My latest patch, attempting to fix the refcount leak in a minimal
way turned out to add a new bug.
Whenever the bind operation fails before we attempt to grab
a reference count on a device, we might release the device refcount
of a prior successful bind() operation.
syzbot was not happy about this [1].
Note to stable teams:
Make sure commit b37a46683739 ("netdevice: add the case if dev is NULL")
is already present in your trees.
af_key: add __GFP_ZERO flag for compose_sadb_supported in function pfkey_register
Add __GFP_ZERO flag for compose_sadb_supported in function pfkey_register
to initialize the buffer of supp_skb to fix a kernel-info-leak issue.
1) Function pfkey_register calls compose_sadb_supported to request
a sk_buff. 2) compose_sadb_supported calls alloc_sbk to allocate
a sk_buff, but it doesn't zero it. 3) If auth_len is greater 0, then
compose_sadb_supported treats the memory as a struct sadb_supported and
begins to initialize. But it just initializes the field sadb_supported_len
and field sadb_supported_exttype without field sadb_supported_reserved.
In order to not add fancy protection for drv_priv we can move
htc_handle->drv_priv initialization at the end of the
ath9k_htc_probe_device() and add helper macro to make
all *_STAT_* macros NULL safe, since syzbot has reported related NULL
deref in that macros [1]
Commit 6dce5aa59e0b ("PCI: xgene: Use inbound resources for setup")
killed PCIe on my XGene-1 box (a Mustang board). The machine itself
is still alive, but half of its storage (over NVMe) is gone, and the
NVMe driver just times out.
Note that this machine boots with a device tree provided by the
UEFI firmware (2016 vintage), which could well be non conformant
with the spec, hence the breakage.
With the patch reverted, the box boots 5.17-rc8 with flying colors.
Link: https://<email address hidden>
Link: https://<email address hidden>
Fixes: 6dce5aa59e0b ("PCI: xgene: Use inbound resources for setup")
Signed-off-by: Marc Zyngier <email address hidden>
Signed-off-by: Lorenzo Pieralisi <email address hidden>
Cc: <email address hidden>
Cc: Rob Herring <email address hidden>
Cc: Toan Le <email address hidden>
Cc: Lorenzo Pieralisi <email address hidden>
Cc: Krzysztof Wilczyński <email address hidden>
Cc: Bjorn Helgaas <email address hidden>
Cc: Stéphane Graber <email address hidden>
Cc: dann frazier <email address hidden>
(backported from commit 1874b6d7ab1bdc900e8398026350313ac29caddb)
[dannf: minor context adjustment]
Signed-off-by: dann frazier <email address hidden>
Acked-by: Tim Gardner <email address hidden>
Acked-by: Luke Nowakowski-Krijger <email address hidden>
Signed-off-by: Stefan Bader <email address hidden>
Commit c7a75d07827a ("PCI: xgene: Fix IB window setup") tried to
fix the damages that 6dce5aa59e0b ("PCI: xgene: Use inbound resources
for setup") caused, but actually didn't improve anything for some
plarforms (at least Mustang and m400 are still broken).
Given that 6dce5aa59e0b has been reverted, revert this patch as well,
restoring the PCIe support on XGene to its pre-5.5, working state.
Link: https://<email address hidden>
Link: https://<email address hidden>
Fixes: c7a75d07827a ("PCI: xgene: Fix IB window setup")
Signed-off-by: Marc Zyngier <email address hidden>
Signed-off-by: Lorenzo Pieralisi <email address hidden>
Cc: <email address hidden>
Cc: Rob Herring <email address hidden>
Cc: Toan Le <email address hidden>
Cc: Lorenzo Pieralisi <email address hidden>
Cc: Krzysztof Wilczyński <email address hidden>
Cc: Bjorn Helgaas <email address hidden>
Cc: Stéphane Graber <email address hidden>
Cc: dann frazier <email address hidden>
(cherry picked from commit 825da4e9cec68713fbb02dc6f71fe1bf65fe8050)
Signed-off-by: dann frazier <email address hidden>
Acked-by: Tim Gardner <email address hidden>
Acked-by: Luke Nowakowski-Krijger <email address hidden>
Signed-off-by: Stefan Bader <email address hidden>
This reverts commit 3829acb7f33d2bcf746b2df598c9a3066713fc2d.
With the broken test case, "underlay in a VRF", fixed in the
test_vxlan_under_vrf.sh, we should remove the SAUCE patches that mark
this test failure as an expected failure to catch regressions in the
future. This could reduce maintenance cost as well.
Signed-off-by: Po-Hsu Lin <email address hidden>
Acked-by: Luke Nowakowski-Krijger <email address hidden>
Acked-by: Tim Gardner <email address hidden>
This reverts commit 7d9c6353c36a560ca73ca1bea35e16f62b4bec69.
With the broken test case, "underlay in a VRF", fixed in the
test_vxlan_under_vrf.sh, we should remove the SAUCE patches that mark
this test failure as an expected failure to catch regressions in the
future. This could reduce maintenance cost as well.
Signed-off-by: Po-Hsu Lin <email address hidden>
Acked-by: Luke Nowakowski-Krijger <email address hidden>
Acked-by: Tim Gardner <email address hidden>
s390/mm: use non-quiescing sske for KVM switch to keyed guest
s390/mm: use non-quiescing sske for KVM switch to keyed guest
From: Christian Borntraeger <email address hidden>
The switch to a keyed guest does not require a classic sske as the other
guest CPUs are not accessing the key before the switch is complete.
By using the NQ SSKE things are faster especially with multiple guests.
Signed-off-by: Christian Borntraeger <email address hidden>
Suggested-by: Janis Schoetterl-Glausch <email address hidden>
Reviewed-by: Claudio Imbrenda <email address hidden>
Link: https://<email address hidden>
Signed-off-by: Christian Borntraeger <email address hidden>
Signed-off-by: Heiko Carstens <email address hidden>
(cherry picked from commit 3ae11dbcfac906a8c3a480e98660a823130dc16a)
Signed-off-by: Frank Heimes <email address hidden>
Acked-by: Stefan Bader <email address hidden>
Acked-by: Tim Gardner <email address hidden>
Signed-off-by: Stefan Bader <email address hidden>