From: Sasha Levin Date: Fri, 7 Aug 2020 20:10:52 +0000 (-0400) Subject: Fixes for 4.9 X-Git-Tag: v4.19.139~24 X-Git-Url: http://git.ipfire.org/?a=commitdiff_plain;h=7ca7f63fe4b9cb3da8905356634c4eaeac7a5d08;p=thirdparty%2Fkernel%2Fstable-queue.git Fixes for 4.9 Signed-off-by: Sasha Levin --- diff --git a/queue-4.9/atm-fix-atm_dev-refcnt-leaks-in-atmtcp_remove_persis.patch b/queue-4.9/atm-fix-atm_dev-refcnt-leaks-in-atmtcp_remove_persis.patch new file mode 100644 index 00000000000..ea2656db2fc --- /dev/null +++ b/queue-4.9/atm-fix-atm_dev-refcnt-leaks-in-atmtcp_remove_persis.patch @@ -0,0 +1,54 @@ +From d5190b4a942702f0ab352109cbc285403233a7b1 Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Wed, 29 Jul 2020 21:06:59 +0800 +Subject: atm: fix atm_dev refcnt leaks in atmtcp_remove_persistent + +From: Xin Xiong + +[ Upstream commit 51875dad43b44241b46a569493f1e4bfa0386d86 ] + +atmtcp_remove_persistent() invokes atm_dev_lookup(), which returns a +reference of atm_dev with increased refcount or NULL if fails. + +The refcount leaks issues occur in two error handling paths. If +dev_data->persist is zero or PRIV(dev)->vcc isn't NULL, the function +returns 0 without decreasing the refcount kept by a local variable, +resulting in refcount leaks. + +Fix the issue by adding atm_dev_put() before returning 0 both when +dev_data->persist is zero or PRIV(dev)->vcc isn't NULL. + +Signed-off-by: Xin Xiong +Signed-off-by: Xiyu Yang +Signed-off-by: Xin Tan +Signed-off-by: David S. Miller +Signed-off-by: Sasha Levin +--- + drivers/atm/atmtcp.c | 10 ++++++++-- + 1 file changed, 8 insertions(+), 2 deletions(-) + +diff --git a/drivers/atm/atmtcp.c b/drivers/atm/atmtcp.c +index 480fa6ffbc090..04fca6db273ef 100644 +--- a/drivers/atm/atmtcp.c ++++ b/drivers/atm/atmtcp.c +@@ -432,9 +432,15 @@ static int atmtcp_remove_persistent(int itf) + return -EMEDIUMTYPE; + } + dev_data = PRIV(dev); +- if (!dev_data->persist) return 0; ++ if (!dev_data->persist) { ++ atm_dev_put(dev); ++ return 0; ++ } + dev_data->persist = 0; +- if (PRIV(dev)->vcc) return 0; ++ if (PRIV(dev)->vcc) { ++ atm_dev_put(dev); ++ return 0; ++ } + kfree(dev_data); + atm_dev_put(dev); + atm_dev_deregister(dev); +-- +2.25.1 + diff --git a/queue-4.9/cfg80211-check-vendor-command-doit-pointer-before-us.patch b/queue-4.9/cfg80211-check-vendor-command-doit-pointer-before-us.patch new file mode 100644 index 00000000000..fbfd351870d --- /dev/null +++ b/queue-4.9/cfg80211-check-vendor-command-doit-pointer-before-us.patch @@ -0,0 +1,50 @@ +From f13a01ac17f88346e52469f04bab524c1d35d892 Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Mon, 6 Jul 2020 17:13:53 -0400 +Subject: cfg80211: check vendor command doit pointer before use + +From: Julian Squires + +[ Upstream commit 4052d3d2e8f47a15053320bbcbe365d15610437d ] + +In the case where a vendor command does not implement doit, and has no +flags set, doit would not be validated and a NULL pointer dereference +would occur, for example when invoking the vendor command via iw. + +I encountered this while developing new vendor commands. Perhaps in +practice it is advisable to always implement doit along with dumpit, +but it seems reasonable to me to always check doit anyway, not just +when NEED_WDEV. + +Signed-off-by: Julian Squires +Link: https://lore.kernel.org/r/20200706211353.2366470-1-julian@cipht.net +Signed-off-by: Johannes Berg +Signed-off-by: Sasha Levin +--- + net/wireless/nl80211.c | 6 +++--- + 1 file changed, 3 insertions(+), 3 deletions(-) + +diff --git a/net/wireless/nl80211.c b/net/wireless/nl80211.c +index 0048f90944ddf..e107754e29a77 100644 +--- a/net/wireless/nl80211.c ++++ b/net/wireless/nl80211.c +@@ -11317,13 +11317,13 @@ static int nl80211_vendor_cmd(struct sk_buff *skb, struct genl_info *info) + if (!wdev->netdev && !wdev->p2p_started) + return -ENETDOWN; + } +- +- if (!vcmd->doit) +- return -EOPNOTSUPP; + } else { + wdev = NULL; + } + ++ if (!vcmd->doit) ++ return -EOPNOTSUPP; ++ + if (info->attrs[NL80211_ATTR_VENDOR_DATA]) { + data = nla_data(info->attrs[NL80211_ATTR_VENDOR_DATA]); + len = nla_len(info->attrs[NL80211_ATTR_VENDOR_DATA]); +-- +2.25.1 + diff --git a/queue-4.9/drm-nouveau-fbcon-fix-module-unload-when-fbcon-init-.patch b/queue-4.9/drm-nouveau-fbcon-fix-module-unload-when-fbcon-init-.patch new file mode 100644 index 00000000000..6bf9b6ec53e --- /dev/null +++ b/queue-4.9/drm-nouveau-fbcon-fix-module-unload-when-fbcon-init-.patch @@ -0,0 +1,33 @@ +From 5c4f3407b19ce8a44e469ee25f3c13fd179fe62d Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Fri, 24 Jul 2020 17:01:39 +1000 +Subject: drm/nouveau/fbcon: fix module unload when fbcon init has failed for + some reason + +From: Ben Skeggs + +[ Upstream commit 498595abf5bd51f0ae074cec565d888778ea558f ] + +Stale pointer was tripping up the unload path. + +Signed-off-by: Ben Skeggs +Signed-off-by: Sasha Levin +--- + drivers/gpu/drm/nouveau/nouveau_fbcon.c | 1 + + 1 file changed, 1 insertion(+) + +diff --git a/drivers/gpu/drm/nouveau/nouveau_fbcon.c b/drivers/gpu/drm/nouveau/nouveau_fbcon.c +index 2b79e27dd89c6..275abc424ce25 100644 +--- a/drivers/gpu/drm/nouveau/nouveau_fbcon.c ++++ b/drivers/gpu/drm/nouveau/nouveau_fbcon.c +@@ -584,6 +584,7 @@ fini: + drm_fb_helper_fini(&fbcon->helper); + free: + kfree(fbcon); ++ drm->fbcon = NULL; + return ret; + } + +-- +2.25.1 + diff --git a/queue-4.9/igb-reinit_locked-should-be-called-with-rtnl_lock.patch b/queue-4.9/igb-reinit_locked-should-be-called-with-rtnl_lock.patch new file mode 100644 index 00000000000..cb9723bab3c --- /dev/null +++ b/queue-4.9/igb-reinit_locked-should-be-called-with-rtnl_lock.patch @@ -0,0 +1,92 @@ +From 9c794c708912228e27d80dd17f693d83f12d419a Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Thu, 2 Jul 2020 15:39:06 -0700 +Subject: igb: reinit_locked() should be called with rtnl_lock + +From: Francesco Ruggeri + +[ Upstream commit 024a8168b749db7a4aa40a5fbdfa04bf7e77c1c0 ] + +We observed two panics involving races with igb_reset_task. +The first panic is caused by this race condition: + + kworker reboot -f + + igb_reset_task + igb_reinit_locked + igb_down + napi_synchronize + __igb_shutdown + igb_clear_interrupt_scheme + igb_free_q_vectors + igb_free_q_vector + adapter->q_vector[v_idx] = NULL; + napi_disable + Panics trying to access + adapter->q_vector[v_idx].napi_state + +The second panic (a divide error) is caused by this race: + +kworker reboot -f tx packet + +igb_reset_task + __igb_shutdown + rtnl_lock() + ... + igb_clear_interrupt_scheme + igb_free_q_vectors + adapter->num_tx_queues = 0 + ... + rtnl_unlock() +rtnl_lock() +igb_reinit_locked +igb_down +igb_up +netif_tx_start_all_queues + dev_hard_start_xmit + igb_xmit_frame + igb_tx_queue_mapping + Panics on + r_idx % adapter->num_tx_queues + +This commit applies to igb_reset_task the same changes that +were applied to ixgbe in commit 2f90b8657ec9 ("ixgbe: this patch +adds support for DCB to the kernel and ixgbe driver"), +commit 8f4c5c9fb87a ("ixgbe: reinit_locked() should be called with +rtnl_lock") and commit 88adce4ea8f9 ("ixgbe: fix possible race in +reset subtask"). + +Signed-off-by: Francesco Ruggeri +Tested-by: Aaron Brown +Signed-off-by: Tony Nguyen +Signed-off-by: Sasha Levin +--- + drivers/net/ethernet/intel/igb/igb_main.c | 9 +++++++++ + 1 file changed, 9 insertions(+) + +diff --git a/drivers/net/ethernet/intel/igb/igb_main.c b/drivers/net/ethernet/intel/igb/igb_main.c +index 90eab0521be12..43fb77c6c51ed 100644 +--- a/drivers/net/ethernet/intel/igb/igb_main.c ++++ b/drivers/net/ethernet/intel/igb/igb_main.c +@@ -5381,9 +5381,18 @@ static void igb_reset_task(struct work_struct *work) + struct igb_adapter *adapter; + adapter = container_of(work, struct igb_adapter, reset_task); + ++ rtnl_lock(); ++ /* If we're already down or resetting, just bail */ ++ if (test_bit(__IGB_DOWN, &adapter->state) || ++ test_bit(__IGB_RESETTING, &adapter->state)) { ++ rtnl_unlock(); ++ return; ++ } ++ + igb_dump(adapter); + netdev_err(adapter->netdev, "Reset adapter\n"); + igb_reinit_locked(adapter); ++ rtnl_unlock(); + } + + /** +-- +2.25.1 + diff --git a/queue-4.9/net-9p-validate-fds-in-p9_fd_open.patch b/queue-4.9/net-9p-validate-fds-in-p9_fd_open.patch new file mode 100644 index 00000000000..e09a32e833f --- /dev/null +++ b/queue-4.9/net-9p-validate-fds-in-p9_fd_open.patch @@ -0,0 +1,70 @@ +From 72c9370ef7a0ff4ef7d770bfd268b6d7fda4630d Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Fri, 10 Jul 2020 10:57:22 +0200 +Subject: net/9p: validate fds in p9_fd_open + +From: Christoph Hellwig + +[ Upstream commit a39c46067c845a8a2d7144836e9468b7f072343e ] + +p9_fd_open just fgets file descriptors passed in from userspace, but +doesn't verify that they are valid for read or writing. This gets +cought down in the VFS when actually attempting a read or write, but +a new warning added in linux-next upsets syzcaller. + +Fix this by just verifying the fds early on. + +Link: http://lkml.kernel.org/r/20200710085722.435850-1-hch@lst.de +Reported-by: syzbot+e6f77e16ff68b2434a2c@syzkaller.appspotmail.com +Signed-off-by: Christoph Hellwig +[Dominique: amend goto as per Doug Nazar's review] +Signed-off-by: Dominique Martinet +Signed-off-by: Sasha Levin +--- + net/9p/trans_fd.c | 24 ++++++++++++++++-------- + 1 file changed, 16 insertions(+), 8 deletions(-) + +diff --git a/net/9p/trans_fd.c b/net/9p/trans_fd.c +index b0f47563f0bf3..bad27b0ec65d6 100644 +--- a/net/9p/trans_fd.c ++++ b/net/9p/trans_fd.c +@@ -815,20 +815,28 @@ static int p9_fd_open(struct p9_client *client, int rfd, int wfd) + return -ENOMEM; + + ts->rd = fget(rfd); ++ if (!ts->rd) ++ goto out_free_ts; ++ if (!(ts->rd->f_mode & FMODE_READ)) ++ goto out_put_rd; + ts->wr = fget(wfd); +- if (!ts->rd || !ts->wr) { +- if (ts->rd) +- fput(ts->rd); +- if (ts->wr) +- fput(ts->wr); +- kfree(ts); +- return -EIO; +- } ++ if (!ts->wr) ++ goto out_put_rd; ++ if (!(ts->wr->f_mode & FMODE_WRITE)) ++ goto out_put_wr; + + client->trans = ts; + client->status = Connected; + + return 0; ++ ++out_put_wr: ++ fput(ts->wr); ++out_put_rd: ++ fput(ts->rd); ++out_free_ts: ++ kfree(ts); ++ return -EIO; + } + + static int p9_socket_open(struct p9_client *client, struct socket *csocket) +-- +2.25.1 + diff --git a/queue-4.9/series b/queue-4.9/series index f439c5c8fcb..c6b08dc93c6 100644 --- a/queue-4.9/series +++ b/queue-4.9/series @@ -61,3 +61,9 @@ leds-wm831x-status-fix-use-after-free-on-unbind.patch leds-da903x-fix-use-after-free-on-unbind.patch leds-lm3533-fix-use-after-free-on-unbind.patch leds-88pm860x-fix-use-after-free-on-unbind.patch +net-9p-validate-fds-in-p9_fd_open.patch +drm-nouveau-fbcon-fix-module-unload-when-fbcon-init-.patch +cfg80211-check-vendor-command-doit-pointer-before-us.patch +igb-reinit_locked-should-be-called-with-rtnl_lock.patch +atm-fix-atm_dev-refcnt-leaks-in-atmtcp_remove_persis.patch +tools-lib-traceevent-fix-memory-leak-in-process_dyna.patch diff --git a/queue-4.9/tools-lib-traceevent-fix-memory-leak-in-process_dyna.patch b/queue-4.9/tools-lib-traceevent-fix-memory-leak-in-process_dyna.patch new file mode 100644 index 00000000000..bcda06b0989 --- /dev/null +++ b/queue-4.9/tools-lib-traceevent-fix-memory-leak-in-process_dyna.patch @@ -0,0 +1,72 @@ +From 3f043a9a7517ff04151a936d4e71a60745e87d6e Mon Sep 17 00:00:00 2001 +From: Sasha Levin +Date: Thu, 30 Jul 2020 11:02:36 -0400 +Subject: tools lib traceevent: Fix memory leak in process_dynamic_array_len + +From: Philippe Duplessis-Guindon + +[ Upstream commit e24c6447ccb7b1a01f9bf0aec94939e6450c0b4d ] + +I compiled with AddressSanitizer and I had these memory leaks while I +was using the tep_parse_format function: + + Direct leak of 28 byte(s) in 4 object(s) allocated from: + #0 0x7fb07db49ffe in __interceptor_realloc (/lib/x86_64-linux-gnu/libasan.so.5+0x10dffe) + #1 0x7fb07a724228 in extend_token /home/pduplessis/repo/linux/tools/lib/traceevent/event-parse.c:985 + #2 0x7fb07a724c21 in __read_token /home/pduplessis/repo/linux/tools/lib/traceevent/event-parse.c:1140 + #3 0x7fb07a724f78 in read_token /home/pduplessis/repo/linux/tools/lib/traceevent/event-parse.c:1206 + #4 0x7fb07a725191 in __read_expect_type /home/pduplessis/repo/linux/tools/lib/traceevent/event-parse.c:1291 + #5 0x7fb07a7251df in read_expect_type /home/pduplessis/repo/linux/tools/lib/traceevent/event-parse.c:1299 + #6 0x7fb07a72e6c8 in process_dynamic_array_len /home/pduplessis/repo/linux/tools/lib/traceevent/event-parse.c:2849 + #7 0x7fb07a7304b8 in process_function /home/pduplessis/repo/linux/tools/lib/traceevent/event-parse.c:3161 + #8 0x7fb07a730900 in process_arg_token /home/pduplessis/repo/linux/tools/lib/traceevent/event-parse.c:3207 + #9 0x7fb07a727c0b in process_arg /home/pduplessis/repo/linux/tools/lib/traceevent/event-parse.c:1786 + #10 0x7fb07a731080 in event_read_print_args /home/pduplessis/repo/linux/tools/lib/traceevent/event-parse.c:3285 + #11 0x7fb07a731722 in event_read_print /home/pduplessis/repo/linux/tools/lib/traceevent/event-parse.c:3369 + #12 0x7fb07a740054 in __tep_parse_format /home/pduplessis/repo/linux/tools/lib/traceevent/event-parse.c:6335 + #13 0x7fb07a74047a in __parse_event /home/pduplessis/repo/linux/tools/lib/traceevent/event-parse.c:6389 + #14 0x7fb07a740536 in tep_parse_format /home/pduplessis/repo/linux/tools/lib/traceevent/event-parse.c:6431 + #15 0x7fb07a785acf in parse_event ../../../src/fs-src/fs.c:251 + #16 0x7fb07a785ccd in parse_systems ../../../src/fs-src/fs.c:284 + #17 0x7fb07a786fb3 in read_metadata ../../../src/fs-src/fs.c:593 + #18 0x7fb07a78760e in ftrace_fs_source_init ../../../src/fs-src/fs.c:727 + #19 0x7fb07d90c19c in add_component_with_init_method_data ../../../../src/lib/graph/graph.c:1048 + #20 0x7fb07d90c87b in add_source_component_with_initialize_method_data ../../../../src/lib/graph/graph.c:1127 + #21 0x7fb07d90c92a in bt_graph_add_source_component ../../../../src/lib/graph/graph.c:1152 + #22 0x55db11aa632e in cmd_run_ctx_create_components_from_config_components ../../../src/cli/babeltrace2.c:2252 + #23 0x55db11aa6fda in cmd_run_ctx_create_components ../../../src/cli/babeltrace2.c:2347 + #24 0x55db11aa780c in cmd_run ../../../src/cli/babeltrace2.c:2461 + #25 0x55db11aa8a7d in main ../../../src/cli/babeltrace2.c:2673 + #26 0x7fb07d5460b2 in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x270b2) + +The token variable in the process_dynamic_array_len function is +allocated in the read_expect_type function, but is not freed before +calling the read_token function. + +Free the token variable before calling read_token in order to plug the +leak. + +Signed-off-by: Philippe Duplessis-Guindon +Reviewed-by: Steven Rostedt (VMware) +Link: https://lore.kernel.org/linux-trace-devel/20200730150236.5392-1-pduplessis@efficios.com +Signed-off-by: Arnaldo Carvalho de Melo +Signed-off-by: Sasha Levin +--- + tools/lib/traceevent/event-parse.c | 1 + + 1 file changed, 1 insertion(+) + +diff --git a/tools/lib/traceevent/event-parse.c b/tools/lib/traceevent/event-parse.c +index 62f4cacf253ab..b9db8739487dc 100644 +--- a/tools/lib/traceevent/event-parse.c ++++ b/tools/lib/traceevent/event-parse.c +@@ -2764,6 +2764,7 @@ process_dynamic_array_len(struct event_format *event, struct print_arg *arg, + if (read_expected(EVENT_DELIM, ")") < 0) + goto out_err; + ++ free_token(token); + type = read_token(&token); + *tok = token; + +-- +2.25.1 +