int fmatch = FMATCH_WRONG; int foptsize, optnum; @@ -155,6 +156,9 @@ static bool nf_osf_match_one(const struct sk_buff *skb, } } + if (fmatch != FMATCH_OK) + ctx->optp = optpinit; + return fmatch == FMATCH_OK; } From 206b8cc514d7ff2b79dd2d5ad939adc7c493f07a Mon Sep 17 00:00:00 2001 From: Anders Roxell Date: Wed, 23 Jan 2019 12:48:11 +0100 Subject: [PATCH 59/79] netfilter: ipt_CLUSTERIP: fix warning unused variable cn MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When CONFIG_PROC_FS isn't set the variable cn isn't used. net/ipv4/netfilter/ipt_CLUSTERIP.c: In function ‘clusterip_net_exit’: net/ipv4/netfilter/ipt_CLUSTERIP.c:849:24: warning: unused variable ‘cn’ [-Wunused-variable] struct clusterip_net *cn = clusterip_pernet(net); ^~ Rework so the variable 'cn' is declared inside "#ifdef CONFIG_PROC_FS". Fixes: b12f7bad5ad3 ("netfilter: ipt_CLUSTERIP: remove wrong WARN_ON_ONCE in netns exit routine") Signed-off-by: Anders Roxell Signed-off-by: Pablo Neira Ayuso --- net/ipv4/netfilter/ipt_CLUSTERIP.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/ipv4/netfilter/ipt_CLUSTERIP.c b/net/ipv4/netfilter/ipt_CLUSTERIP.c index b61977db9b7f..2a909e5f9ba0 100644 --- a/net/ipv4/netfilter/ipt_CLUSTERIP.c +++ b/net/ipv4/netfilter/ipt_CLUSTERIP.c @@ -846,9 +846,9 @@ static int clusterip_net_init(struct net *net) static void clusterip_net_exit(struct net *net) { +#ifdef CONFIG_PROC_FS struct clusterip_net *cn = clusterip_pernet(net); -#ifdef CONFIG_PROC_FS mutex_lock(&cn->mutex); proc_remove(cn->procdir); cn->procdir = NULL; From 4aa9fc2a435abe95a1e8d7f8c7b3d6356514b37a Mon Sep 17 00:00:00 2001 From: Michal Hocko Date: Fri, 25 Jan 2019 19:08:58 +0100 Subject: [PATCH 60/79] Revert "mm, memory_hotplug: initialize struct pages for the full memory section" This reverts commit 2830bf6f05fb3e05bc4743274b806c821807a684. The underlying assumption that one sparse section belongs into a single numa node doesn't hold really. Robert Shteynfeld has reported a boot failure. The boot log was not captured but his memory layout is as follows: Early memory node ranges node 1: [mem 0x0000000000001000-0x0000000000090fff] node 1: [mem 0x0000000000100000-0x00000000dbdf8fff] node 1: [mem 0x0000000100000000-0x0000001423ffffff] node 0: [mem 0x0000001424000000-0x0000002023ffffff] This means that node0 starts in the middle of a memory section which is also in node1. memmap_init_zone tries to initialize padding of a section even when it is outside of the given pfn range because there are code paths (e.g. memory hotplug) which assume that the full worth of memory section is always initialized. In this particular case, though, such a range is already intialized and most likely already managed by the page allocator. Scribbling over those pages corrupts the internal state and likely blows up when any of those pages gets used. Reported-by: Robert Shteynfeld Fixes: 2830bf6f05fb ("mm, memory_hotplug: initialize struct pages for the full memory section") Cc: stable@kernel.org Signed-off-by: Michal Hocko Signed-off-by: Linus Torvalds --- mm/page_alloc.c | 12 ------------ 1 file changed, 12 deletions(-) diff --git a/mm/page_alloc.c b/mm/page_alloc.c index d295c9bc01a8..35fdde041f5c 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -5701,18 +5701,6 @@ void __meminit memmap_init_zone(unsigned long size, int nid, unsigned long zone, cond_resched(); } } -#ifdef CONFIG_SPARSEMEM - /* - * If the zone does not span the rest of the section then - * we should at least initialize those pages. Otherwise we - * could blow up on a poisoned page in some paths which depend - * on full sections being initialized (e.g. memory hotplug). - */ - while (end_pfn % PAGES_PER_SECTION) { - __init_single_page(pfn_to_page(end_pfn), end_pfn, zone, nid); - end_pfn++; - } -#endif } #ifdef CONFIG_ZONE_DEVICE From ca899324305d953c19abb583b96f0dd81274b0b2 Mon Sep 17 00:00:00 2001 From: Yang Wei Date: Mon, 28 Jan 2019 22:42:25 +0800 Subject: [PATCH 61/79] net: i825xx: replace dev_kfree_skb_irq by dev_consume_skb_irq for drop profiles dev_consume_skb_irq() should be called in i596_interrupt() when skb xmit done. It makes drop profiles(dropwatch, perf) more friendly. Signed-off-by: Yang Wei Signed-off-by: David S. Miller --- drivers/net/ethernet/i825xx/82596.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/i825xx/82596.c b/drivers/net/ethernet/i825xx/82596.c index d719668a6684..92929750f832 100644 --- a/drivers/net/ethernet/i825xx/82596.c +++ b/drivers/net/ethernet/i825xx/82596.c @@ -1310,7 +1310,7 @@ static irqreturn_t i596_interrupt(int irq, void *dev_id) dev->stats.tx_aborted_errors++; } - dev_kfree_skb_irq(skb); + dev_consume_skb_irq(skb); tx_cmd->cmd.command = 0; /* Mark free */ break; From 9e71a15d8b5bbce25c637f7f8833cd3f45b65646 Mon Sep 17 00:00:00 2001 From: Manish Chopra Date: Mon, 28 Jan 2019 10:05:04 -0800 Subject: [PATCH 62/79] qed: Fix bug in tx promiscuous mode settings When running tx switched traffic between VNICs created via a bridge(to which VFs are added), adapter drops the unicast packets in tx flow due to VNIC's ucast mac being unknown to it. But VF interfaces being in promiscuous mode should have caused adapter to accept all the unknown ucast packets. Later, it was found that driver doesn't really configure tx promiscuous mode settings to accept all unknown unicast macs. This patch fixes tx promiscuous mode settings to accept all unknown/unmatched unicast macs and works out the scenario. Signed-off-by: Manish Chopra Signed-off-by: Ariel Elior Signed-off-by: David S. Miller --- drivers/net/ethernet/qlogic/qed/qed_l2.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/qlogic/qed/qed_l2.c b/drivers/net/ethernet/qlogic/qed/qed_l2.c index 67c02ea93906..b8baa6fcef8e 100644 --- a/drivers/net/ethernet/qlogic/qed/qed_l2.c +++ b/drivers/net/ethernet/qlogic/qed/qed_l2.c @@ -609,6 +609,10 @@ qed_sp_update_accept_mode(struct qed_hwfn *p_hwfn, (!!(accept_filter & QED_ACCEPT_MCAST_MATCHED) && !!(accept_filter & QED_ACCEPT_MCAST_UNMATCHED))); + SET_FIELD(state, ETH_VPORT_TX_MODE_UCAST_ACCEPT_ALL, + (!!(accept_filter & QED_ACCEPT_UCAST_MATCHED) && + !!(accept_filter & QED_ACCEPT_UCAST_UNMATCHED))); + SET_FIELD(state, ETH_VPORT_TX_MODE_BCAST_ACCEPT_ALL, !!(accept_filter & QED_ACCEPT_BCAST)); @@ -2688,7 +2692,8 @@ static int qed_configure_filter_rx_mode(struct qed_dev *cdev, if (type == QED_FILTER_RX_MODE_TYPE_PROMISC) { accept_flags.rx_accept_filter |= QED_ACCEPT_UCAST_UNMATCHED | QED_ACCEPT_MCAST_UNMATCHED; - accept_flags.tx_accept_filter |= QED_ACCEPT_MCAST_UNMATCHED; + accept_flags.tx_accept_filter |= QED_ACCEPT_UCAST_UNMATCHED | + QED_ACCEPT_MCAST_UNMATCHED; } else if (type == QED_FILTER_RX_MODE_TYPE_MULTI_PROMISC) { accept_flags.rx_accept_filter |= QED_ACCEPT_MCAST_UNMATCHED; accept_flags.tx_accept_filter |= QED_ACCEPT_MCAST_UNMATCHED; From ff9296966e5e00b0d0d00477b2365a178f0f06a3 Mon Sep 17 00:00:00 2001 From: Manish Chopra Date: Mon, 28 Jan 2019 10:05:05 -0800 Subject: [PATCH 63/79] qed: Fix LACP pdu drops for VFs VF is always configured to drop control frames (with reserved mac addresses) but to work LACP on the VFs, it would require LACP control frames to be forwarded or transmitted successfully. This patch fixes this in such a way that trusted VFs (marked through ndo_set_vf_trust) would be allowed to pass the control frames such as LACP pdus. Signed-off-by: Manish Chopra Signed-off-by: Ariel Elior Signed-off-by: David S. Miller --- drivers/net/ethernet/qlogic/qed/qed_l2.c | 5 +++++ drivers/net/ethernet/qlogic/qed/qed_l2.h | 3 +++ drivers/net/ethernet/qlogic/qed/qed_sriov.c | 10 ++++++++-- 3 files changed, 16 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/qlogic/qed/qed_l2.c b/drivers/net/ethernet/qlogic/qed/qed_l2.c index b8baa6fcef8e..e68ca83ae915 100644 --- a/drivers/net/ethernet/qlogic/qed/qed_l2.c +++ b/drivers/net/ethernet/qlogic/qed/qed_l2.c @@ -748,6 +748,11 @@ int qed_sp_vport_update(struct qed_hwfn *p_hwfn, return rc; } + if (p_params->update_ctl_frame_check) { + p_cmn->ctl_frame_mac_check_en = p_params->mac_chk_en; + p_cmn->ctl_frame_ethtype_check_en = p_params->ethtype_chk_en; + } + /* Update mcast bins for VFs, PF doesn't use this functionality */ qed_sp_update_mcast_bin(p_hwfn, p_ramrod, p_params); diff --git a/drivers/net/ethernet/qlogic/qed/qed_l2.h b/drivers/net/ethernet/qlogic/qed/qed_l2.h index 8d80f1095d17..7127d5aaac42 100644 --- a/drivers/net/ethernet/qlogic/qed/qed_l2.h +++ b/drivers/net/ethernet/qlogic/qed/qed_l2.h @@ -219,6 +219,9 @@ struct qed_sp_vport_update_params { struct qed_rss_params *rss_params; struct qed_filter_accept_flags accept_flags; struct qed_sge_tpa_params *sge_tpa_params; + u8 update_ctl_frame_check; + u8 mac_chk_en; + u8 ethtype_chk_en; }; int qed_sp_vport_update(struct qed_hwfn *p_hwfn, diff --git a/drivers/net/ethernet/qlogic/qed/qed_sriov.c b/drivers/net/ethernet/qlogic/qed/qed_sriov.c index ca6290fa0f30..71a7af134dd8 100644 --- a/drivers/net/ethernet/qlogic/qed/qed_sriov.c +++ b/drivers/net/ethernet/qlogic/qed/qed_sriov.c @@ -1969,7 +1969,9 @@ static void qed_iov_vf_mbx_start_vport(struct qed_hwfn *p_hwfn, params.vport_id = vf->vport_id; params.max_buffers_per_cqe = start->max_buffers_per_cqe; params.mtu = vf->mtu; - params.check_mac = true; + + /* Non trusted VFs should enable control frame filtering */ + params.check_mac = !vf->p_vf_info.is_trusted_configured; rc = qed_sp_eth_vport_start(p_hwfn, ¶ms); if (rc) { @@ -5130,6 +5132,9 @@ static void qed_iov_handle_trust_change(struct qed_hwfn *hwfn) params.opaque_fid = vf->opaque_fid; params.vport_id = vf->vport_id; + params.update_ctl_frame_check = 1; + params.mac_chk_en = !vf_info->is_trusted_configured; + if (vf_info->rx_accept_mode & mask) { flags->update_rx_mode_config = 1; flags->rx_accept_filter = vf_info->rx_accept_mode; @@ -5147,7 +5152,8 @@ static void qed_iov_handle_trust_change(struct qed_hwfn *hwfn) } if (flags->update_rx_mode_config || - flags->update_tx_mode_config) + flags->update_tx_mode_config || + params.update_ctl_frame_check) qed_sp_vport_update(hwfn, ¶ms, QED_SPQ_MODE_EBLOCK, NULL); } From 327852ec64205bb651be391a069784872098a3b2 Mon Sep 17 00:00:00 2001 From: Manish Chopra Date: Mon, 28 Jan 2019 10:05:06 -0800 Subject: [PATCH 64/79] qed: Fix VF probe failure while FLR VFs may hit VF-PF channel timeout while probing, as in some cases it was observed that VF FLR and VF "acquire" message transaction (i.e first message from VF to PF in VF's probe flow) could occur simultaneously which could lead VF to fail sending "acquire" message to PF as VF is marked disabled from HW perspective due to FLR, which will result into channel timeout and VF probe failure. In such cases, try retrying VF "acquire" message so that in later attempts it could be successful to pass message to PF after the VF FLR is completed and can be probed successfully. Signed-off-by: Manish Chopra Signed-off-by: Ariel Elior Signed-off-by: David S. Miller --- drivers/net/ethernet/qlogic/qed/qed_vf.c | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/drivers/net/ethernet/qlogic/qed/qed_vf.c b/drivers/net/ethernet/qlogic/qed/qed_vf.c index b6cccf44bf40..5dda547772c1 100644 --- a/drivers/net/ethernet/qlogic/qed/qed_vf.c +++ b/drivers/net/ethernet/qlogic/qed/qed_vf.c @@ -261,6 +261,7 @@ static int qed_vf_pf_acquire(struct qed_hwfn *p_hwfn) struct pfvf_acquire_resp_tlv *resp = &p_iov->pf2vf_reply->acquire_resp; struct pf_vf_pfdev_info *pfdev_info = &resp->pfdev_info; struct vf_pf_resc_request *p_resc; + u8 retry_cnt = VF_ACQUIRE_THRESH; bool resources_acquired = false; struct vfpf_acquire_tlv *req; int rc = 0, attempts = 0; @@ -314,6 +315,15 @@ static int qed_vf_pf_acquire(struct qed_hwfn *p_hwfn) /* send acquire request */ rc = qed_send_msg2pf(p_hwfn, &resp->hdr.status, sizeof(*resp)); + + /* Re-try acquire in case of vf-pf hw channel timeout */ + if (retry_cnt && rc == -EBUSY) { + DP_VERBOSE(p_hwfn, QED_MSG_IOV, + "VF retrying to acquire due to VPC timeout\n"); + retry_cnt--; + continue; + } + if (rc) goto exit; From 7c81626a3c37e4ac320b8ad785694ba498f24794 Mon Sep 17 00:00:00 2001 From: Manish Chopra Date: Mon, 28 Jan 2019 10:05:07 -0800 Subject: [PATCH 65/79] qed: Fix system crash in ll2 xmit Cache number of fragments in the skb locally as in case of linear skb (with zero fragments), tx completion (or freeing of skb) may happen before driver tries to get number of frgaments from the skb which could lead to stale access to an already freed skb. Signed-off-by: Manish Chopra Signed-off-by: Ariel Elior Signed-off-by: David S. Miller --- drivers/net/ethernet/qlogic/qed/qed_ll2.c | 20 +++++++++++++++----- 1 file changed, 15 insertions(+), 5 deletions(-) diff --git a/drivers/net/ethernet/qlogic/qed/qed_ll2.c b/drivers/net/ethernet/qlogic/qed/qed_ll2.c index d9237c65a838..b5f419b71287 100644 --- a/drivers/net/ethernet/qlogic/qed/qed_ll2.c +++ b/drivers/net/ethernet/qlogic/qed/qed_ll2.c @@ -2451,19 +2451,24 @@ static int qed_ll2_start_xmit(struct qed_dev *cdev, struct sk_buff *skb, { struct qed_ll2_tx_pkt_info pkt; const skb_frag_t *frag; + u8 flags = 0, nr_frags; int rc = -EINVAL, i; dma_addr_t mapping; u16 vlan = 0; - u8 flags = 0; if (unlikely(skb->ip_summed != CHECKSUM_NONE)) { DP_INFO(cdev, "Cannot transmit a checksummed packet\n"); return -EINVAL; } - if (1 + skb_shinfo(skb)->nr_frags > CORE_LL2_TX_MAX_BDS_PER_PACKET) { + /* Cache number of fragments from SKB since SKB may be freed by + * the completion routine after calling qed_ll2_prepare_tx_packet() + */ + nr_frags = skb_shinfo(skb)->nr_frags; + + if (1 + nr_frags > CORE_LL2_TX_MAX_BDS_PER_PACKET) { DP_ERR(cdev, "Cannot transmit a packet with %d fragments\n", - 1 + skb_shinfo(skb)->nr_frags); + 1 + nr_frags); return -EINVAL; } @@ -2485,7 +2490,7 @@ static int qed_ll2_start_xmit(struct qed_dev *cdev, struct sk_buff *skb, } memset(&pkt, 0, sizeof(pkt)); - pkt.num_of_bds = 1 + skb_shinfo(skb)->nr_frags; + pkt.num_of_bds = 1 + nr_frags; pkt.vlan = vlan; pkt.bd_flags = flags; pkt.tx_dest = QED_LL2_TX_DEST_NW; @@ -2496,12 +2501,17 @@ static int qed_ll2_start_xmit(struct qed_dev *cdev, struct sk_buff *skb, test_bit(QED_LL2_XMIT_FLAGS_FIP_DISCOVERY, &xmit_flags)) pkt.remove_stag = true; + /* qed_ll2_prepare_tx_packet() may actually send the packet if + * there are no fragments in the skb and subsequently the completion + * routine may run and free the SKB, so no dereferencing the SKB + * beyond this point unless skb has any fragments. + */ rc = qed_ll2_prepare_tx_packet(&cdev->hwfns[0], cdev->ll2->handle, &pkt, 1); if (rc) goto err; - for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) { + for (i = 0; i < nr_frags; i++) { frag = &skb_shinfo(skb)->frags[i]; mapping = skb_frag_dma_map(&cdev->pdev->dev, frag, 0, From ffb057f98928aa099b08e419bbe5afc26ec9f448 Mon Sep 17 00:00:00 2001 From: Manish Chopra Date: Mon, 28 Jan 2019 10:05:08 -0800 Subject: [PATCH 66/79] qed: Fix stack out of bounds bug KASAN reported following bug in qed_init_qm_get_idx_from_flags due to inappropriate casting of "pq_flags". Fix the type of "pq_flags". [ 196.624707] BUG: KASAN: stack-out-of-bounds in qed_init_qm_get_idx_from_flags+0x1a4/0x1b8 [qed] [ 196.624712] Read of size 8 at addr ffff809b00bc7360 by task kworker/0:9/1712 [ 196.624714] [ 196.624720] CPU: 0 PID: 1712 Comm: kworker/0:9 Not tainted 4.18.0-60.el8.aarch64+debug #1 [ 196.624723] Hardware name: To be filled by O.E.M. Saber/Saber, BIOS 0ACKL024 09/26/2018 [ 196.624733] Workqueue: events work_for_cpu_fn [ 196.624738] Call trace: [ 196.624742] dump_backtrace+0x0/0x2f8 [ 196.624745] show_stack+0x24/0x30 [ 196.624749] dump_stack+0xe0/0x11c [ 196.624755] print_address_description+0x68/0x260 [ 196.624759] kasan_report+0x178/0x340 [ 196.624762] __asan_report_load_n_noabort+0x38/0x48 [ 196.624786] qed_init_qm_get_idx_from_flags+0x1a4/0x1b8 [qed] [ 196.624808] qed_init_qm_info+0xec0/0x2200 [qed] [ 196.624830] qed_resc_alloc+0x284/0x7e8 [qed] [ 196.624853] qed_slowpath_start+0x6cc/0x1ae8 [qed] [ 196.624864] __qede_probe.isra.10+0x1cc/0x12c0 [qede] [ 196.624874] qede_probe+0x78/0xf0 [qede] [ 196.624879] local_pci_probe+0xc4/0x180 [ 196.624882] work_for_cpu_fn+0x54/0x98 [ 196.624885] process_one_work+0x758/0x1900 [ 196.624888] worker_thread+0x4e0/0xd18 [ 196.624892] kthread+0x2c8/0x350 [ 196.624897] ret_from_fork+0x10/0x18 [ 196.624899] [ 196.624902] Allocated by task 2: [ 196.624906] kasan_kmalloc.part.1+0x40/0x108 [ 196.624909] kasan_kmalloc+0xb4/0xc8 [ 196.624913] kasan_slab_alloc+0x14/0x20 [ 196.624916] kmem_cache_alloc_node+0x1dc/0x480 [ 196.624921] copy_process.isra.1.part.2+0x1d8/0x4a98 [ 196.624924] _do_fork+0x150/0xfa0 [ 196.624926] kernel_thread+0x48/0x58 [ 196.624930] kthreadd+0x3a4/0x5a0 [ 196.624932] ret_from_fork+0x10/0x18 [ 196.624934] [ 196.624937] Freed by task 0: [ 196.624938] (stack is not available) [ 196.624940] [ 196.624943] The buggy address belongs to the object at ffff809b00bc0000 [ 196.624943] which belongs to the cache thread_stack of size 32768 [ 196.624946] The buggy address is located 29536 bytes inside of [ 196.624946] 32768-byte region [ffff809b00bc0000, ffff809b00bc8000) [ 196.624948] The buggy address belongs to the page: [ 196.624952] page:ffff7fe026c02e00 count:1 mapcount:0 mapping:ffff809b4001c000 index:0x0 compound_mapcount: 0 [ 196.624960] flags: 0xfffff8000008100(slab|head) [ 196.624967] raw: 0fffff8000008100 dead000000000100 dead000000000200 ffff809b4001c000 [ 196.624970] raw: 0000000000000000 0000000000080008 00000001ffffffff 0000000000000000 [ 196.624973] page dumped because: kasan: bad access detected [ 196.624974] [ 196.624976] Memory state around the buggy address: [ 196.624980] ffff809b00bc7200: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 [ 196.624983] ffff809b00bc7280: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 [ 196.624985] >ffff809b00bc7300: 00 00 00 00 00 00 00 00 f1 f1 f1 f1 04 f2 f2 f2 [ 196.624988] ^ [ 196.624990] ffff809b00bc7380: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 [ 196.624993] ffff809b00bc7400: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 [ 196.624995] ================================================================== Signed-off-by: Manish Chopra Signed-off-by: Ariel Elior Signed-off-by: David S. Miller --- drivers/net/ethernet/qlogic/qed/qed_dev.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/net/ethernet/qlogic/qed/qed_dev.c b/drivers/net/ethernet/qlogic/qed/qed_dev.c index 8f6551421945..2ecaaaa4469a 100644 --- a/drivers/net/ethernet/qlogic/qed/qed_dev.c +++ b/drivers/net/ethernet/qlogic/qed/qed_dev.c @@ -795,19 +795,19 @@ static void qed_init_qm_pq(struct qed_hwfn *p_hwfn, /* get pq index according to PQ_FLAGS */ static u16 *qed_init_qm_get_idx_from_flags(struct qed_hwfn *p_hwfn, - u32 pq_flags) + unsigned long pq_flags) { struct qed_qm_info *qm_info = &p_hwfn->qm_info; /* Can't have multiple flags set here */ - if (bitmap_weight((unsigned long *)&pq_flags, + if (bitmap_weight(&pq_flags, sizeof(pq_flags) * BITS_PER_BYTE) > 1) { - DP_ERR(p_hwfn, "requested multiple pq flags 0x%x\n", pq_flags); + DP_ERR(p_hwfn, "requested multiple pq flags 0x%lx\n", pq_flags); goto err; } if (!(qed_get_pq_flags(p_hwfn) & pq_flags)) { - DP_ERR(p_hwfn, "pq flag 0x%x is not set\n", pq_flags); + DP_ERR(p_hwfn, "pq flag 0x%lx is not set\n", pq_flags); goto err; } From b46a0bf78ad7b150ef5910da83859f7f5a514ffd Mon Sep 17 00:00:00 2001 From: Jason Wang Date: Mon, 28 Jan 2019 15:05:05 +0800 Subject: [PATCH 67/79] vhost: fix OOB in get_rx_bufs() After batched used ring updating was introduced in commit e2b3b35eb989 ("vhost_net: batch used ring update in rx"). We tend to batch heads in vq->heads for more than one packet. But the quota passed to get_rx_bufs() was not correctly limited, which can result a OOB write in vq->heads. headcount = get_rx_bufs(vq, vq->heads + nvq->done_idx, vhost_len, &in, vq_log, &log, likely(mergeable) ? UIO_MAXIOV : 1); UIO_MAXIOV was still used which is wrong since we could have batched used in vq->heads, this will cause OOB if the next buffer needs more than 960 (1024 (UIO_MAXIOV) - 64 (VHOST_NET_BATCH)) heads after we've batched 64 (VHOST_NET_BATCH) heads: Acked-by: Stefan Hajnoczi ============================================================================= BUG kmalloc-8k (Tainted: G B ): Redzone overwritten ----------------------------------------------------------------------------- INFO: 0x00000000fd93b7a2-0x00000000f0713384. First byte 0xa9 instead of 0xcc INFO: Allocated in alloc_pd+0x22/0x60 age=3933677 cpu=2 pid=2674 kmem_cache_alloc_trace+0xbb/0x140 alloc_pd+0x22/0x60 gen8_ppgtt_create+0x11d/0x5f0 i915_ppgtt_create+0x16/0x80 i915_gem_create_context+0x248/0x390 i915_gem_context_create_ioctl+0x4b/0xe0 drm_ioctl_kernel+0xa5/0xf0 drm_ioctl+0x2ed/0x3a0 do_vfs_ioctl+0x9f/0x620 ksys_ioctl+0x6b/0x80 __x64_sys_ioctl+0x11/0x20 do_syscall_64+0x43/0xf0 entry_SYSCALL_64_after_hwframe+0x44/0xa9 INFO: Slab 0x00000000d13e87af objects=3 used=3 fp=0x (null) flags=0x200000000010201 INFO: Object 0x0000000003278802 @offset=17064 fp=0x00000000e2e6652b Fixing this by allocating UIO_MAXIOV + VHOST_NET_BATCH iovs for vhost-net. This is done through set the limitation through vhost_dev_init(), then set_owner can allocate the number of iov in a per device manner. This fixes CVE-2018-16880. Fixes: e2b3b35eb989 ("vhost_net: batch used ring update in rx") Signed-off-by: Jason Wang Signed-off-by: David S. Miller --- drivers/vhost/net.c | 3 ++- drivers/vhost/scsi.c | 2 +- drivers/vhost/vhost.c | 7 ++++--- drivers/vhost/vhost.h | 4 +++- drivers/vhost/vsock.c | 2 +- 5 files changed, 11 insertions(+), 7 deletions(-) diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c index bca86bf7189f..df51a35cf537 100644 --- a/drivers/vhost/net.c +++ b/drivers/vhost/net.c @@ -1337,7 +1337,8 @@ static int vhost_net_open(struct inode *inode, struct file *f) n->vqs[i].rx_ring = NULL; vhost_net_buf_init(&n->vqs[i].rxq); } - vhost_dev_init(dev, vqs, VHOST_NET_VQ_MAX); + vhost_dev_init(dev, vqs, VHOST_NET_VQ_MAX, + UIO_MAXIOV + VHOST_NET_BATCH); vhost_poll_init(n->poll + VHOST_NET_VQ_TX, handle_tx_net, EPOLLOUT, dev); vhost_poll_init(n->poll + VHOST_NET_VQ_RX, handle_rx_net, EPOLLIN, dev); diff --git a/drivers/vhost/scsi.c b/drivers/vhost/scsi.c index 344684f3e2e4..23593cb23dd0 100644 --- a/drivers/vhost/scsi.c +++ b/drivers/vhost/scsi.c @@ -1627,7 +1627,7 @@ static int vhost_scsi_open(struct inode *inode, struct file *f) vqs[i] = &vs->vqs[i].vq; vs->vqs[i].vq.handle_kick = vhost_scsi_handle_kick; } - vhost_dev_init(&vs->dev, vqs, VHOST_SCSI_MAX_VQ); + vhost_dev_init(&vs->dev, vqs, VHOST_SCSI_MAX_VQ, UIO_MAXIOV); vhost_scsi_init_inflight(vs, NULL); diff --git a/drivers/vhost/vhost.c b/drivers/vhost/vhost.c index 15a216cdd507..24a129fcdd61 100644 --- a/drivers/vhost/vhost.c +++ b/drivers/vhost/vhost.c @@ -390,9 +390,9 @@ static long vhost_dev_alloc_iovecs(struct vhost_dev *dev) vq->indirect = kmalloc_array(UIO_MAXIOV, sizeof(*vq->indirect), GFP_KERNEL); - vq->log = kmalloc_array(UIO_MAXIOV, sizeof(*vq->log), + vq->log = kmalloc_array(dev->iov_limit, sizeof(*vq->log), GFP_KERNEL); - vq->heads = kmalloc_array(UIO_MAXIOV, sizeof(*vq->heads), + vq->heads = kmalloc_array(dev->iov_limit, sizeof(*vq->heads), GFP_KERNEL); if (!vq->indirect || !vq->log || !vq->heads) goto err_nomem; @@ -414,7 +414,7 @@ static void vhost_dev_free_iovecs(struct vhost_dev *dev) } void vhost_dev_init(struct vhost_dev *dev, - struct vhost_virtqueue **vqs, int nvqs) + struct vhost_virtqueue **vqs, int nvqs, int iov_limit) { struct vhost_virtqueue *vq; int i; @@ -427,6 +427,7 @@ void vhost_dev_init(struct vhost_dev *dev, dev->iotlb = NULL; dev->mm = NULL; dev->worker = NULL; + dev->iov_limit = iov_limit; init_llist_head(&dev->work_list); init_waitqueue_head(&dev->wait); INIT_LIST_HEAD(&dev->read_list); diff --git a/drivers/vhost/vhost.h b/drivers/vhost/vhost.h index 1b675dad5e05..9490e7ddb340 100644 --- a/drivers/vhost/vhost.h +++ b/drivers/vhost/vhost.h @@ -170,9 +170,11 @@ struct vhost_dev { struct list_head read_list; struct list_head pending_list; wait_queue_head_t wait; + int iov_limit; }; -void vhost_dev_init(struct vhost_dev *, struct vhost_virtqueue **vqs, int nvqs); +void vhost_dev_init(struct vhost_dev *, struct vhost_virtqueue **vqs, + int nvqs, int iov_limit); long vhost_dev_set_owner(struct vhost_dev *dev); bool vhost_dev_has_owner(struct vhost_dev *dev); long vhost_dev_check_owner(struct vhost_dev *); diff --git a/drivers/vhost/vsock.c b/drivers/vhost/vsock.c index 3fbc068eaa9b..bb5fc0e9fbc2 100644 --- a/drivers/vhost/vsock.c +++ b/drivers/vhost/vsock.c @@ -531,7 +531,7 @@ static int vhost_vsock_dev_open(struct inode *inode, struct file *file) vsock->vqs[VSOCK_VQ_TX].handle_kick = vhost_vsock_handle_tx_kick; vsock->vqs[VSOCK_VQ_RX].handle_kick = vhost_vsock_handle_rx_kick; - vhost_dev_init(&vsock->dev, vqs, ARRAY_SIZE(vsock->vqs)); + vhost_dev_init(&vsock->dev, vqs, ARRAY_SIZE(vsock->vqs), UIO_MAXIOV); file->private_data = vsock; spin_lock_init(&vsock->send_pkt_list_lock); From 32eb67b93c9e3cd62cb423e30b090cdd4aa8d275 Mon Sep 17 00:00:00 2001 From: Dave Watson Date: Sun, 27 Jan 2019 00:57:38 +0000 Subject: [PATCH 68/79] net: tls: Save iv in tls_rec for async crypto requests aead_request_set_crypt takes an iv pointer, and we change the iv soon after setting it. Some async crypto algorithms don't save the iv, so we need to save it in the tls_rec for async requests. Found by hardcoding x64 aesni to use async crypto manager (to test the async codepath), however I don't think this combination can happen in the wild. Presumably other hardware offloads will need this fix, but there have been no user reports. Fixes: a42055e8d2c30 ("Add support for async encryption of records...") Signed-off-by: Dave Watson Signed-off-by: David S. Miller --- include/net/tls.h | 2 ++ net/tls/tls_sw.c | 4 +++- 2 files changed, 5 insertions(+), 1 deletion(-) diff --git a/include/net/tls.h b/include/net/tls.h index 2a6ac8d642af..1486b60c4de8 100644 --- a/include/net/tls.h +++ b/include/net/tls.h @@ -120,6 +120,8 @@ struct tls_rec { struct scatterlist sg_aead_out[2]; char aad_space[TLS_AAD_SPACE_SIZE]; + u8 iv_data[TLS_CIPHER_AES_GCM_128_IV_SIZE + + TLS_CIPHER_AES_GCM_128_SALT_SIZE]; struct aead_request aead_req; u8 aead_req_ctx[]; }; diff --git a/net/tls/tls_sw.c b/net/tls/tls_sw.c index 11cdc8f7db63..7e963560edef 100644 --- a/net/tls/tls_sw.c +++ b/net/tls/tls_sw.c @@ -439,6 +439,8 @@ static int tls_do_encryption(struct sock *sk, struct scatterlist *sge = sk_msg_elem(msg_en, start); int rc; + memcpy(rec->iv_data, tls_ctx->tx.iv, sizeof(rec->iv_data)); + sge->offset += tls_ctx->tx.prepend_size; sge->length -= tls_ctx->tx.prepend_size; @@ -448,7 +450,7 @@ static int tls_do_encryption(struct sock *sk, aead_request_set_ad(aead_req, TLS_AAD_SPACE_SIZE); aead_request_set_crypt(aead_req, rec->sg_aead_in, rec->sg_aead_out, - data_len, tls_ctx->tx.iv); + data_len, rec->iv_data); aead_request_set_callback(aead_req, CRYPTO_TFM_REQ_MAY_BACKLOG, tls_encrypt_done, sk); From 1023121375c6b0b3dc00334983c762ba2b76cb19 Mon Sep 17 00:00:00 2001 From: Dave Watson Date: Sun, 27 Jan 2019 00:59:03 +0000 Subject: [PATCH 69/79] net: tls: Fix deadlock in free_resources tx If there are outstanding async tx requests (when crypto returns EINPROGRESS), there is a potential deadlock: the tx work acquires the lock, while we cancel_delayed_work_sync() while holding the lock. Drop the lock while waiting for the work to complete. Fixes: a42055e8d2c30 ("Add support for async encryption of records...") Signed-off-by: Dave Watson Signed-off-by: David S. Miller --- net/tls/tls_sw.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/net/tls/tls_sw.c b/net/tls/tls_sw.c index 7e963560edef..bf5b54b513bc 100644 --- a/net/tls/tls_sw.c +++ b/net/tls/tls_sw.c @@ -1794,7 +1794,9 @@ void tls_sw_free_resources_tx(struct sock *sk) if (atomic_read(&ctx->encrypt_pending)) crypto_wait_req(-EINPROGRESS, &ctx->async_wait); + release_sock(sk); cancel_delayed_work_sync(&ctx->tx_work.work); + lock_sock(sk); /* Tx whatever records we can transmit and abandon the rest */ tls_tx_records(sk, -1); From f48af114895b2f9cdc218ed75666ade83380255b Mon Sep 17 00:00:00 2001 From: Yang Wei Date: Sun, 27 Jan 2019 23:56:34 +0800 Subject: [PATCH 70/79] net: alteon: replace dev_kfree_skb_irq by dev_consume_skb_irq dev_consume_skb_irq() should be called in ace_tx_int() when xmit done. It makes drop profiles more friendly. Signed-off-by: Yang Wei Signed-off-by: David S. Miller --- drivers/net/ethernet/alteon/acenic.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/alteon/acenic.c b/drivers/net/ethernet/alteon/acenic.c index 4f11f98347ed..1827ef1f6d55 100644 --- a/drivers/net/ethernet/alteon/acenic.c +++ b/drivers/net/ethernet/alteon/acenic.c @@ -2059,7 +2059,7 @@ static inline void ace_tx_int(struct net_device *dev, if (skb) { dev->stats.tx_packets++; dev->stats.tx_bytes += skb->len; - dev_kfree_skb_irq(skb); + dev_consume_skb_irq(skb); info->skb = NULL; } From 3afa73dd7c8b586d9d01a96063d537e1550d1287 Mon Sep 17 00:00:00 2001 From: Yang Wei Date: Sun, 27 Jan 2019 23:58:25 +0800 Subject: [PATCH 71/79] net: amd8111e: replace dev_kfree_skb_irq by dev_consume_skb_irq dev_consume_skb_irq() should be called in amd8111e_tx() when xmit done. It makes drop profiles more friendly. Signed-off-by: Yang Wei Signed-off-by: David S. Miller --- drivers/net/ethernet/amd/amd8111e.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/amd/amd8111e.c b/drivers/net/ethernet/amd/amd8111e.c index a90080f12e67..e548c0ae2e00 100644 --- a/drivers/net/ethernet/amd/amd8111e.c +++ b/drivers/net/ethernet/amd/amd8111e.c @@ -666,7 +666,7 @@ static int amd8111e_tx(struct net_device *dev) pci_unmap_single(lp->pci_dev, lp->tx_dma_addr[tx_index], lp->tx_skbuff[tx_index]->len, PCI_DMA_TODEVICE); - dev_kfree_skb_irq (lp->tx_skbuff[tx_index]); + dev_consume_skb_irq(lp->tx_skbuff[tx_index]); lp->tx_skbuff[tx_index] = NULL; lp->tx_dma_addr[tx_index] = 0; } From 100091156c1a4da11c0764265b9e328d7858f2bb Mon Sep 17 00:00:00 2001 From: Yang Wei Date: Tue, 29 Jan 2019 07:39:13 +0800 Subject: [PATCH 72/79] net: apple: replace dev_kfree_skb_irq by dev_consume_skb_irq for drop profiles dev_consume_skb_irq() should be called in bmac_txdma_intr() when xmit done. It makes drop profiles more friendly. Signed-off-by: Yang Wei Signed-off-by: David S. Miller --- drivers/net/ethernet/apple/bmac.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/apple/bmac.c b/drivers/net/ethernet/apple/bmac.c index 6a8e2567f2bd..4d3855ceb500 100644 --- a/drivers/net/ethernet/apple/bmac.c +++ b/drivers/net/ethernet/apple/bmac.c @@ -777,7 +777,7 @@ static irqreturn_t bmac_txdma_intr(int irq, void *dev_id) if (bp->tx_bufs[bp->tx_empty]) { ++dev->stats.tx_packets; - dev_kfree_skb_irq(bp->tx_bufs[bp->tx_empty]); + dev_consume_skb_irq(bp->tx_bufs[bp->tx_empty]); } bp->tx_bufs[bp->tx_empty] = NULL; bp->tx_fullup = 0; From b3379a424dde97755ad7a6f4405648a9ea13d6e3 Mon Sep 17 00:00:00 2001 From: Yang Wei Date: Tue, 29 Jan 2019 07:40:10 +0800 Subject: [PATCH 73/79] net: ti: replace dev_kfree_skb_irq by dev_consume_skb_irq for drop profiles dev_consume_skb_irq() should be called in cpmac_end_xmit() when xmit done. It makes drop profiles more friendly. Signed-off-by: Yang Wei Signed-off-by: David S. Miller --- drivers/net/ethernet/ti/cpmac.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/ti/cpmac.c b/drivers/net/ethernet/ti/cpmac.c index 810dfc7de1f9..e2d47b24a869 100644 --- a/drivers/net/ethernet/ti/cpmac.c +++ b/drivers/net/ethernet/ti/cpmac.c @@ -608,7 +608,7 @@ static void cpmac_end_xmit(struct net_device *dev, int queue) netdev_dbg(dev, "sent 0x%p, len=%d\n", desc->skb, desc->skb->len); - dev_kfree_skb_irq(desc->skb); + dev_consume_skb_irq(desc->skb); desc->skb = NULL; if (__netif_subqueue_stopped(dev, queue)) netif_wake_subqueue(dev, queue); From e501070e4db0b67a4c17a5557d1e9d098f3db310 Mon Sep 17 00:00:00 2001 From: Harini Katakam Date: Tue, 29 Jan 2019 15:20:03 +0530 Subject: [PATCH 74/79] net: macb: Apply RXUBR workaround only to versions with errata The interrupt handler contains a workaround for RX hang applicable to Zynq and AT91RM9200 only. Subsequent versions do not need this workaround. This workaround unnecessarily resets RX whenever RX used bit read is observed, which can be often under heavy traffic. There is no other action performed on RX UBR interrupt. Hence introduce a CAPS mask; enable this interrupt and workaround only on affected versions. Signed-off-by: Harini Katakam Signed-off-by: David S. Miller --- drivers/net/ethernet/cadence/macb.h | 3 +++ drivers/net/ethernet/cadence/macb_main.c | 28 ++++++++++++++---------- 2 files changed, 20 insertions(+), 11 deletions(-) diff --git a/drivers/net/ethernet/cadence/macb.h b/drivers/net/ethernet/cadence/macb.h index 3d45f4c92cf6..9bbaad9f3d63 100644 --- a/drivers/net/ethernet/cadence/macb.h +++ b/drivers/net/ethernet/cadence/macb.h @@ -643,6 +643,7 @@ #define MACB_CAPS_JUMBO 0x00000020 #define MACB_CAPS_GEM_HAS_PTP 0x00000040 #define MACB_CAPS_BD_RD_PREFETCH 0x00000080 +#define MACB_CAPS_NEEDS_RSTONUBR 0x00000100 #define MACB_CAPS_FIFO_MODE 0x10000000 #define MACB_CAPS_GIGABIT_MODE_AVAILABLE 0x20000000 #define MACB_CAPS_SG_DISABLED 0x40000000 @@ -1214,6 +1215,8 @@ struct macb { int rx_bd_rd_prefetch; int tx_bd_rd_prefetch; + + u32 rx_intr_mask; }; #ifdef CONFIG_MACB_USE_HWSTAMP diff --git a/drivers/net/ethernet/cadence/macb_main.c b/drivers/net/ethernet/cadence/macb_main.c index 66cc7927061a..2b2882615e8b 100644 --- a/drivers/net/ethernet/cadence/macb_main.c +++ b/drivers/net/ethernet/cadence/macb_main.c @@ -56,8 +56,7 @@ /* level of occupied TX descriptors under which we wake up TX process */ #define MACB_TX_WAKEUP_THRESH(bp) (3 * (bp)->tx_ring_size / 4) -#define MACB_RX_INT_FLAGS (MACB_BIT(RCOMP) | MACB_BIT(RXUBR) \ - | MACB_BIT(ISR_ROVR)) +#define MACB_RX_INT_FLAGS (MACB_BIT(RCOMP) | MACB_BIT(ISR_ROVR)) #define MACB_TX_ERR_FLAGS (MACB_BIT(ISR_TUND) \ | MACB_BIT(ISR_RLE) \ | MACB_BIT(TXERR)) @@ -1270,7 +1269,7 @@ static int macb_poll(struct napi_struct *napi, int budget) queue_writel(queue, ISR, MACB_BIT(RCOMP)); napi_reschedule(napi); } else { - queue_writel(queue, IER, MACB_RX_INT_FLAGS); + queue_writel(queue, IER, bp->rx_intr_mask); } } @@ -1288,7 +1287,7 @@ static void macb_hresp_error_task(unsigned long data) u32 ctrl; for (q = 0, queue = bp->queues; q < bp->num_queues; ++q, ++queue) { - queue_writel(queue, IDR, MACB_RX_INT_FLAGS | + queue_writel(queue, IDR, bp->rx_intr_mask | MACB_TX_INT_FLAGS | MACB_BIT(HRESP)); } @@ -1318,7 +1317,7 @@ static void macb_hresp_error_task(unsigned long data) /* Enable interrupts */ queue_writel(queue, IER, - MACB_RX_INT_FLAGS | + bp->rx_intr_mask | MACB_TX_INT_FLAGS | MACB_BIT(HRESP)); } @@ -1372,14 +1371,14 @@ static irqreturn_t macb_interrupt(int irq, void *dev_id) (unsigned int)(queue - bp->queues), (unsigned long)status); - if (status & MACB_RX_INT_FLAGS) { + if (status & bp->rx_intr_mask) { /* There's no point taking any more interrupts * until we have processed the buffers. The * scheduling call may fail if the poll routine * is already scheduled, so disable interrupts * now. */ - queue_writel(queue, IDR, MACB_RX_INT_FLAGS); + queue_writel(queue, IDR, bp->rx_intr_mask); if (bp->caps & MACB_CAPS_ISR_CLEAR_ON_WRITE) queue_writel(queue, ISR, MACB_BIT(RCOMP)); @@ -1412,8 +1411,9 @@ static irqreturn_t macb_interrupt(int irq, void *dev_id) /* There is a hardware issue under heavy load where DMA can * stop, this causes endless "used buffer descriptor read" * interrupts but it can be cleared by re-enabling RX. See - * the at91 manual, section 41.3.1 or the Zynq manual - * section 16.7.4 for details. + * the at91rm9200 manual, section 41.3.1 or the Zynq manual + * section 16.7.4 for details. RXUBR is only enabled for + * these two versions. */ if (status & MACB_BIT(RXUBR)) { ctrl = macb_readl(bp, NCR); @@ -2259,7 +2259,7 @@ static void macb_init_hw(struct macb *bp) /* Enable interrupts */ queue_writel(queue, IER, - MACB_RX_INT_FLAGS | + bp->rx_intr_mask | MACB_TX_INT_FLAGS | MACB_BIT(HRESP)); } @@ -3907,6 +3907,7 @@ static const struct macb_config sama5d4_config = { }; static const struct macb_config emac_config = { + .caps = MACB_CAPS_NEEDS_RSTONUBR, .clk_init = at91ether_clk_init, .init = at91ether_init, }; @@ -3928,7 +3929,8 @@ static const struct macb_config zynqmp_config = { }; static const struct macb_config zynq_config = { - .caps = MACB_CAPS_GIGABIT_MODE_AVAILABLE | MACB_CAPS_NO_GIGABIT_HALF, + .caps = MACB_CAPS_GIGABIT_MODE_AVAILABLE | MACB_CAPS_NO_GIGABIT_HALF | + MACB_CAPS_NEEDS_RSTONUBR, .dma_burst_length = 16, .clk_init = macb_clk_init, .init = macb_init, @@ -4083,6 +4085,10 @@ static int macb_probe(struct platform_device *pdev) macb_dma_desc_get_size(bp); } + bp->rx_intr_mask = MACB_RX_INT_FLAGS; + if (bp->caps & MACB_CAPS_NEEDS_RSTONUBR) + bp->rx_intr_mask |= MACB_BIT(RXUBR); + mac = of_get_mac_address(np); if (mac) { ether_addr_copy(bp->dev->dev_addr, mac); From 896cebc0e261a29a6f0b20ab21d1bfc06959e91d Mon Sep 17 00:00:00 2001 From: Yang Wei Date: Tue, 29 Jan 2019 22:40:51 +0800 Subject: [PATCH 75/79] net: 8139cp: replace dev_kfree_skb_irq by dev_consume_skb_irq for drop profiles dev_consume_skb_irq() should be called in cp_tx() when skb xmit done. It makes drop profiles(dropwatch, perf) more friendly. Signed-off-by: Yang Wei Signed-off-by: David S. Miller --- drivers/net/ethernet/realtek/8139cp.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/realtek/8139cp.c b/drivers/net/ethernet/realtek/8139cp.c index 44f6e4873aad..4f910c4f67b0 100644 --- a/drivers/net/ethernet/realtek/8139cp.c +++ b/drivers/net/ethernet/realtek/8139cp.c @@ -691,7 +691,7 @@ static void cp_tx (struct cp_private *cp) } bytes_compl += skb->len; pkts_compl++; - dev_kfree_skb_irq(skb); + dev_consume_skb_irq(skb); } cp->tx_skb[tx_tail] = NULL; From e339f8631eb70fb48b2801e9e66a267c9a730a6a Mon Sep 17 00:00:00 2001 From: Yang Wei Date: Tue, 29 Jan 2019 23:32:22 +0800 Subject: [PATCH 76/79] net: caif: call dev_consume_skb_any when skb xmit done The skb shouled be consumed when xmit done, it makes drop profiles (dropwatch, perf) more friendly. dev_kfree_skb_irq()/kfree_skb() shouled be replaced by dev_consume_skb_any(), it makes code cleaner. Signed-off-by: Yang Wei Signed-off-by: David S. Miller --- drivers/net/caif/caif_serial.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/drivers/net/caif/caif_serial.c b/drivers/net/caif/caif_serial.c index a0f954f36c09..44e6c7b1b222 100644 --- a/drivers/net/caif/caif_serial.c +++ b/drivers/net/caif/caif_serial.c @@ -257,10 +257,7 @@ static int handle_tx(struct ser_device *ser) if (skb->len == 0) { struct sk_buff *tmp = skb_dequeue(&ser->head); WARN_ON(tmp != skb); - if (in_interrupt()) - dev_kfree_skb_irq(skb); - else - kfree_skb(skb); + dev_consume_skb_any(skb); } } /* Send flow off if queue is empty */ From 0f0ed8282e5bfdc87cdd562e58f3d90d893e7ee5 Mon Sep 17 00:00:00 2001 From: Yang Wei Date: Tue, 29 Jan 2019 23:04:40 +0800 Subject: [PATCH 77/79] net: b44: replace dev_kfree_skb_xxx by dev_consume_skb_xxx for drop profiles The skb should be freed by dev_consume_skb_any() in b44_start_xmit() when bounce_skb is used. The skb is be replaced by bounce_skb, so the original skb should be consumed(not drop). dev_consume_skb_irq() should be called in b44_tx() when skb xmit done. It makes drop profiles(dropwatch, perf) more friendly. Signed-off-by: Yang Wei Signed-off-by: David S. Miller --- drivers/net/ethernet/broadcom/b44.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/broadcom/b44.c b/drivers/net/ethernet/broadcom/b44.c index f44808959ff3..97ab0dd25552 100644 --- a/drivers/net/ethernet/broadcom/b44.c +++ b/drivers/net/ethernet/broadcom/b44.c @@ -638,7 +638,7 @@ static void b44_tx(struct b44 *bp) bytes_compl += skb->len; pkts_compl++; - dev_kfree_skb_irq(skb); + dev_consume_skb_irq(skb); } netdev_completed_queue(bp->dev, pkts_compl, bytes_compl); @@ -1012,7 +1012,7 @@ static netdev_tx_t b44_start_xmit(struct sk_buff *skb, struct net_device *dev) } skb_copy_from_linear_data(skb, skb_put(bounce_skb, len), len); - dev_kfree_skb_any(skb); + dev_consume_skb_any(skb); skb = bounce_skb; } From 35edfdc77f683c8fd27d7732af06cf6489af60a5 Mon Sep 17 00:00:00 2001 From: Josh Elsasser Date: Sat, 26 Jan 2019 14:38:33 -0800 Subject: [PATCH 78/79] net: set default network namespace in init_dummy_netdev() Assign a default net namespace to netdevs created by init_dummy_netdev(). Fixes a NULL pointer dereference caused by busy-polling a socket bound to an iwlwifi wireless device, which bumps the per-net BUSYPOLLRXPACKETS stat if napi_poll() received packets: BUG: unable to handle kernel NULL pointer dereference at 0000000000000190 IP: napi_busy_loop+0xd6/0x200 Call Trace: sock_poll+0x5e/0x80 do_sys_poll+0x324/0x5a0 SyS_poll+0x6c/0xf0 do_syscall_64+0x6b/0x1f0 entry_SYSCALL_64_after_hwframe+0x3d/0xa2 Fixes: 7db6b048da3b ("net: Commonize busy polling code to focus on napi_id instead of socket") Signed-off-by: Josh Elsasser Signed-off-by: David S. Miller --- net/core/dev.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/net/core/dev.c b/net/core/dev.c index 82f20022259d..8e276e0192a1 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -8712,6 +8712,9 @@ int init_dummy_netdev(struct net_device *dev) set_bit(__LINK_STATE_PRESENT, &dev->state); set_bit(__LINK_STATE_START, &dev->state); + /* napi_busy_loop stats accounting wants this */ + dev_net_set(dev, &init_net); + /* Note : We dont allocate pcpu_refcnt for dummy devices, * because users of this 'device' dont need to change * its refcount. From d07e1e0febe10b65eecd3205ad3bd1e999754887 Mon Sep 17 00:00:00 2001 From: Jesper Dangaard Brouer Date: Tue, 29 Jan 2019 14:22:33 +0100 Subject: [PATCH 79/79] MAINTAINERS: Add entry for XDP (eXpress Data Path) Add multiple people as maintainers for XDP, sorted alphabetically. XDP is also tied to driver level support and code, but we cannot add all drivers to the list. Instead K: and N: match on 'xdp' in hope to catch some of those changes in drivers. Signed-off-by: Jesper Dangaard Brouer Acked-by: John Fastabend Acked-by: Jakub Kicinski Acked-by: David S. Miller Acked-by: Daniel Borkmann Acked-by: Alexei Starovoitov Signed-off-by: David S. Miller --- MAINTAINERS | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) diff --git a/MAINTAINERS b/MAINTAINERS index 9f64f8d3740e..2d3c1918f1b0 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -16673,6 +16673,24 @@ T: git git://linuxtv.org/media_tree.git S: Maintained F: drivers/media/tuners/tuner-xc2028.* +XDP (eXpress Data Path) +M: Alexei Starovoitov +M: Daniel Borkmann +M: David S. Miller +M: Jakub Kicinski +M: Jesper Dangaard Brouer +M: John Fastabend +L: netdev@vger.kernel.org +L: xdp-newbies@vger.kernel.org +S: Supported +F: net/core/xdp.c +F: include/net/xdp.h +F: kernel/bpf/devmap.c +F: kernel/bpf/cpumap.c +F: include/trace/events/xdp.h +K: xdp +N: xdp + XDP SOCKETS (AF_XDP) M: Björn Töpel M: Magnus Karlsson