February 2018 - Linux-stable-mirror

Patch "ALSA: caiaq: Add a sanity check for invalid EPs" has been added to the 4.14-stable tree

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled ALSA: caiaq: Add a sanity check for invalid EPs to the 4.14-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… The filename of the patch is: alsa-caiaq-add-a-sanity-check-for-invalid-eps.patch and it can be found in the queue-4.14 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable(a)vger.kernel.org> know about it. >From 58fc7f73a85d45a47057dad2af53502fdf6cf778 Mon Sep 17 00:00:00 2001 From: Takashi Iwai <tiwai(a)suse.de> Date: Wed, 4 Oct 2017 15:07:21 +0200 Subject: ALSA: caiaq: Add a sanity check for invalid EPs From: Takashi Iwai <tiwai(a)suse.de> commit 58fc7f73a85d45a47057dad2af53502fdf6cf778 upstream. As syzkaller spotted, currently caiaq driver submits a URB with the fixed EP without checking whether it's actually available, which may result in a kernel warning like: usb 1-1: BOGUS urb xfer, pipe 3 != type 1 ------------[ cut here ]------------ WARNING: CPU: 1 PID: 1150 at drivers/usb/core/urb.c:449 usb_submit_urb+0xf8a/0x11d0 Modules linked in: CPU: 1 PID: 1150 Comm: kworker/1:1 Not tainted 4.14.0-rc2-42660-g24b7bd59eec0 #277 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011 Workqueue: usb_hub_wq hub_event Call Trace: init_card sound/usb/caiaq/device.c:467 snd_probe+0x81c/0x1150 sound/usb/caiaq/device.c:525 usb_probe_interface+0x35d/0x8e0 drivers/usb/core/driver.c:361 .... This patch adds a sanity check of validity of EPs at the device initialization phase for avoiding the call with an invalid EP. Reported-by: Andrey Konovalov <andreyknvl(a)google.com> Tested-by: Andrey Konovalov <andreyknvl(a)google.com> Signed-off-by: Takashi Iwai <tiwai(a)suse.de> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- sound/usb/caiaq/device.c | 7 +++++++ 1 file changed, 7 insertions(+) --- a/sound/usb/caiaq/device.c +++ b/sound/usb/caiaq/device.c @@ -461,6 +461,13 @@ static int init_card(struct snd_usb_caia cdev->midi_out_buf, EP1_BUFSIZE, snd_usb_caiaq_midi_output_done, cdev); + /* sanity checks of EPs before actually submitting */ + if (usb_urb_ep_type_check(&cdev->ep1_in_urb) || + usb_urb_ep_type_check(&cdev->midi_out_urb)) { + dev_err(dev, "invalid EPs\n"); + return -EINVAL; + } + init_waitqueue_head(&cdev->ep1_wait_queue); init_waitqueue_head(&cdev->prepare_wait_queue); Patches currently in stable-queue which might be from tiwai(a)suse.de are queue-4.14/alsa-bcd2000-add-a-sanity-check-for-invalid-eps.patch queue-4.14/alsa-caiaq-add-a-sanity-check-for-invalid-eps.patch queue-4.14/alsa-line6-add-a-sanity-check-for-invalid-eps.patch queue-4.14/usb-core-add-a-helper-function-to-check-the-validity-of-ep-type-in-urb.patch

6 years, 10 months

1
0
0 0

Patch "ALSA: bcd2000: Add a sanity check for invalid EPs" has been added to the 4.14-stable tree

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled ALSA: bcd2000: Add a sanity check for invalid EPs to the 4.14-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… The filename of the patch is: alsa-bcd2000-add-a-sanity-check-for-invalid-eps.patch and it can be found in the queue-4.14 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable(a)vger.kernel.org> know about it. >From 6815a0b444572527256f0d0efd8efe3ddede6018 Mon Sep 17 00:00:00 2001 From: Takashi Iwai <tiwai(a)suse.de> Date: Wed, 4 Oct 2017 15:03:40 +0200 Subject: ALSA: bcd2000: Add a sanity check for invalid EPs From: Takashi Iwai <tiwai(a)suse.de> commit 6815a0b444572527256f0d0efd8efe3ddede6018 upstream. As syzkaller spotted, currently bcd2000 driver submits a URB with the fixed EP without checking whether it's actually available, which may result in a kernel warning like: usb 1-1: BOGUS urb xfer, pipe 1 != type 3 ------------[ cut here ]------------ WARNING: CPU: 0 PID: 1846 at drivers/usb/core/urb.c:449 usb_submit_urb+0xf8a/0x11d0 Modules linked in: CPU: 0 PID: 1846 Comm: kworker/0:2 Not tainted 4.14.0-rc2-42613-g1488251d1a98 #238 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011 Workqueue: usb_hub_wq hub_event Call Trace: bcd2000_init_device sound/usb/bcd2000/bcd2000.c:289 bcd2000_init_midi sound/usb/bcd2000/bcd2000.c:345 bcd2000_probe+0xe64/0x19e0 sound/usb/bcd2000/bcd2000.c:406 usb_probe_interface+0x35d/0x8e0 drivers/usb/core/driver.c:361 .... This patch adds a sanity check of validity of EPs at the device initialization phase for avoiding the call with an invalid EP. Reported-by: Andrey Konovalov <andreyknvl(a)google.com> Tested-by: Andrey Konovalov <andreyknvl(a)google.com> Signed-off-by: Takashi Iwai <tiwai(a)suse.de> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- sound/usb/bcd2000/bcd2000.c | 7 +++++++ 1 file changed, 7 insertions(+) --- a/sound/usb/bcd2000/bcd2000.c +++ b/sound/usb/bcd2000/bcd2000.c @@ -342,6 +342,13 @@ static int bcd2000_init_midi(struct bcd2 bcd2k->midi_out_buf, BUFSIZE, bcd2000_output_complete, bcd2k, 1); + /* sanity checks of EPs before actually submitting */ + if (usb_urb_ep_type_check(bcd2k->midi_in_urb) || + usb_urb_ep_type_check(bcd2k->midi_out_urb)) { + dev_err(&bcd2k->dev->dev, "invalid MIDI EP\n"); + return -EINVAL; + } + bcd2000_init_device(bcd2k); return 0; Patches currently in stable-queue which might be from tiwai(a)suse.de are queue-4.14/alsa-bcd2000-add-a-sanity-check-for-invalid-eps.patch queue-4.14/alsa-caiaq-add-a-sanity-check-for-invalid-eps.patch queue-4.14/alsa-line6-add-a-sanity-check-for-invalid-eps.patch queue-4.14/usb-core-add-a-helper-function-to-check-the-validity-of-ep-type-in-urb.patch

6 years, 10 months

1
0
0 0

Patch "tun: fix tun_napi_alloc_frags() frag allocator" has been added to the 4.15-stable tree

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled tun: fix tun_napi_alloc_frags() frag allocator to the 4.15-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… The filename of the patch is: tun-fix-tun_napi_alloc_frags-frag-allocator.patch and it can be found in the queue-4.15 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable(a)vger.kernel.org> know about it. >From 43a08e0f58b3f236165029710a4e3b303815253b Mon Sep 17 00:00:00 2001 From: Eric Dumazet <edumazet(a)google.com> Date: Thu, 15 Feb 2018 14:47:15 -0800 Subject: tun: fix tun_napi_alloc_frags() frag allocator MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit From: Eric Dumazet <edumazet(a)google.com> commit 43a08e0f58b3f236165029710a4e3b303815253b upstream. <Mark Rutland reported> While fuzzing arm64 v4.16-rc1 with Syzkaller, I've been hitting a misaligned atomic in __skb_clone: atomic_inc(&(skb_shinfo(skb)->dataref)); where dataref doesn't have the required natural alignment, and the atomic operation faults. e.g. i often see it aligned to a single byte boundary rather than a four byte boundary. AFAICT, the skb_shared_info is misaligned at the instant it's allocated in __napi_alloc_skb() __napi_alloc_skb() </end of report> Problem is caused by tun_napi_alloc_frags() using napi_alloc_frag() with user provided seg sizes, leading to other users of this API getting unaligned page fragments. Since we would like to not necessarily add paddings or alignments to the frags that tun_napi_alloc_frags() attaches to the skb, switch to another page frag allocator. As a bonus skb_page_frag_refill() can use GFP_KERNEL allocations, meaning that we can not deplete memory reserves as easily. Fixes: 90e33d459407 ("tun: enable napi_gro_frags() for TUN/TAP driver") Signed-off-by: Eric Dumazet <edumazet(a)google.com> Reported-by: Mark Rutland <mark.rutland(a)arm.com> Tested-by: Mark Rutland <mark.rutland(a)arm.com> Signed-off-by: David S. Miller <davem(a)davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- drivers/net/tun.c | 16 ++++++---------- 1 file changed, 6 insertions(+), 10 deletions(-) --- a/drivers/net/tun.c +++ b/drivers/net/tun.c @@ -1317,27 +1317,23 @@ static struct sk_buff *tun_napi_alloc_fr skb->truesize += skb->data_len; for (i = 1; i < it->nr_segs; i++) { + struct page_frag *pfrag = &current->task_frag; size_t fragsz = it->iov[i].iov_len; - unsigned long offset; - struct page *page; - void *data; if (fragsz == 0 || fragsz > PAGE_SIZE) { err = -EINVAL; goto free; } - local_bh_disable(); - data = napi_alloc_frag(fragsz); - local_bh_enable(); - if (!data) { + if (!skb_page_frag_refill(fragsz, pfrag, GFP_KERNEL)) { err = -ENOMEM; goto free; } - page = virt_to_head_page(data); - offset = data - page_address(page); - skb_fill_page_desc(skb, i - 1, page, offset, fragsz); + skb_fill_page_desc(skb, i - 1, pfrag->page, + pfrag->offset, fragsz); + page_ref_inc(pfrag->page); + pfrag->offset += fragsz; } return skb; Patches currently in stable-queue which might be from edumazet(a)google.com are queue-4.15/net_sched-gen_estimator-fix-lockdep-splat.patch queue-4.15/tun-fix-tun_napi_alloc_frags-frag-allocator.patch queue-4.15/netfilter-x_tables-avoid-out-of-bounds-reads-in-xt_request_find_-match-target.patch queue-4.15/netfilter-xt_rateest-acquire-xt_rateest_mutex-for-hash-insert.patch

6 years, 10 months

1
0
0 0

Patch "staging: android: ion: Switch from WARN to pr_warn" has been added to the 4.15-stable tree

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled staging: android: ion: Switch from WARN to pr_warn to the 4.15-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… The filename of the patch is: staging-android-ion-switch-from-warn-to-pr_warn.patch and it can be found in the queue-4.15 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable(a)vger.kernel.org> know about it. >From e4e179a844f52e907e550f887d0a2171f1508af1 Mon Sep 17 00:00:00 2001 From: Laura Abbott <labbott(a)redhat.com> Date: Fri, 5 Jan 2018 11:14:09 -0800 Subject: staging: android: ion: Switch from WARN to pr_warn From: Laura Abbott <labbott(a)redhat.com> commit e4e179a844f52e907e550f887d0a2171f1508af1 upstream. Syzbot reported a warning with Ion: WARNING: CPU: 0 PID: 3502 at drivers/staging/android/ion/ion-ioctl.c:73 ion_ioctl+0x2db/0x380 drivers/staging/android/ion/ion-ioctl.c:73 Kernel panic - not syncing: panic_on_warn set ... This is a warning that validation of the ioctl fields failed. This was deliberately added as a warning to make it very obvious to developers that something needed to be fixed. In reality, this is overkill and disturbs fuzzing. Switch to pr_warn for a message instead. Reported-by: syzbot+fa2d5f63ee5904a0115a(a)syzkaller.appspotmail.com Reported-by: syzbot <syzkaller(a)googlegroups.com> Signed-off-by: Laura Abbott <labbott(a)redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- drivers/staging/android/ion/ion-ioctl.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) --- a/drivers/staging/android/ion/ion-ioctl.c +++ b/drivers/staging/android/ion/ion-ioctl.c @@ -70,8 +70,10 @@ long ion_ioctl(struct file *filp, unsign return -EFAULT; ret = validate_ioctl_arg(cmd, &data); - if (WARN_ON_ONCE(ret)) + if (ret) { + pr_warn_once("%s: ioctl validate failed\n", __func__); return ret; + } if (!(dir & _IOC_WRITE)) memset(&data, 0, sizeof(data)); Patches currently in stable-queue which might be from labbott(a)redhat.com are queue-4.15/staging-android-ion-switch-from-warn-to-pr_warn.patch queue-4.15/staging-android-ion-add-__gfp_nowarn-for-system-contig-heap.patch

6 years, 10 months

1
0
0 0

Patch "selinux: skip bounded transition processing if the policy isn't loaded" has been added to the 4.15-stable tree

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled selinux: skip bounded transition processing if the policy isn't loaded to the 4.15-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… The filename of the patch is: selinux-skip-bounded-transition-processing-if-the-policy-isn-t-loaded.patch and it can be found in the queue-4.15 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable(a)vger.kernel.org> know about it. >From 4b14752ec4e0d87126e636384cf37c8dd9df157c Mon Sep 17 00:00:00 2001 From: Paul Moore <paul(a)paul-moore.com> Date: Tue, 5 Dec 2017 17:17:43 -0500 Subject: selinux: skip bounded transition processing if the policy isn't loaded From: Paul Moore <paul(a)paul-moore.com> commit 4b14752ec4e0d87126e636384cf37c8dd9df157c upstream. We can't do anything reasonable in security_bounded_transition() if we don't have a policy loaded, and in fact we could run into problems with some of the code inside expecting a policy. Fix these problems like we do many others in security/selinux/ss/services.c by checking to see if the policy is loaded (ss_initialized) and returning quickly if it isn't. Reported-by: syzbot <syzkaller-bugs(a)googlegroups.com> Signed-off-by: Paul Moore <paul(a)paul-moore.com> Acked-by: Stephen Smalley <sds(a)tycho.nsa.gov> Reviewed-by: James Morris <james.l.morris(a)oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- security/selinux/ss/services.c | 3 +++ 1 file changed, 3 insertions(+) --- a/security/selinux/ss/services.c +++ b/security/selinux/ss/services.c @@ -867,6 +867,9 @@ int security_bounded_transition(u32 old_ int index; int rc; + if (!ss_initialized) + return 0; + read_lock(&policy_rwlock); rc = -EINVAL; Patches currently in stable-queue which might be from paul(a)paul-moore.com are queue-4.15/selinux-skip-bounded-transition-processing-if-the-policy-isn-t-loaded.patch queue-4.15/selinux-ensure-the-context-is-nul-terminated-in-security_context_to_sid_core.patch

6 years, 10 months

1
0
0 0

Patch "staging: android: ion: Add __GFP_NOWARN for system contig heap" has been added to the 4.15-stable tree

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled staging: android: ion: Add __GFP_NOWARN for system contig heap to the 4.15-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… The filename of the patch is: staging-android-ion-add-__gfp_nowarn-for-system-contig-heap.patch and it can be found in the queue-4.15 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable(a)vger.kernel.org> know about it. >From 0c75f10312a35b149b2cebb1832316b35c2337ca Mon Sep 17 00:00:00 2001 From: Laura Abbott <labbott(a)redhat.com> Date: Fri, 5 Jan 2018 11:14:08 -0800 Subject: staging: android: ion: Add __GFP_NOWARN for system contig heap From: Laura Abbott <labbott(a)redhat.com> commit 0c75f10312a35b149b2cebb1832316b35c2337ca upstream. syzbot reported a warning from Ion: WARNING: CPU: 1 PID: 3485 at mm/page_alloc.c:3926 ... __alloc_pages_nodemask+0x9fb/0xd80 mm/page_alloc.c:4252 alloc_pages_current+0xb6/0x1e0 mm/mempolicy.c:2036 alloc_pages include/linux/gfp.h:492 [inline] ion_system_contig_heap_allocate+0x40/0x2c0 drivers/staging/android/ion/ion_system_heap.c:374 ion_buffer_create drivers/staging/android/ion/ion.c:93 [inline] ion_alloc+0x2c1/0x9e0 drivers/staging/android/ion/ion.c:420 ion_ioctl+0x26d/0x380 drivers/staging/android/ion/ion-ioctl.c:84 vfs_ioctl fs/ioctl.c:46 [inline] do_vfs_ioctl+0x1b1/0x1520 fs/ioctl.c:686 SYSC_ioctl fs/ioctl.c:701 [inline] SyS_ioctl+0x8f/0xc0 fs/ioctl.c:692 This is a warning about attempting to allocate order > MAX_ORDER. This is coming from a userspace Ion allocation request. Since userspace is free to request however much memory it wants (and the kernel is free to deny its allocation), silence the allocation attempt with __GFP_NOWARN in case it fails. Reported-by: syzbot+76e7efc4748495855a4d(a)syzkaller.appspotmail.com Reported-by: syzbot <syzkaller(a)googlegroups.com> Signed-off-by: Laura Abbott <labbott(a)redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- drivers/staging/android/ion/ion_system_heap.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) --- a/drivers/staging/android/ion/ion_system_heap.c +++ b/drivers/staging/android/ion/ion_system_heap.c @@ -371,7 +371,7 @@ static int ion_system_contig_heap_alloca unsigned long i; int ret; - page = alloc_pages(low_order_gfp_flags, order); + page = alloc_pages(low_order_gfp_flags | __GFP_NOWARN, order); if (!page) return -ENOMEM; Patches currently in stable-queue which might be from labbott(a)redhat.com are queue-4.15/staging-android-ion-switch-from-warn-to-pr_warn.patch queue-4.15/staging-android-ion-add-__gfp_nowarn-for-system-contig-heap.patch

6 years, 10 months

1
0
0 0

Patch "selinux: ensure the context is NUL terminated in security_context_to_sid_core()" has been added to the 4.15-stable tree

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled selinux: ensure the context is NUL terminated in security_context_to_sid_core() to the 4.15-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… The filename of the patch is: selinux-ensure-the-context-is-nul-terminated-in-security_context_to_sid_core.patch and it can be found in the queue-4.15 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable(a)vger.kernel.org> know about it. >From ef28df55ac27e1e5cd122e19fa311d886d47a756 Mon Sep 17 00:00:00 2001 From: Paul Moore <paul(a)paul-moore.com> Date: Tue, 28 Nov 2017 18:51:12 -0500 Subject: selinux: ensure the context is NUL terminated in security_context_to_sid_core() From: Paul Moore <paul(a)paul-moore.com> commit ef28df55ac27e1e5cd122e19fa311d886d47a756 upstream. The syzbot/syzkaller automated tests found a problem in security_context_to_sid_core() during early boot (before we load the SELinux policy) where we could potentially feed context strings without NUL terminators into the strcmp() function. We already guard against this during normal operation (after the SELinux policy has been loaded) by making a copy of the context strings and explicitly adding a NUL terminator to the end. The patch extends this protection to the early boot case (no loaded policy) by moving the context copy earlier in security_context_to_sid_core(). Reported-by: syzbot <syzkaller(a)googlegroups.com> Signed-off-by: Paul Moore <paul(a)paul-moore.com> Reviewed-By: William Roberts <william.c.roberts(a)intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- security/selinux/ss/services.c | 18 ++++++++---------- 1 file changed, 8 insertions(+), 10 deletions(-) --- a/security/selinux/ss/services.c +++ b/security/selinux/ss/services.c @@ -1413,27 +1413,25 @@ static int security_context_to_sid_core( if (!scontext_len) return -EINVAL; + /* Copy the string to allow changes and ensure a NUL terminator */ + scontext2 = kmemdup_nul(scontext, scontext_len, gfp_flags); + if (!scontext2) + return -ENOMEM; + if (!ss_initialized) { int i; for (i = 1; i < SECINITSID_NUM; i++) { - if (!strcmp(initial_sid_to_string[i], scontext)) { + if (!strcmp(initial_sid_to_string[i], scontext2)) { *sid = i; - return 0; + goto out; } } *sid = SECINITSID_KERNEL; - return 0; + goto out; } *sid = SECSID_NULL; - /* Copy the string so that we can modify the copy as we parse it. */ - scontext2 = kmalloc(scontext_len + 1, gfp_flags); - if (!scontext2) - return -ENOMEM; - memcpy(scontext2, scontext, scontext_len); - scontext2[scontext_len] = 0; - if (force) { /* Save another copy for storing in uninterpreted form */ rc = -ENOMEM; Patches currently in stable-queue which might be from paul(a)paul-moore.com are queue-4.15/selinux-skip-bounded-transition-processing-if-the-policy-isn-t-loaded.patch queue-4.15/selinux-ensure-the-context-is-nul-terminated-in-security_context_to_sid_core.patch

6 years, 10 months

1
0
0 0

Patch "rds: tcp: correctly sequence cleanup on netns deletion." has been added to the 4.15-stable tree

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled rds: tcp: correctly sequence cleanup on netns deletion. to the 4.15-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… The filename of the patch is: rds-tcp-correctly-sequence-cleanup-on-netns-deletion.patch and it can be found in the queue-4.15 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable(a)vger.kernel.org> know about it. >From 681648e67d43cf269c5590ecf021ed481f4551fc Mon Sep 17 00:00:00 2001 From: Sowmini Varadhan <sowmini.varadhan(a)oracle.com> Date: Thu, 30 Nov 2017 11:11:28 -0800 Subject: rds: tcp: correctly sequence cleanup on netns deletion. From: Sowmini Varadhan <sowmini.varadhan(a)oracle.com> commit 681648e67d43cf269c5590ecf021ed481f4551fc upstream. Commit 8edc3affc077 ("rds: tcp: Take explicit refcounts on struct net") introduces a regression in rds-tcp netns cleanup. The cleanup_net(), (and thus rds_tcp_dev_event notification) is only called from put_net() when all netns refcounts go to 0, but this cannot happen if the rds_connection itself is holding a c_net ref that it expects to release in rds_tcp_kill_sock. Instead, the rds_tcp_kill_sock callback should make sure to tear down state carefully, ensuring that the socket teardown is only done after all data-structures and workqs that depend on it are quiesced. The original motivation for commit 8edc3affc077 ("rds: tcp: Take explicit refcounts on struct net") was to resolve a race condition reported by syzkaller where workqs for tx/rx/connect were triggered after the namespace was deleted. Those worker threads should have been cancelled/flushed before socket tear-down and indeed, rds_conn_path_destroy() does try to sequence this by doing /* cancel cp_send_w */ /* cancel cp_recv_w */ /* flush cp_down_w */ /* free data structures */ Here the "flush cp_down_w" will trigger rds_conn_shutdown and thus invoke rds_tcp_conn_path_shutdown() to close the tcp socket, so that we ought to have satisfied the requirement that "socket-close is done after all other dependent state is quiesced". However, rds_conn_shutdown has a bug in that it *always* triggers the reconnect workq (and if connection is successful, we always restart tx/rx workqs so with the right timing, we risk the race conditions reported by syzkaller). Netns deletion is like module teardown- no need to restart a reconnect in this case. We can use the c_destroy_in_prog bit to avoid restarting the reconnect. Fixes: 8edc3affc077 ("rds: tcp: Take explicit refcounts on struct net") Signed-off-by: Sowmini Varadhan <sowmini.varadhan(a)oracle.com> Acked-by: Santosh Shilimkar <santosh.shilimkar(a)oracle.com> Signed-off-by: David S. Miller <davem(a)davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- net/rds/connection.c | 3 ++- net/rds/rds.h | 6 +++--- net/rds/tcp.c | 4 ++-- 3 files changed, 7 insertions(+), 6 deletions(-) --- a/net/rds/connection.c +++ b/net/rds/connection.c @@ -366,6 +366,8 @@ void rds_conn_shutdown(struct rds_conn_p * to the conn hash, so we never trigger a reconnect on this * conn - the reconnect is always triggered by the active peer. */ cancel_delayed_work_sync(&cp->cp_conn_w); + if (conn->c_destroy_in_prog) + return; rcu_read_lock(); if (!hlist_unhashed(&conn->c_hash_node)) { rcu_read_unlock(); @@ -445,7 +447,6 @@ void rds_conn_destroy(struct rds_connect */ rds_cong_remove_conn(conn); - put_net(conn->c_net); kfree(conn->c_path); kmem_cache_free(rds_conn_slab, conn); --- a/net/rds/rds.h +++ b/net/rds/rds.h @@ -150,7 +150,7 @@ struct rds_connection { /* Protocol version */ unsigned int c_version; - struct net *c_net; + possible_net_t c_net; struct list_head c_map_item; unsigned long c_map_queued; @@ -165,13 +165,13 @@ struct rds_connection { static inline struct net *rds_conn_net(struct rds_connection *conn) { - return conn->c_net; + return read_pnet(&conn->c_net); } static inline void rds_conn_net_set(struct rds_connection *conn, struct net *net) { - conn->c_net = get_net(net); + write_pnet(&conn->c_net, net); } #define RDS_FLAG_CONG_BITMAP 0x01 --- a/net/rds/tcp.c +++ b/net/rds/tcp.c @@ -528,7 +528,7 @@ static void rds_tcp_kill_sock(struct net rds_tcp_listen_stop(lsock, &rtn->rds_tcp_accept_w); spin_lock_irq(&rds_tcp_conn_lock); list_for_each_entry_safe(tc, _tc, &rds_tcp_conn_list, t_tcp_node) { - struct net *c_net = tc->t_cpath->cp_conn->c_net; + struct net *c_net = read_pnet(&tc->t_cpath->cp_conn->c_net); if (net != c_net || !tc->t_sock) continue; @@ -587,7 +587,7 @@ static void rds_tcp_sysctl_reset(struct spin_lock_irq(&rds_tcp_conn_lock); list_for_each_entry_safe(tc, _tc, &rds_tcp_conn_list, t_tcp_node) { - struct net *c_net = tc->t_cpath->cp_conn->c_net; + struct net *c_net = read_pnet(&tc->t_cpath->cp_conn->c_net); if (net != c_net || !tc->t_sock) continue; Patches currently in stable-queue which might be from sowmini.varadhan(a)oracle.com are queue-4.15/rds-tcp-atomically-purge-entries-from-rds_tcp_conn_list-during-netns-delete.patch queue-4.15/rds-tcp-correctly-sequence-cleanup-on-netns-deletion.patch

6 years, 10 months

1
0
0 0

Patch "rds: tcp: atomically purge entries from rds_tcp_conn_list during netns delete" has been added to the 4.15-stable tree

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled rds: tcp: atomically purge entries from rds_tcp_conn_list during netns delete to the 4.15-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… The filename of the patch is: rds-tcp-atomically-purge-entries-from-rds_tcp_conn_list-during-netns-delete.patch and it can be found in the queue-4.15 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable(a)vger.kernel.org> know about it. >From f10b4cff98c6977668434fbf5dd58695eeca2897 Mon Sep 17 00:00:00 2001 From: Sowmini Varadhan <sowmini.varadhan(a)oracle.com> Date: Thu, 30 Nov 2017 11:11:29 -0800 Subject: rds: tcp: atomically purge entries from rds_tcp_conn_list during netns delete From: Sowmini Varadhan <sowmini.varadhan(a)oracle.com> commit f10b4cff98c6977668434fbf5dd58695eeca2897 upstream. The rds_tcp_kill_sock() function parses the rds_tcp_conn_list to find the rds_connection entries marked for deletion as part of the netns deletion under the protection of the rds_tcp_conn_lock. Since the rds_tcp_conn_list tracks rds_tcp_connections (which have a 1:1 mapping with rds_conn_path), multiple tc entries in the rds_tcp_conn_list will map to a single rds_connection, and will be deleted as part of the rds_conn_destroy() operation that is done outside the rds_tcp_conn_lock. The rds_tcp_conn_list traversal done under the protection of rds_tcp_conn_lock should not leave any doomed tc entries in the list after the rds_tcp_conn_lock is released, else another concurrently executiong netns delete (for a differnt netns) thread may trip on these entries. Reported-by: syzbot <syzkaller(a)googlegroups.com> Signed-off-by: Sowmini Varadhan <sowmini.varadhan(a)oracle.com> Acked-by: Santosh Shilimkar <santosh.shilimkar(a)oracle.com> Signed-off-by: David S. Miller <davem(a)davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- net/rds/tcp.c | 9 +++++++-- net/rds/tcp.h | 1 + 2 files changed, 8 insertions(+), 2 deletions(-) --- a/net/rds/tcp.c +++ b/net/rds/tcp.c @@ -307,7 +307,8 @@ static void rds_tcp_conn_free(void *arg) rdsdebug("freeing tc %p\n", tc); spin_lock_irqsave(&rds_tcp_conn_lock, flags); - list_del(&tc->t_tcp_node); + if (!tc->t_tcp_node_detached) + list_del(&tc->t_tcp_node); spin_unlock_irqrestore(&rds_tcp_conn_lock, flags); kmem_cache_free(rds_tcp_conn_slab, tc); @@ -532,8 +533,12 @@ static void rds_tcp_kill_sock(struct net if (net != c_net || !tc->t_sock) continue; - if (!list_has_conn(&tmp_list, tc->t_cpath->cp_conn)) + if (!list_has_conn(&tmp_list, tc->t_cpath->cp_conn)) { list_move_tail(&tc->t_tcp_node, &tmp_list); + } else { + list_del(&tc->t_tcp_node); + tc->t_tcp_node_detached = true; + } } spin_unlock_irq(&rds_tcp_conn_lock); list_for_each_entry_safe(tc, _tc, &tmp_list, t_tcp_node) { --- a/net/rds/tcp.h +++ b/net/rds/tcp.h @@ -12,6 +12,7 @@ struct rds_tcp_incoming { struct rds_tcp_connection { struct list_head t_tcp_node; + bool t_tcp_node_detached; struct rds_conn_path *t_cpath; /* t_conn_path_lock synchronizes the connection establishment between * rds_tcp_accept_one and rds_tcp_conn_path_connect Patches currently in stable-queue which might be from sowmini.varadhan(a)oracle.com are queue-4.15/rds-tcp-atomically-purge-entries-from-rds_tcp_conn_list-during-netns-delete.patch queue-4.15/rds-tcp-correctly-sequence-cleanup-on-netns-deletion.patch

6 years, 10 months

1
0
0 0

Patch "ptr_ring: fail early if queue occupies more than KMALLOC_MAX_SIZE" has been added to the 4.15-stable tree

by gregkh＠linuxfoundation.org

This is a note to let you know that I've just added the patch titled ptr_ring: fail early if queue occupies more than KMALLOC_MAX_SIZE to the 4.15-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=sum… The filename of the patch is: ptr_ring-fail-early-if-queue-occupies-more-than-kmalloc_max_size.patch and it can be found in the queue-4.15 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable(a)vger.kernel.org> know about it. >From 6e6e41c3112276288ccaf80c70916779b84bb276 Mon Sep 17 00:00:00 2001 From: Jason Wang <jasowang(a)redhat.com> Date: Fri, 9 Feb 2018 17:45:49 +0800 Subject: ptr_ring: fail early if queue occupies more than KMALLOC_MAX_SIZE From: Jason Wang <jasowang(a)redhat.com> commit 6e6e41c3112276288ccaf80c70916779b84bb276 upstream. To avoid slab to warn about exceeded size, fail early if queue occupies more than KMALLOC_MAX_SIZE. Reported-by: syzbot+e4d4f9ddd4295539735d(a)syzkaller.appspotmail.com Fixes: 2e0ab8ca83c12 ("ptr_ring: array based FIFO for pointers") Signed-off-by: Jason Wang <jasowang(a)redhat.com> Acked-by: Michael S. Tsirkin <mst(a)redhat.com> Signed-off-by: David S. Miller <davem(a)davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> --- include/linux/ptr_ring.h | 2 ++ 1 file changed, 2 insertions(+) --- a/include/linux/ptr_ring.h +++ b/include/linux/ptr_ring.h @@ -453,6 +453,8 @@ static inline int ptr_ring_consume_batch static inline void **__ptr_ring_init_queue_alloc(unsigned int size, gfp_t gfp) { + if (size * sizeof(void *) > KMALLOC_MAX_SIZE) + return NULL; return kcalloc(size, sizeof(void *), gfp); } Patches currently in stable-queue which might be from jasowang(a)redhat.com are queue-4.15/ptr_ring-try-vmalloc-when-kmalloc-fails.patch queue-4.15/ptr_ring-fail-early-if-queue-occupies-more-than-kmalloc_max_size.patch

6 years, 10 months

1
0
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror February 2018