commit a13b67c9a0 upstream.
Christian Theune says:
I upgraded from 6.1.38 to 6.1.55 this morning and it broke my traffic shaping script,
leaving me with a non-functional uplink on a remote router.
A 'rt' curve cannot be used as a inner curve (parent class), but we were
allowing such configurations since the qdisc was introduced. Such
configurations would trigger a UAF as Budimir explains:
The parent will have vttree_insert() called on it in init_vf(),
but will not have vttree_remove() called on it in update_vf()
because it does not have the HFSC_FSC flag set.
The qdisc always assumes that inner classes have the HFSC_FSC flag set.
This is by design as it doesn't make sense 'qdisc wise' for an 'rt'
curve to be an inner curve.
Budimir's original patch disallows users to add classes with a 'rt'
parent, but this is too strict as it breaks users that have been using
'rt' as a inner class. Another approach, taken by this patch, is to
upgrade the inner 'rt' into a 'sc', warning the user in the process.
It avoids the UAF reported by Budimir while also being more permissive
to bad scripts/users/code using 'rt' as a inner class.
Users checking the `tc class ls [...]` or `tc class get [...]` dumps would
observe the curve change and are potentially breaking with this change.
v1->v2: https://lore.kernel.org/all/20231013151057.2611860-1-pctammela@mojatatu.com/
- Correct 'Fixes' tag and merge with revert (Jakub)
Cc: Christian Theune <ct@flyingcircus.io>
Cc: Budimir Markovic <markovicbudimir@gmail.com>
Fixes: b3d26c5702 ("net/sched: sch_hfsc: Ensure inner classes have fsc curve")
Signed-off-by: Pedro Tammela <pctammela@mojatatu.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Link: https://lore.kernel.org/r/20231017143602.3191556-1-pctammela@mojatatu.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit d93f3f9927 upstream.
Since 429e3d123d ("bonding: Fix extraction of ports from the packet
headers"), header offsets used to compute a hash in bond_xmit_hash() are
relative to skb->data and not skb->head. If the tail of the header buffer
of an skb really needs to be advanced and the operation is successful, the
pointer to the data must be returned (and not a pointer to the head of the
buffer).
Fixes: 429e3d123d ("bonding: Fix extraction of ports from the packet headers")
Signed-off-by: Jiri Wiesner <jwiesner@suse.de>
Acked-by: Jay Vosburgh <jay.vosburgh@canonical.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 61b40cefe5 upstream.
In bcm_sf2_mdio_register(), the class_find_device() will call get_device()
to increment reference count for priv->master_mii_bus->dev if
of_mdio_find_bus() succeeds. If mdiobus_alloc() or mdiobus_register()
fails, it will call get_device() twice without decrement reference count
for the device. And it is the same if bcm_sf2_mdio_register() succeeds but
fails in bcm_sf2_sw_probe(), or if bcm_sf2_sw_probe() succeeds. If the
reference count has not decremented to zero, the dev related resource will
not be freed.
So remove the get_device() in bcm_sf2_mdio_register(), and call
put_device() if mdiobus_alloc() or mdiobus_register() fails and in
bcm_sf2_mdio_unregister() to solve the issue.
And as Simon suggested, unwind from errors for bcm_sf2_mdio_register() and
just return 0 if it succeeds to make it cleaner.
Fixes: 461cd1b03e ("net: dsa: bcm_sf2: Register our slave MDIO bus")
Signed-off-by: Jinjie Ruan <ruanjinjie@huawei.com>
Suggested-by: Simon Horman <horms@kernel.org>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Link: https://lore.kernel.org/r/20231011032419.2423290-1-ruanjinjie@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit fc6f716a50 upstream.
The hardware provides the indexes of the first and the last available
queue and VF. From the indexes, the driver calculates the numbers of
queues and VFs. In theory, a faulty device might say the last index is
smaller than the first index. In that case, the driver's calculation
would underflow, it would attempt to write to non-existent registers
outside of the ioremapped range and crash.
I ran into this not by having a faulty device, but by an operator error.
I accidentally ran a QE test meant for i40e devices on an ice device.
The test used 'echo i40e > /sys/...ice PCI device.../driver_override',
bound the driver to the device and crashed in one of the wr32 calls in
i40e_clear_hw.
Add checks to prevent underflows in the calculations of num_queues and
num_vfs. With this fix, the wrong device probing reports errors and
returns a failure without crashing.
Fixes: 838d41d92a ("i40e: clear all queues and interrupts")
Signed-off-by: Michal Schmidt <mschmidt@redhat.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel)
Link: https://lore.kernel.org/r/20231011233334.336092-2-jacob.e.keller@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit a0ca6b9dfe upstream.
Sometimes Tx is completed immediately after doorbell is updated, which
causes Tx completion routing to update completion bytes before the
same packet bytes are updated in sent bytes in transmit function, hence
hitting BUG_ON() in dql_completed(). To avoid this, update BQL
sent bytes before ringing doorbell.
Fixes: 37d79d0596 ("octeon_ep: add Tx/Rx processing and interrupt support")
Signed-off-by: Shinas Rasheed <srasheed@marvell.com>
Link: https://lore.kernel.org/r/20231017105030.2310966-1-srasheed@marvell.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 8702cf12e6 upstream.
syzbot reported a warning [0] introduced by commit c48ef9c4ae ("tcp: Fix
bind() regression for v4-mapped-v6 non-wildcard address.").
After the cited commit, a v4 socket's address matches the corresponding
v4-mapped-v6 tb2 in inet_bind2_bucket_match_addr(), not vice versa.
During X.X.X.X -> ::ffff:X.X.X.X order bind()s, the second bind() uses
bhash and conflicts properly without checking bhash2 so that we need not
check if a v4-mapped-v6 sk matches the corresponding v4 address tb2 in
inet_bind2_bucket_match_addr(). However, the repro shows that we need
to check that in a no-conflict case.
The repro bind()s two sockets to the 2-tuples using SO_REUSEPORT and calls
listen() for the first socket:
from socket import *
s1 = socket()
s1.setsockopt(SOL_SOCKET, SO_REUSEPORT, 1)
s1.bind(('127.0.0.1', 0))
s2 = socket(AF_INET6)
s2.setsockopt(SOL_SOCKET, SO_REUSEPORT, 1)
s2.bind(('::ffff:127.0.0.1', s1.getsockname()[1]))
s1.listen()
The second socket should belong to the first socket's tb2, but the second
bind() creates another tb2 bucket because inet_bind2_bucket_find() returns
NULL in inet_csk_get_port() as the v4-mapped-v6 sk does not match the
corresponding v4 address tb2.
bhash2[] -> tb2(::ffff:X.X.X.X) -> tb2(X.X.X.X)
Then, listen() for the first socket calls inet_csk_get_port(), where the
v4 address matches the v4-mapped-v6 tb2 and WARN_ON() is triggered.
To avoid that, we need to check if v4-mapped-v6 sk address matches with
the corresponding v4 address tb2 in inet_bind2_bucket_match().
The same checks are needed in inet_bind2_bucket_addr_match() too, so we
can move all checks there and call it from inet_bind2_bucket_match().
Note that now tb->family is just an address family of tb->(v6_)?rcv_saddr
and not of sockets in the bucket. This could be refactored later by
defining tb->rcv_saddr as tb->v6_rcv_saddr.s6_addr32[3] and prepending
::ffff: when creating v4 tb2.
[0]:
WARNING: CPU: 0 PID: 5049 at net/ipv4/inet_connection_sock.c:587 inet_csk_get_port+0xf96/0x2350 net/ipv4/inet_connection_sock.c:587
Modules linked in:
CPU: 0 PID: 5049 Comm: syz-executor288 Not tainted 6.6.0-rc2-syzkaller-00018-g2cf0f7156238 #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 08/04/2023
RIP: 0010:inet_csk_get_port+0xf96/0x2350 net/ipv4/inet_connection_sock.c:587
Code: 7c 24 08 e8 4c b6 8a 01 31 d2 be 88 01 00 00 48 c7 c7 e0 94 ae 8b e8 59 2e a3 f8 2e 2e 2e 31 c0 e9 04 fe ff ff e8 ca 88 d0 f8 <0f> 0b e9 0f f9 ff ff e8 be 88 d0 f8 49 8d 7e 48 e8 65 ca 5a 00 31
RSP: 0018:ffffc90003abfbf0 EFLAGS: 00010293
RAX: 0000000000000000 RBX: ffff888026429100 RCX: 0000000000000000
RDX: ffff88807edcbb80 RSI: ffffffff88b73d66 RDI: ffff888026c49f38
RBP: ffff888026c49f30 R08: 0000000000000005 R09: 0000000000000000
R10: 0000000000000001 R11: 0000000000000000 R12: ffffffff9260f200
R13: ffff888026c49880 R14: 0000000000000000 R15: ffff888026429100
FS: 00005555557d5380(0000) GS:ffff8880b9800000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 000000000045ad50 CR3: 0000000025754000 CR4: 00000000003506f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
<TASK>
inet_csk_listen_start+0x155/0x360 net/ipv4/inet_connection_sock.c:1256
__inet_listen_sk+0x1b8/0x5c0 net/ipv4/af_inet.c:217
inet_listen+0x93/0xd0 net/ipv4/af_inet.c:239
__sys_listen+0x194/0x270 net/socket.c:1866
__do_sys_listen net/socket.c:1875 [inline]
__se_sys_listen net/socket.c:1873 [inline]
__x64_sys_listen+0x53/0x80 net/socket.c:1873
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x38/0xb0 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x63/0xcd
RIP: 0033:0x7f3a5bce3af9
Code: 28 00 00 00 75 05 48 83 c4 28 c3 e8 c1 17 00 00 90 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 b8 ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007ffc1a1c79e8 EFLAGS: 00000246 ORIG_RAX: 0000000000000032
RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007f3a5bce3af9
RDX: 00007f3a5bce3af9 RSI: 0000000000000000 RDI: 0000000000000003
RBP: 00007f3a5bd565f0 R08: 0000000000000006 R09: 0000000000000006
R10: 0000000000000006 R11: 0000000000000246 R12: 0000000000000001
R13: 431bde82d7b634db R14: 0000000000000001 R15: 0000000000000001
</TASK>
Fixes: c48ef9c4ae ("tcp: Fix bind() regression for v4-mapped-v6 non-wildcard address.")
Reported-by: syzbot+71e724675ba3958edb31@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=71e724675ba3958edb31
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20231010013814.70571-1-kuniyu@amazon.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit f921a4a5bf upstream.
In commit 75eefc6c59 ("tcp: tsq: add a shortcut in tcp_small_queue_check()")
we allowed to send an skb regardless of TSQ limits being hit if rtx queue
was empty or had a single skb, in order to better fill the pipe
when/if TX completions were slow.
Then later, commit 75c119afe1 ("tcp: implement rb-tree based
retransmit queue") accidentally removed the special case for
one skb in rtx queue.
Stefan Wahren reported a regression in single TCP flow throughput
using a 100Mbit fec link, starting from commit 65466904b0 ("tcp: adjust
TSO packet sizes based on min_rtt"). This last commit only made the
regression more visible, because it locked the TCP flow on a particular
behavior where TSQ prevented two skbs being pushed downstream,
adding silences on the wire between each TSO packet.
Many thanks to Stefan for his invaluable help !
Fixes: 75c119afe1 ("tcp: implement rb-tree based retransmit queue")
Link: https://lore.kernel.org/netdev/7f31ddc8-9971-495e-a1f6-819df542e0af@gmx.net/
Reported-by: Stefan Wahren <wahrenst@gmx.net>
Tested-by: Stefan Wahren <wahrenst@gmx.net>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Link: https://lore.kernel.org/r/20231017124526.4060202-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 1c2709cfff upstream.
We discovered from packet traces of slow loss recovery on kernels with
the default HZ=250 setting (and min_rtt < 1ms) that after reordering,
when receiving a SACKed sequence range, the RACK reordering timer was
firing after about 16ms rather than the desired value of roughly
min_rtt/4 + 2ms. The problem is largely due to the RACK reorder timer
calculation adding in TCP_TIMEOUT_MIN, which is 2 jiffies. On kernels
with HZ=250, this is 2*4ms = 8ms. The TLP timer calculation has the
exact same issue.
This commit fixes the TLP transmit timer and RACK reordering timer
floor calculation to more closely match the intended 2ms floor even on
kernels with HZ=250. It does this by adding in a new
TCP_TIMEOUT_MIN_US floor of 2000 us and then converting to jiffies,
instead of the current approach of converting to jiffies and then
adding th TCP_TIMEOUT_MIN value of 2 jiffies.
Our testing has verified that on kernels with HZ=1000, as expected,
this does not produce significant changes in behavior, but on kernels
with the default HZ=250 the latency improvement can be large. For
example, our tests show that for HZ=250 kernels at low RTTs this fix
roughly halves the latency for the RACK reorder timer: instead of
mostly firing at 16ms it mostly fires at 8ms.
Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Fixes: bb4d991a28 ("tcp: adjust tail loss probe timeout")
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20231015174700.2206872-1-ncardwell.sw@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit b2f750c3a8 upstream.
When either reset- or shutdown-gpio have are initially deasserted,
e.g. after a reboot - or when the hardware does not include pull-down,
there will be a short toggle of both IOs to logical 0 and back to 1.
It seems that the rfkill default is unblocked, so the driver should not
glitch to output low during probe.
It can lead e.g. to unexpected lte modem reconnect:
[1] root@localhost:~# dmesg | grep "usb 2-1"
[ 2.136124] usb 2-1: new SuperSpeed USB device number 2 using xhci-hcd
[ 21.215278] usb 2-1: USB disconnect, device number 2
[ 28.833977] usb 2-1: new SuperSpeed USB device number 3 using xhci-hcd
The glitch has been discovered on an arm64 board, now that device-tree
support for the rfkill-gpio driver has finally appeared :).
Change the flags for devm_gpiod_get_optional from GPIOD_OUT_LOW to
GPIOD_ASIS to avoid any glitches.
The rfkill driver will set the intended value during rfkill_sync_work.
Fixes: 7176ba23f8 ("net: rfkill: add generic gpio rfkill driver")
Signed-off-by: Josua Mayer <josua@solid-run.com>
Link: https://lore.kernel.org/r/20231004163928.14609-1-josua@solid-run.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit dad4e491e3 upstream.
In esp_remove_trailer(), to avoid an unexpected result returned by
pskb_trim, we should check the return value of pskb_trim().
Signed-off-by: Ma Ke <make_ruc2021@163.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 513f61e219 upstream.
In esp_remove_trailer(), to avoid an unexpected result returned by
pskb_trim, we should check the return value of pskb_trim().
Signed-off-by: Ma Ke <make_ruc2021@163.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 91d20ab9d9 upstream.
Since wiphy work items can run pretty much arbitrary
code in the stack/driver, it can take longer to run
all of this, so we shouldn't be using system_wq via
schedule_work(). Also, we lock the wiphy (which is
the reason this exists), so use system_unbound_wq.
Reported-and-tested-by: Kalle Valo <kvalo@kernel.org>
Fixes: a3ee4dc84c ("wifi: cfg80211: add a work abstraction with special semantics")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 2f3389c738 upstream.
Driver allocates the LL2 rx buffers from kmalloc()
area to construct the skb using slab_build_skb()
The required size allocation seems to have overlooked
for accounting both skb_shared_info size and device
placement padding bytes which results into the below
panic when doing skb_put() for a standard MTU sized frame.
skbuff: skb_over_panic: text:ffffffffc0b0225f len:1514 put:1514
head:ff3dabceaf39c000 data:ff3dabceaf39c042 tail:0x62c end:0x566
dev:<NULL>
…
skb_panic+0x48/0x4a
skb_put.cold+0x10/0x10
qed_ll2b_complete_rx_packet+0x14f/0x260 [qed]
qed_ll2_rxq_handle_completion.constprop.0+0x169/0x200 [qed]
qed_ll2_rxq_completion+0xba/0x320 [qed]
qed_int_sp_dpc+0x1a7/0x1e0 [qed]
This patch fixes this by accouting skb_shared_info and device
placement padding size bytes when allocating the buffers.
Cc: David S. Miller <davem@davemloft.net>
Fixes: 0a7fb11c23 ("qed: Add Light L2 support")
Signed-off-by: Manish Chopra <manishc@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit da29b94ed3 upstream.
Add the missing code to release resources on bind errors, including the
references taken by wcd938x_sdw_device_get() which also need to be
dropped on unbind().
Fixes: 16572522ae ("ASoC: codecs: wcd938x-sdw: add SoundWire driver")
Cc: stable@vger.kernel.org # 5.14
Cc: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Signed-off-by: Johan Hovold <johan+linaro@kernel.org>
Link: https://lore.kernel.org/r/20231003155558.27079-4-johan+linaro@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit fa2f8a991b upstream.
Make sure to deregister the component before tearing down the resources
it depends on during unbind().
Fixes: 16572522ae ("ASoC: codecs: wcd938x-sdw: add SoundWire driver")
Cc: stable@vger.kernel.org # 5.14
Cc: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Signed-off-by: Johan Hovold <johan+linaro@kernel.org>
Link: https://lore.kernel.org/r/20231003155558.27079-3-johan+linaro@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit c5c0383082 upstream.
Make sure to balance the runtime PM operations, including the disable
count, on probe errors and on driver unbind.
Fixes: 16572522ae ("ASoC: codecs: wcd938x-sdw: add SoundWire driver")
Cc: stable@vger.kernel.org # 5.14
Cc: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Signed-off-by: Johan Hovold <johan+linaro@kernel.org>
Link: https://lore.kernel.org/r/20231003155558.27079-8-johan+linaro@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit f0dfdcbe70 upstream.
Make sure to deregister the component when the driver is being unbound
and before the underlying device-managed resources are freed.
Fixes: 16572522ae ("ASoC: codecs: wcd938x-sdw: add SoundWire driver")
Cc: stable@vger.kernel.org # 5.14
Cc: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Signed-off-by: Johan Hovold <johan+linaro@kernel.org>
Link: https://lore.kernel.org/r/20231003155558.27079-7-johan+linaro@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit c8c0a03ec1 upstream.
ASUS platform Headset Mic was disable by default.
Assigned verb table for Mic pin will enable it.
Signed-off-by: Kailang Yang <kailang@realtek.com>
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/1155d914c20c40569f56d36c79254879@realtek.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit dcc583c225 upstream.
The MediaTek DRM driver implements GEM PRIME vmap by fetching the
sg_table for the object, iterating through the pages, and then
vmapping them. In essence, unlike the GEM DMA helpers which vmap
when the object is first created or imported, the MediaTek version
does it on request.
Unfortunately, the code never correctly frees the sg_table contents.
This results in a kernel memory leak. On a Hayato device with a text
console on the internal display, this results in the system running
out of memory in a few days from all the console screen cursor updates.
Add sg_free_table() to correctly free the contents of the sg_table. This
was missing despite explicitly required by mtk_gem_prime_get_sg_table().
Also move the "out" shortcut label to after the kfree() call for the
sg_table. Having sg_free_table() together with kfree() makes more sense.
The shortcut is only used when the object already has a kernel address,
in which case the pointer is NULL and kfree() does nothing. Hence this
change causes no functional change.
Fixes: 3df64d7b0a ("drm/mediatek: Implement gem prime vmap/vunmap function")
Cc: <stable@vger.kernel.org>
Signed-off-by: Chen-Yu Tsai <wenst@chromium.org>
Reviewed-by: CK Hu <ck.hu@mediatek.com>
Link: https://patchwork.kernel.org/project/dri-devel/patch/20231004083226.1940055-1-wenst@chromium.org/
Signed-off-by: Chun-Kuang Hu <chunkuang.hu@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit e339c6d628 upstream.
If we can't find a free fence register to handle a fault in the GMADR
range just return VM_FAULT_NOPAGE without populating the PTE so that
userspace will retry the access and trigger another fault. Eventually
we should find a free fence and the fault will get properly handled.
A further improvement idea might be to reserve a fence (or one per CPU?)
for the express purpose of handling faults without having to retry. But
that would require some additional work.
Looks like this may have gotten broken originally by
commit 39965b3766 ("drm/i915: don't trash the gtt when running out of fences")
as that changed the errno to -EDEADLK which wasn't handle by the gtt
fault code either. But later in commit 2feeb52859 ("drm/i915/gt: Fix
-EDEADLK handling regression") I changed it again to -ENOBUFS as -EDEADLK
was now getting used for the ww mutex dance. So this fix only makes
sense after that last commit.
Cc: stable@vger.kernel.org
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/9479
Fixes: 2feeb52859 ("drm/i915/gt: Fix -EDEADLK handling regression")
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20231012132801.16292-1-ville.syrjala@linux.intel.com
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
(cherry picked from commit 7f403caabe)
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit d920abd1e7 upstream.
From Alon:
"Due to a logical bug in the NVMe-oF/TCP subsystem in the Linux kernel,
a malicious user can cause a UAF and a double free, which may lead to
RCE (may also lead to an LPE in case the attacker already has local
privileges)."
Hence, when a queue initialization fails after the ahash requests are
allocated, it is guaranteed that the queue removal async work will be
called, hence leave the deallocation to the queue removal.
Also, be extra careful not to continue processing the socket, so set
queue rcv_state to NVMET_TCP_RECV_ERR upon a socket error.
Cc: stable@vger.kernel.org
Reported-by: Alon Zahavi <zahavi.alon@gmail.com>
Tested-by: Alon Zahavi <zahavi.alon@gmail.com>
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit d351c1ea2d upstream.
mcast packets get looped back to the local machine.
Such packets have a 0-length mac header, we should treat
this like "mac header not set" and abort rule evaluation.
As-is, we just copy data from the network header instead.
Fixes: 96518518cc ("netfilter: add nftables")
Reported-by: Blažej Krajňák <krajnak@levonet.sk>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 8e7e27b2ee upstream.
Here is a BUG report about linux-6.1 from syzbot, but it still remains
within upstream:
BUG: KASAN: slab-out-of-bounds in ntfs_list_ea fs/ntfs3/xattr.c:191 [inline]
BUG: KASAN: slab-out-of-bounds in ntfs_listxattr+0x401/0x570 fs/ntfs3/xattr.c:710
Read of size 1 at addr ffff888021acaf3d by task syz-executor128/3632
Call Trace:
kasan_report+0x139/0x170 mm/kasan/report.c:495
ntfs_list_ea fs/ntfs3/xattr.c:191 [inline]
ntfs_listxattr+0x401/0x570 fs/ntfs3/xattr.c:710
vfs_listxattr fs/xattr.c:457 [inline]
listxattr+0x293/0x2d0 fs/xattr.c:804
path_listxattr fs/xattr.c:828 [inline]
__do_sys_llistxattr fs/xattr.c:846 [inline]
Before derefering field members of `ea` in unpacked_ea_size(), we need to
check whether the EA_FULL struct is located in access validate range.
Similarly, when derefering `ea->name` field member, we need to check
whethe the ea->name is located in access validate range, too.
Fixes: be71b5cba2 ("fs/ntfs3: Add attrib operations")
Reported-by: syzbot+9fcea5ef6dc4dc72d334@syzkaller.appspotmail.com
Signed-off-by: Zeng Heng <zengheng4@huawei.com>
[almaz.alexandrovich@paragon-software.com: took the ret variable out of the loop block]
Signed-off-by: Konstantin Komarov <almaz.alexandrovich@paragon-software.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 1f9b94af92 upstream.
Upon investigation of the C reproducer provided by Syzbot, it seemed
the reproducer was trying to mount a corrupted NTFS filesystem, then
issue a rename syscall to some nodes in the filesystem. This can be
shown by modifying the reproducer to only include the mount syscall,
and investigating the filesystem by e.g. `ls` and `rm` commands. As a
result, during the problematic call to `hdr_fine_e`, the `inode` being
supplied did not go through `indx_init`, hence the `cmp` function
pointer was never set.
The fix is simply to check whether `cmp` is not set, and return NULL
if that's the case, in order to be consistent with other error
scenarios of the `hdr_find_e` method. The rationale behind this patch
is that:
- We should prevent crashing the kernel even if the mounted filesystem
is corrupted. Any syscalls made on the filesystem could return
invalid, but the kernel should be able to sustain these calls.
- Only very specific corruption would lead to this bug, so it would be
a pretty rare case in actual usage anyways. Therefore, introducing a
check to specifically protect against this bug seems appropriate.
Because of its rarity, an `unlikely` clause is used to wrap around
this nullity check.
Reported-by: syzbot+60cf892fc31d1f4358fc@syzkaller.appspotmail.com
Signed-off-by: Ziqi Zhao <astrajoan@yahoo.com>
Signed-off-by: Konstantin Komarov <almaz.alexandrovich@paragon-software.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 6db8a37dfc upstream.
The MPTCP protocol can acquire the subflow-level socket lock and
cause the tcp backlog usage. When inserting new skbs into the
backlog, the stack will try to coalesce them.
Currently, we have no check in place to ensure that such coalescing
will respect the MPTCP-level DSS, and that may cause data stream
corruption, as reported by Christoph.
Address the issue by adding the relevant admission check for coalescing
in tcp_add_backlog().
Note the issue is not easy to reproduce, as the MPTCP protocol tries
hard to avoid acquiring the subflow-level socket lock.
Fixes: 648ef4b886 ("mptcp: Implement MPTCP receive path")
Cc: stable@vger.kernel.org
Reported-by: Christoph Paasch <cpaasch@apple.com>
Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/420
Reviewed-by: Mat Martineau <martineau@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <martineau@kernel.org>
Link: https://lore.kernel.org/r/20231018-send-net-20231018-v1-2-17ecb002e41d@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 03adc61eda upstream.
An io_uring openat operation can update an audit reference count
from multiple threads resulting in the call trace below.
A call to io_uring_submit() with a single openat op with a flag of
IOSQE_ASYNC results in the following reference count updates.
These first part of the system call performs two increments that do not race.
do_syscall_64()
__do_sys_io_uring_enter()
io_submit_sqes()
io_openat_prep()
__io_openat_prep()
getname()
getname_flags() /* update 1 (increment) */
__audit_getname() /* update 2 (increment) */
The openat op is queued to an io_uring worker thread which starts the
opportunity for a race. The system call exit performs one decrement.
do_syscall_64()
syscall_exit_to_user_mode()
syscall_exit_to_user_mode_prepare()
__audit_syscall_exit()
audit_reset_context()
putname() /* update 3 (decrement) */
The io_uring worker thread performs one increment and two decrements.
These updates can race with the system call decrement.
io_wqe_worker()
io_worker_handle_work()
io_wq_submit_work()
io_issue_sqe()
io_openat()
io_openat2()
do_filp_open()
path_openat()
__audit_inode() /* update 4 (increment) */
putname() /* update 5 (decrement) */
__audit_uring_exit()
audit_reset_context()
putname() /* update 6 (decrement) */
The fix is to change the refcnt member of struct audit_names
from int to atomic_t.
kernel BUG at fs/namei.c:262!
Call Trace:
...
? putname+0x68/0x70
audit_reset_context.part.0.constprop.0+0xe1/0x300
__audit_uring_exit+0xda/0x1c0
io_issue_sqe+0x1f3/0x450
? lock_timer_base+0x3b/0xd0
io_wq_submit_work+0x8d/0x2b0
? __try_to_del_timer_sync+0x67/0xa0
io_worker_handle_work+0x17c/0x2b0
io_wqe_worker+0x10a/0x350
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/lkml/MW2PR2101MB1033FFF044A258F84AEAA584F1C9A@MW2PR2101MB1033.namprd21.prod.outlook.com/
Fixes: 5bd2182d58 ("audit,io_uring,io-wq: add some basic audit support to io_uring")
Signed-off-by: Dan Clash <daclash@linux.microsoft.com>
Link: https://lore.kernel.org/r/20231012215518.GA4048@linuxonhyperv3.guj3yctzbm1etfxqx2vob5hsef.xx.internal.cloudapp.net
Reviewed-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 3fdc6087df upstream.
svm_leave_nested() similar to a nested VM exit, get the vCPU out of nested
mode and thus should end the local inhibition of AVIC on this vCPU.
Failure to do so, can lead to hangs on guest reboot.
Raise the KVM_REQ_APICV_UPDATE request to refresh the AVIC state of the
current vCPU in this case.
Fixes: f44509f849 ("KVM: x86: SVM: allow AVIC to co-exist with a nested guest running")
Cc: stable@vger.kernel.org
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20230928173354.217464-4-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 2dcf37abf9 upstream.
In later revisions of AMD's APM, there is a new 'incomplete IPI' exit code:
"Invalid IPI Vector - The vector for the specified IPI was set to an
illegal value (VEC < 16)"
Note that tests on Zen2 machine show that this VM exit doesn't happen and
instead AVIC just does nothing.
Add support for this exit code by doing nothing, instead of filling
the kernel log with errors.
Also replace an unthrottled 'pr_err()' if another unknown incomplete
IPI exit happens with vcpu_unimpl()
(e.g in case AMD adds yet another 'Invalid IPI' exit reason)
Cc: <stable@vger.kernel.org>
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20230928173354.217464-3-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 8647c52e95 upstream.
Mask off xfeatures that aren't exposed to the guest only when saving guest
state via KVM_GET_XSAVE{2} instead of modifying user_xfeatures directly.
Preserving the maximal set of xfeatures in user_xfeatures restores KVM's
ABI for KVM_SET_XSAVE, which prior to commit ad856280dd ("x86/kvm/fpu:
Limit guest user_xfeatures to supported bits of XCR0") allowed userspace
to load xfeatures that are supported by the host, irrespective of what
xfeatures are exposed to the guest.
There is no known use case where userspace *intentionally* loads xfeatures
that aren't exposed to the guest, but the bug fixed by commit ad856280dd
was specifically that KVM_GET_SAVE{2} would save xfeatures that weren't
exposed to the guest, e.g. would lead to userspace unintentionally loading
guest-unsupported xfeatures when live migrating a VM.
Restricting KVM_SET_XSAVE to guest-supported xfeatures is especially
problematic for QEMU-based setups, as QEMU has a bug where instead of
terminating the VM if KVM_SET_XSAVE fails, QEMU instead simply stops
loading guest state, i.e. resumes the guest after live migration with
incomplete guest state, and ultimately results in guest data corruption.
Note, letting userspace restore all host-supported xfeatures does not fix
setups where a VM is migrated from a host *without* commit ad856280dd,
to a target with a subset of host-supported xfeatures. However there is
no way to safely address that scenario, e.g. KVM could silently drop the
unsupported features, but that would be a clear violation of KVM's ABI and
so would require userspace to opt-in, at which point userspace could
simply be updated to sanitize the to-be-loaded XSAVE state.
Reported-by: Tyler Stachecki <stachecki.tyler@gmail.com>
Closes: https://lore.kernel.org/all/20230914010003.358162-1-tstachecki@bloomberg.net
Fixes: ad856280dd ("x86/kvm/fpu: Limit guest user_xfeatures to supported bits of XCR0")
Cc: stable@vger.kernel.org
Cc: Leonardo Bras <leobras@redhat.com>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Message-Id: <20230928001956.924301-3-seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 18164f66e6 upstream.
Plumb an xfeatures mask into __copy_xstate_to_uabi_buf() so that KVM can
constrain which xfeatures are saved into the userspace buffer without
having to modify the user_xfeatures field in KVM's guest_fpu state.
KVM's ABI for KVM_GET_XSAVE{2} is that features that are not exposed to
guest must not show up in the effective xstate_bv field of the buffer.
Saving only the guest-supported xfeatures allows userspace to load the
saved state on a different host with a fewer xfeatures, so long as the
target host supports the xfeatures that are exposed to the guest.
KVM currently sets user_xfeatures directly to restrict KVM_GET_XSAVE{2} to
the set of guest-supported xfeatures, but doing so broke KVM's historical
ABI for KVM_SET_XSAVE, which allows userspace to load any xfeatures that
are supported by the *host*.
Cc: stable@vger.kernel.org
Signed-off-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20230928001956.924301-2-seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Upstream commit: 63e44bc520
Check the memory operand of INS/OUTS before emulating the instruction.
The #VC exception can get raised from user-space, but the memory operand
can be manipulated to access kernel memory before the emulation actually
begins and after the exception handler has run.
[ bp: Massage commit message. ]
Fixes: 597cfe4821 ("x86/boot/compressed/64: Setup a GHCB-based VC Exception handler")
Reported-by: Tom Dohrmann <erbse.13@gmx.de>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Cc: <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Upstream commit: b9cb9c4558
Check the IO permission bitmap (if present) before emulating IOIO #VC
exceptions for user-space. These permissions are checked by hardware
already before the #VC is raised, but due to the VC-handler decoding
race it needs to be checked again in software.
Fixes: 25189d08e5 ("x86/sev-es: Add support for handling IOIO exceptions")
Reported-by: Tom Dohrmann <erbse.13@gmx.de>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Tested-by: Tom Dohrmann <erbse.13@gmx.de>
Cc: <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Upstream commit: a37cd2a59d
A virt scenario can be constructed where MMIO memory can be user memory.
When that happens, a race condition opens between when the hardware
raises the #VC and when the #VC handler gets to emulate the instruction.
If the MOVS is replaced with a MOVS accessing kernel memory in that
small race window, then write to kernel memory happens as the access
checks are not done at emulation time.
Disable MMIO emulation in user mode temporarily until a sensible use
case appears and justifies properly handling the race window.
Fixes: 0118b604c2 ("x86/sev-es: Handle MMIO String Instructions")
Reported-by: Tom Dohrmann <erbse.13@gmx.de>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Tested-by: Tom Dohrmann <erbse.13@gmx.de>
Cc: <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit a16eb25b09 upstream.
Per the SDM, "When the local APIC handles a performance-monitoring
counters interrupt, it automatically sets the mask flag in the LVT
performance counter register." Add this behavior to KVM's local APIC
emulation.
Failure to mask the LVTPC entry results in spurious PMIs, e.g. when
running Linux as a guest, PMI handlers that do a "late_ack" spew a large
number of "dazed and confused" spurious NMI warnings.
Fixes: f5132b0138 ("KVM: Expose a version 2 architectural PMU to a guests")
Cc: stable@vger.kernel.org
Signed-off-by: Jim Mattson <jmattson@google.com>
Tested-by: Mingwei Zhang <mizhang@google.com>
Signed-off-by: Mingwei Zhang <mizhang@google.com>
Link: https://lore.kernel.org/r/20230925173448.3518223-3-mizhang@google.com
[sean: massage changelog, correct Fixes]
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit c6df843348 upstream.
Not all regmaps have a name so make sure to check for that to avoid
dereferencing a NULL pointer when dev_get_regmap() is used to lookup a
named regmap.
Fixes: e84861fec3 ("regmap: dev_get_regmap_match(): fix string comparison")
Cc: stable@vger.kernel.org # 5.8
Cc: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Johan Hovold <johan+linaro@kernel.org>
Link: https://lore.kernel.org/r/20231006082104.16707-1-johan+linaro@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 8a540e990d upstream.
Commit f6fca3917b "btrfs: store chunk size in space-info struct"
broke data chunk allocations on non-zoned multi-device filesystems when
using default chunk_size. Commit 5da431b71d "btrfs: fix the max chunk
size and stripe length calculation" partially fixed that, and this patch
completes the fix for that case.
After commit f6fca3917b and 5da431b71d, the sequence of events for
a data chunk allocation on a non-zoned filesystem is:
1. btrfs_create_chunk calls init_alloc_chunk_ctl, which copies
space_info->chunk_size (default 10 GiB) to ctl->max_stripe_len
unmodified. Before f6fca3917b, ctl->max_stripe_len value was
1 GiB for non-zoned data chunks and not configurable.
2. btrfs_create_chunk calls gather_device_info which consumes
and produces more fields of chunk_ctl.
3. gather_device_info multiplies ctl->max_stripe_len by
ctl->dev_stripes (which is 1 in all cases except dup)
and calls find_free_dev_extent with that number as num_bytes.
4. find_free_dev_extent locates the first dev_extent hole on
a device which is at least as large as num_bytes. With default
max_chunk_size from f6fca3917b, it finds the first hole which is
longer than 10 GiB, or the largest hole if that hole is shorter
than 10 GiB. This is different from the pre-f6fca3917b4d
behavior, where num_bytes is 1 GiB, and find_free_dev_extent
may choose a different hole.
5. gather_device_info repeats step 4 with all devices to find
the first or largest dev_extent hole that can be allocated on
each device.
6. gather_device_info sorts the device list by the hole size
on each device, using total unallocated space on each device to
break ties, then returns to btrfs_create_chunk with the list.
7. btrfs_create_chunk calls decide_stripe_size_regular.
8. decide_stripe_size_regular finds the largest stripe_len that
fits across the first nr_devs device dev_extent holes that were
found by gather_device_info (and satisfies other constraints
on stripe_len that are not relevant here).
9. decide_stripe_size_regular caps the length of the stripe it
computed at 1 GiB. This cap appeared in 5da431b71d to correct
one of the other regressions introduced in f6fca3917b.
10. btrfs_create_chunk creates a new chunk with the above
computed size and number of devices.
At step 4, gather_device_info() has found a location where stripe up to
10 GiB in length could be allocated on several devices, and selected
which devices should have a dev_extent allocated on them, but at step
9, only 1 GiB of the space that was found on each device can be used.
This mismatch causes new suboptimal chunk allocation cases that did not
occur in pre-f6fca3917b4d kernels.
Consider a filesystem using raid1 profile with 3 devices. After some
balances, device 1 has 10x 1 GiB unallocated space, while devices 2
and 3 have 1x 10 GiB unallocated space, i.e. the same total amount of
space, but distributed across different numbers of dev_extent holes.
For visualization, let's ignore all the chunks that were allocated before
this point, and focus on the remaining holes:
Device 1: [_] [_] [_] [_] [_] [_] [_] [_] [_] [_] (10x 1 GiB unallocated)
Device 2: [__________] (10 GiB contig unallocated)
Device 3: [__________] (10 GiB contig unallocated)
Before f6fca3917b, the allocator would fill these optimally by
allocating chunks with dev_extents on devices 1 and 2 ([12]), 1 and 3
([13]), or 2 and 3 ([23]):
[after 0 chunk allocations]
Device 1: [_] [_] [_] [_] [_] [_] [_] [_] [_] [_] (10 GiB)
Device 2: [__________] (10 GiB)
Device 3: [__________] (10 GiB)
[after 1 chunk allocation]
Device 1: [12] [_] [_] [_] [_] [_] [_] [_] [_] [_]
Device 2: [12] [_________] (9 GiB)
Device 3: [__________] (10 GiB)
[after 2 chunk allocations]
Device 1: [12] [13] [_] [_] [_] [_] [_] [_] [_] [_] (8 GiB)
Device 2: [12] [_________] (9 GiB)
Device 3: [13] [_________] (9 GiB)
[after 3 chunk allocations]
Device 1: [12] [13] [12] [_] [_] [_] [_] [_] [_] [_] (7 GiB)
Device 2: [12] [12] [________] (8 GiB)
Device 3: [13] [_________] (9 GiB)
[...]
[after 12 chunk allocations]
Device 1: [12] [13] [12] [13] [12] [13] [12] [13] [_] [_] (2 GiB)
Device 2: [12] [12] [23] [23] [12] [12] [23] [23] [__] (2 GiB)
Device 3: [13] [13] [23] [23] [13] [23] [13] [23] [__] (2 GiB)
[after 13 chunk allocations]
Device 1: [12] [13] [12] [13] [12] [13] [12] [13] [12] [_] (1 GiB)
Device 2: [12] [12] [23] [23] [12] [12] [23] [23] [12] [_] (1 GiB)
Device 3: [13] [13] [23] [23] [13] [23] [13] [23] [__] (2 GiB)
[after 14 chunk allocations]
Device 1: [12] [13] [12] [13] [12] [13] [12] [13] [12] [13] (full)
Device 2: [12] [12] [23] [23] [12] [12] [23] [23] [12] [_] (1 GiB)
Device 3: [13] [13] [23] [23] [13] [23] [13] [23] [13] [_] (1 GiB)
[after 15 chunk allocations]
Device 1: [12] [13] [12] [13] [12] [13] [12] [13] [12] [13] (full)
Device 2: [12] [12] [23] [23] [12] [12] [23] [23] [12] [23] (full)
Device 3: [13] [13] [23] [23] [13] [23] [13] [23] [13] [23] (full)
This allocates all of the space with no waste. The sorting function used
by gather_device_info considers free space holes above 1 GiB in length
to be equal to 1 GiB, so once find_free_dev_extent locates a sufficiently
long hole on each device, all the holes appear equal in the sort, and the
comparison falls back to sorting devices by total free space. This keeps
usable space on each device equal so they can all be filled completely.
After f6fca3917b, the allocator prefers the devices with larger holes
over the devices with more free space, so it makes bad allocation choices:
[after 1 chunk allocation]
Device 1: [_] [_] [_] [_] [_] [_] [_] [_] [_] [_] (10 GiB)
Device 2: [23] [_________] (9 GiB)
Device 3: [23] [_________] (9 GiB)
[after 2 chunk allocations]
Device 1: [_] [_] [_] [_] [_] [_] [_] [_] [_] [_] (10 GiB)
Device 2: [23] [23] [________] (8 GiB)
Device 3: [23] [23] [________] (8 GiB)
[after 3 chunk allocations]
Device 1: [_] [_] [_] [_] [_] [_] [_] [_] [_] [_] (10 GiB)
Device 2: [23] [23] [23] [_______] (7 GiB)
Device 3: [23] [23] [23] [_______] (7 GiB)
[...]
[after 9 chunk allocations]
Device 1: [_] [_] [_] [_] [_] [_] [_] [_] [_] [_] (10 GiB)
Device 2: [23] [23] [23] [23] [23] [23] [23] [23] [23] [_] (1 GiB)
Device 3: [23] [23] [23] [23] [23] [23] [23] [23] [23] [_] (1 GiB)
[after 10 chunk allocations]
Device 1: [12] [_] [_] [_] [_] [_] [_] [_] [_] [_] (9 GiB)
Device 2: [23] [23] [23] [23] [23] [23] [23] [23] [12] (full)
Device 3: [23] [23] [23] [23] [23] [23] [23] [23] [_] (1 GiB)
[after 11 chunk allocations]
Device 1: [12] [13] [_] [_] [_] [_] [_] [_] [_] [_] (8 GiB)
Device 2: [23] [23] [23] [23] [23] [23] [23] [23] [12] (full)
Device 3: [23] [23] [23] [23] [23] [23] [23] [23] [13] (full)
No further allocations are possible, with 8 GiB wasted (4 GiB of data
space). The sort in gather_device_info now considers free space in
holes longer than 1 GiB to be distinct, so it will prefer devices 2 and
3 over device 1 until all but 1 GiB is allocated on devices 2 and 3.
At that point, with only 1 GiB unallocated on every device, the largest
hole length on each device is equal at 1 GiB, so the sort finally moves
to ordering the devices with the most free space, but by this time it
is too late to make use of the free space on device 1.
Note that it's possible to contrive a case where the pre-f6fca3917b4d
allocator fails the same way, but these cases generally have extensive
dev_extent fragmentation as a precondition (e.g. many holes of 768M
in length on one device, and few holes 1 GiB in length on the others).
With the regression in f6fca3917b, bad chunk allocation can occur even
under optimal conditions, when all dev_extent holes are exact multiples
of stripe_len in length, as in the example above.
Also note that post-f6fca3917b4d kernels do treat dev_extent holes
larger than 10 GiB as equal, so the bad behavior won't show up on a
freshly formatted filesystem; however, as the filesystem ages and fills
up, and holes ranging from 1 GiB to 10 GiB in size appear, the problem
can show up as a failure to balance after adding or removing devices,
or an unexpected shortfall in available space due to unequal allocation.
To fix the regression and make data chunk allocation work
again, set ctl->max_stripe_len back to the original SZ_1G, or
space_info->chunk_size if that's smaller (the latter can happen if the
user set space_info->chunk_size to less than 1 GiB via sysfs, or it's
a 32 MiB system chunk with a hardcoded chunk_size and stripe_len).
While researching the background of the earlier commits, I found that an
identical fix was already proposed at:
https://lore.kernel.org/linux-btrfs/de83ac46-a4a3-88d3-85ce-255b7abc5249@gmx.com/
The previous review missed one detail: ctl->max_stripe_len is used
before decide_stripe_size_regular() is called, when it is too late for
the changes in that function to have any effect. ctl->max_stripe_len is
not used directly by decide_stripe_size_regular(), but the parameter
does heavily influence the per-device free space data presented to
the function.
Fixes: f6fca3917b ("btrfs: store chunk size in space-info struct")
CC: stable@vger.kernel.org # 6.1+
Link: https://lore.kernel.org/linux-btrfs/20231007051421.19657-1-ce3g8jdj@umail.furryterror.org/
Reviewed-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: Zygo Blaxell <ce3g8jdj@umail.furryterror.org>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 4abbd2e3c1 upstream.
In the smc_listen_work(), if smc_listen_prfx_check() failed,
the real reason: SMC_CLC_DECL_DIFFPREFIX was dropped, and
SMC_CLC_DECL_NOSMCDEV was returned.
Althrough this is also kind of SMC_CLC_DECL_NOSMCDEV, but return
the real reason is much friendly for debugging.
Fixes: e49300a6bf ("net/smc: add listen processing for SMC-Rv2")
Signed-off-by: Dust Li <dust.li@linux.alibaba.com>
Reviewed-by: Alexandra Winter <wintera@linux.ibm.com>
Reviewed-by: Wenjia Zhang <wenjia@linux.ibm.com>
Link: https://lore.kernel.org/r/20231012123729.29307-1-dust.li@linux.alibaba.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 0288c3e709 upstream.
When the system boots into the crash dump kernel after a panic, the ice
networking device may still have pending transactions that can cause errors
or machine checks when the device is re-enabled. This can prevent the crash
dump kernel from loading the driver or collecting the crash data.
To avoid this issue, perform a function level reset (FLR) on the ice device
via PCIe config space before enabling it on the crash kernel. This will
clear any outstanding transactions and stop all queues and interrupts.
Restore the config space after the FLR, otherwise it was found in testing
that the driver wouldn't load successfully.
The following sequence causes the original issue:
- Load the ice driver with modprobe ice
- Enable SR-IOV with 2 VFs: echo 2 > /sys/class/net/eth0/device/sriov_num_vfs
- Trigger a crash with echo c > /proc/sysrq-trigger
- Load the ice driver again (or let it load automatically) with modprobe ice
- The system crashes again during pcim_enable_device()
Fixes: 837f08fdec ("ice: Add basic driver framework for Intel(R) E800 Series")
Reported-by: Vishal Agrawal <vagrawal@redhat.com>
Reviewed-by: Jay Vosburgh <jay.vosburgh@canonical.com>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Pucha Himasekhar Reddy <himasekharx.reddy.pucha@intel.com> (A Contingent worker at Intel)
Link: https://lore.kernel.org/r/20231011233334.336092-3-jacob.e.keller@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
