A page table upgrade in a kernel section that uses secondary address
mode will mess up the kernel instructions as follows:
Consider the following scenario: two threads are sharing memory.
On CPU1 thread 1 does e.g. strnlen_user(). That gets to
old_fs = enable_sacf_uaccess();
len = strnlen_user_srst(src, size);
and
" la %2,0(%1)\n"
" la %3,0(%0,%1)\n"
" slgr %0,%0\n"
" sacf 256\n"
"0: srst %3,%2\n"
in strnlen_user_srst(). At that point we are in secondary space mode,
control register 1 points to kernel page table and instruction fetching
happens via c1, rather than usual c13. Interrupts are not disabled, for
obvious reasons.
On CPU2 thread 2 does MAP_FIXED mmap(), forcing the upgrade of page table
from 3-level to e.g. 4-level one. We'd allocated new top-level table,
set it up and now we hit this:
notify = 1;
spin_unlock_bh(&mm->page_table_lock);
}
if (notify)
on_each_cpu(__crst_table_upgrade, mm, 0);
OK, we need to actually change over to use of new page table and we
need that to happen in all threads that are currently running. Which
happens to include the thread 1. IPI is delivered and we have
static void __crst_table_upgrade(void *arg)
{
struct mm_struct *mm = arg;
if (current->active_mm == mm)
set_user_asce(mm);
__tlb_flush_local();
}
run on CPU1. That does
static inline void set_user_asce(struct mm_struct *mm)
{
S390_lowcore.user_asce = mm->context.asce;
OK, user page table address updated...
__ctl_load(S390_lowcore.user_asce, 1, 1);
... and control register 1 set to it.
clear_cpu_flag(CIF_ASCE_PRIMARY);
}
IPI is run in home space mode, so it's fine - insns are fetched
using c13, which always points to kernel page table. But as soon
as we return from the interrupt, previous PSW is restored, putting
CPU1 back into secondary space mode, at which point we no longer
get the kernel instructions from the kernel mapping.
The fix is to only fixup the control registers that are currently in use
for user processes during the page table update. We must also disable
interrupts in enable_sacf_uaccess to synchronize the cr and
thread.mm_segment updates against the on_each-cpu.
Fixes: 0aaba41b58bc ("s390: remove all code using the access register mode") Cc: stable@vger.kernel.org # 4.15+ Reported-by: Al Viro <viro@zeniv.linux.org.uk> Reviewed-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
References: CVE-2020-11884 Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When performing rename operation with RENAME_WHITEOUT flag, we will
hold AGF lock to allocate or free extents in manipulating the dirents
firstly, and then doing the xfs_iunlink_remove() call last to hold
AGI lock to modify the tmpfile info, so we the lock order AGI->AGF.
The big problem here is that we have an ordering constraint on AGF
and AGI locking - inode allocation locks the AGI, then can allocate
a new extent for new inodes, locking the AGF after the AGI. Hence
the ordering that is imposed by other parts of the code is AGI before
AGF. So we get an ABBA deadlock between the AGI and AGF here.
In this patch we move the xfs_iunlink_remove() call to
before acquiring the AGF lock to preserve correct AGI/AGF locking
order.
[Minor massage required due to upstream change making xfs_bumplink() a
void function where as in the 4.19.y tree the return value is checked,
even though it is always zero. Only change was to the last code block
removed by the patch. Functionally equivalent to upstream.]
Signed-off-by: kaixuxia <kaixuxia@tencent.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Suraj Jitindar Singh <surajjs@amazon.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
For SCIF and HSCIF interfaces the SCxSR register holds the status of
data that is to be read next from SCxRDR register, But where as for
SCIFA and SCIFB interfaces SCxSR register holds status of data that is
previously read from SCxRDR register.
This patch makes sure the status register is read depending on the port
types so that errors are caught accordingly.
Suspending the bus and host controller while a port is in a over-current
condition may halt the host.
Also keep the roothub running if over-current is active.
For userspace functions using OS Descriptors, if a function also supplies
Extended Property descriptors currently the counts and lengths stored in
the ms_os_descs_ext_prop_{count,name_len,data_len} variables are not
getting reset to 0 during an unbind or when the epfiles are closed. If
the same function is re-bound and the descriptors are re-written, this
results in those count/length variables to monotonically increase
causing the VLA allocation in _ffs_func_bind() to grow larger and larger
at each bind/unbind cycle and eventually fail to allocate.
Fix this by clearing the ms_os_descs_ext_prop count & lengths to 0 in
ffs_data_reset().
A request may not be completed because not all the TRBs are prepared for
it. This happens when we run out of available TRBs. When some TRBs are
completed, the driver needs to prepare the rest of the TRBs for the
request. The check dwc3_gadget_ep_request_completed() shouldn't be
checking the amount of data received but rather the number of pending
TRBs. Revise this request completion check.
A SCSI error handler and block runtime PM must not allocate
memory with GFP_KERNEL. Furthermore they must not wait for
tasks allocating memory with GFP_KERNEL.
That means that they cannot share a workqueue with arbitrary tasks.
Fix this for UAS using a private workqueue.
Signed-off-by: Oliver Neukum <oneukum@suse.com> Fixes: f9dc024a2da1f ("uas: pre_reset and suspend: Fix a few races") Cc: stable <stable@vger.kernel.org> Link: https://lore.kernel.org/r/20200415141750.811-2-oneukum@suse.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Suspend increments a counter, then kills the URBs,
then kills the scheduled work. The scheduled work, however,
may reschedule the URBs. Fix this by having the work
check the counter.
Signed-off-by: Oliver Neukum <oneukum@suse.com> Cc: stable <stable@vger.kernel.org> Tested-by: Jonas Karlsson <jonas.karlsson@actia.se> Link: https://lore.kernel.org/r/20200415151358.32664-1-oneukum@suse.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Even if the actual screen size is bounded in vc_do_resize(), the unicode
buffer is still a little more than twice the size of the glyph buffer
and may exceed MAX_ORDER down the kmalloc() path. This can be triggered
from user space.
Since there is no point having a physically contiguous buffer here,
let's avoid the above issue as well as reducing pressure on high order
allocations by using vmalloc() instead.
The code in vc_do_resize() bounds the memory allocation size to avoid
exceeding MAX_ORDER down the kzalloc() call chain and generating a
runtime warning triggerable from user space. However, not only is it
unwise to use a literal value here, but MAX_ORDER may also be
configurable based on CONFIG_FORCE_MAX_ZONEORDER.
Let's use KMALLOC_MAX_SIZE instead.
Note that prior commit bb1107f7c605 ("mm, slab: make sure that
KMALLOC_MAX_SIZE will fit into MAX_ORDER") the KMALLOC_MAX_SIZE value
could not be relied upon.
comedi_open() invokes comedi_dev_get_from_minor(), which returns a
reference of the COMEDI device to "dev" with increased refcount.
When comedi_open() returns, "dev" becomes invalid, so the refcount
should be decreased to keep refcount balanced.
The reference counting issue happens in one exception handling path of
comedi_open(). When "cfp" allocation is failed, the refcnt increased by
comedi_dev_get_from_minor() is not decreased, causing a refcnt leak.
Fix this issue by calling comedi_dev_put() on this error path when "cfp"
allocation is failed.
Fixes: 20f083c07565 ("staging: comedi: prepare support for per-file read and write subdevices") Signed-off-by: Xiyu Yang <xiyuyang19@fudan.edu.cn> Cc: stable <stable@vger.kernel.org> Signed-off-by: Xin Tan <tanxin.ctf@gmail.com> Signed-off-by: Ian Abbott <abbotti@mev.co.uk> Link: https://lore.kernel.org/r/1587361459-83622-1-git-send-email-xiyuyang19@fudan.edu.cn Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The DT2815 analog output command is 16 bits wide, consisting of the
12-bit sample value in bits 15 to 4, the channel number in bits 3 to 1,
and a voltage or current selector in bit 0. Both bytes of the 16-bit
command need to be written in turn to a single 8-bit data register.
However, the driver currently only writes the low 8-bits. It is broken
and appears to have always been broken.
Electronic copies of the DT2815 User's Manual seem impossible to find
online, but looking at the source code, a best guess for the sequence
the driver intended to use to write the analog output command is as
follows:
1. Wait for the status register to read 0x00.
2. Write the low byte of the command to the data register.
3. Wait for the status register to read 0x80.
4. Write the high byte of the command to the data register.
Step 4 is missing from the driver. Add step 4 to (hopefully) fix the
driver.
Also add a "FIXME" comment about setting bit 0 of the low byte of the
command. Supposedly, it is used to choose between voltage output and
current output, but the current driver always sets it to 1.
If {i,d}-cache-block-size is set and {i,d}-cache-line-size is not, use
the block-size value for both. Per the devicetree spec cache-line-size
is only needed if it differs from the block size.
Originally the code would fallback from block size to line size. An
error message was printed if both properties were missing.
Later the code was refactored to use clearer names and logic but it
inadvertently made line size a required property, meaning on systems
without a line size property we fall back to the default from the
cputable.
On powernv (OPAL) platforms, since the introduction of device tree CPU
features (5a61ef74f269 ("powerpc/64s: Support new device tree binding
for discovering CPU features")), that has led to the wrong value being
used, as the fallback value is incorrect for Power8/Power9 CPUs.
The incorrect values flow through to the VDSO and also to the sysconf
values, SC_LEVEL1_ICACHE_LINESIZE etc.
Fixes: bd067f83b084 ("powerpc/64: Fix naming of cache block vs. cache line") Cc: stable@vger.kernel.org # v4.11+ Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz> Reported-by: Qian Cai <cai@lca.pw>
[mpe: Add even more detail to change log] Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200416221908.7886-1-chris.packham@alliedtelesis.co.nz Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
512a928affd5 ("ARM: imx: build v7_cpu_resume() unconditionally")
introduced an unintended linker error for i.MX6 configurations that have
ARM_CPU_SUSPEND=n which can happen if neither CONFIG_PM, CONFIG_CPU_IDLE,
nor ARM_PSCI_FW are selected.
Fix this by having v7_cpu_resume() compiled only when cpu_resume() it
calls is available as well.
The C declaration for the function remains unguarded to avoid future code
inadvertently using a stub and introducing a regression to the bug the
original commit fixed.
Cc: <stable@vger.kernel.org> Fixes: 512a928affd5 ("ARM: imx: build v7_cpu_resume() unconditionally") Reported-by: Clemens Gruber <clemens.gruber@pqgruber.com> Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de> Tested-by: Roland Hieber <rhi@pengutronix.de> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The iwl_trans_pcie_dyn_txq_free() function only releases the frames
that may be left on the queue by calling iwl_pcie_gen2_txq_unmap(),
but doesn't actually free the DMA ring or byte-count tables for the
queue. This leads to pretty large memory leaks (at least before my
queue size improvements), in particular in monitor/sniffer mode on
channel hopping since this happens on every channel change.
This was also now more evident after the move to a DMA pool for the
byte count tables, showing messages such as
BUG iwlwifi:bc (...): Objects remaining in iwlwifi:bc on __kmem_cache_shutdown()
This fixes https://bugzilla.kernel.org/show_bug.cgi?id=206811.
snd_soc_dapm_kcontrol widget which is created by autodisable control
should contain correct on_val, mask and shift because it is set when the
widget is powered and changed value is applied on registers by following
code in dapm_seq_run_coalesced().
mask |= w->mask << w->shift;
if (w->power)
value |= w->on_val << w->shift;
else
value |= w->off_val << w->shift;
Shift on the mask in dapm_kcontrol_data_alloc() is removed to prevent
double shift.
And, on_val in dapm_kcontrol_set_value() is modified to get correct
value in the dapm_seq_run_coalesced().
Commit 756125289285 ("audit: always check the netlink payload length
in audit_receive_msg()") fixed a number of missing message length
checks, but forgot to check the length of userspace generated audit
records. The good news is that you need CAP_AUDIT_WRITE to submit
userspace audit records, which is generally only given to trusted
processes, so the impact should be limited.
Cc: stable@vger.kernel.org Fixes: 756125289285 ("audit: always check the netlink payload length in audit_receive_msg()") Reported-by: syzbot+49e69b4d71a420ceda3e@syzkaller.appspotmail.com Signed-off-by: Paul Moore <paul@paul-moore.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cyril Roelandt reports that his JMicron JMS566 USB-SATA bridge fails
to handle WRITE commands with the FUA bit set, even though it claims
to support FUA. (Oddly enough, a later version of the same bridge,
version 2.03 as opposed to 1.14, doesn't claim to support FUA. Also
oddly, the bridge _does_ support FUA when using the UAS transport
instead of the Bulk-Only transport -- but this device was blacklisted
for uas in commit bc3bdb12bbb3 ("usb-storage: Disable UAS on JMicron
SATA enclosure") for apparently unrelated reasons.)
This patch adds a usb-storage unusual_devs entry with the BROKEN_FUA
flag. This allows the bridge to work properly with usb-storage.
init_r_port can access pc104 array out of bounds. pc104 is a 2D array
defined to have 4 members. Each member has 8 submembers.
* we can have more than 4 (PCI) boards, i.e. [board] can be OOB
* line is not modulo-ed by anything, so the first line on the second
board can be 4, on the 3rd 12 or alike (depending on previously
registered boards). It's zero only on the first line of the first
board. So even [line] can be OOB, quite soon (with the 2nd registered
board already).
This code is broken for ages, so just avoid the OOB accesses and don't
try to fix it as we would need to find out the correct line number. Use
the default: RS232, if we are out.
Generally, if anyone needs to set the interface types, a module parameter
is past the last thing that should be used for this purpose. The
parameters' description says it's for ISA cards anyway.
If there is a lot(more then 16) of virtio-console devices
or virtio_console module is reloaded
- buffers 'vtermnos' and 'cons_ops' are overflowed.
In older kernels it overruns spinlock which leads to kernel freezing:
https://bugzilla.redhat.com/show_bug.cgi?id=1786239
To reproduce the issue, you can try simple script that
loads/unloads module. Something like this:
while [ 1 ]
do
modprobe virtio_console
sleep 2
modprobe -r virtio_console
sleep 2
done
Description of problem:
Guest get 'Call Trace' when loading module "virtio_console"
and unloading it frequently - clearly reproduced on kernel-4.18.0:
Check that the resolved slot (somewhat confusingly named 'start') is a
valid/allocated slot before doing the final comparison to see if the
specified gfn resides in the associated slot. The resolved slot can be
invalid if the binary search loop terminated because the search index
was incremented beyond the number of used slots.
This bug has existed since the binary search algorithm was introduced,
but went unnoticed because KVM statically allocated memory for the max
number of slots, i.e. the access would only be truly out-of-bounds if
all possible slots were allocated and the specified gfn was less than
the base of the lowest memslot. Commit 36947254e5f98 ("KVM: Dynamically
size memslot array based on number of used slots") eliminated the "all
possible slots allocated" condition and made the bug embarrasingly easy
to hit.
Fixes: 9c1a5d38780e6 ("kvm: optimize GFN to memslot lookup with large slots amount") Reported-by: syzbot+d889b59b2bb87d4047a2@syzkaller.appspotmail.com Cc: stable@vger.kernel.org Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200408064059.8957-2-sean.j.christopherson@intel.com> Reviewed-by: Cornelia Huck <cohuck@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Return the index of the last valid slot from gfn_to_memslot_approx() if
its binary search loop yielded an out-of-bounds index. The index can
be out-of-bounds if the specified gfn is less than the base of the
lowest memslot (which is also the last valid memslot).
Note, the sole caller, kvm_s390_get_cmma(), ensures used_slots is
non-zero.
Fixes: afdad61615cc3 ("KVM: s390: Fix storage attributes migration with memory slots") Cc: stable@vger.kernel.org # 4.19.x: 0774a964ef56: KVM: Fix out of range accesses to memslots Cc: stable@vger.kernel.org # 4.19.x Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200408064059.8957-3-sean.j.christopherson@intel.com> Reviewed-by: Cornelia Huck <cohuck@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
tpm_ibmvtpm_send() can fail during PowerVM Live Partition Mobility resume
with an H_CLOSED return from ibmvtpm_send_crq(). The PAPR says, 'The
"partner partition suspended" transport event disables the associated CRQ
such that any H_SEND_CRQ hcall() to the associated CRQ returns H_Closed
until the CRQ has been explicitly enabled using the H_ENABLE_CRQ hcall.'
This patch adds a check in tpm_ibmvtpm_send() for an H_CLOSED return from
ibmvtpm_send_crq() and in that case calls tpm_ibmvtpm_resume() and
retries the ibmvtpm_send_crq() once.
Cc: stable@vger.kernel.org # 3.7.x Fixes: 132f76294744 ("drivers/char/tpm: Add new device driver to support IBM vTPM") Reported-by: Linh Pham <phaml@us.ibm.com> Reviewed-by: Stefan Berger <stefanb@linux.ibm.com> Signed-off-by: George Wilson <gcwilson@linux.ibm.com> Tested-by: Linh Pham <phaml@us.ibm.com> Reviewed-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
snd_microii_spdif_default_get() invokes snd_usb_lock_shutdown(), which
increases the refcount of the snd_usb_audio object "chip".
When snd_microii_spdif_default_get() returns, local variable "chip"
becomes invalid, so the refcount should be decreased to keep refcount
balanced.
The reference counting issue happens in several exception handling paths
of snd_microii_spdif_default_get(). When those error scenarios occur
such as usb_ifnum_to_if() returns NULL, the function forgets to decrease
the refcnt increased by snd_usb_lock_shutdown(), causing a refcnt leak.
Fix this issue by jumping to "end" label when those error scenarios
occur.
Fixes: 447d6275f0c2 ("ALSA: usb-audio: Add sanity checks for endpoint accesses") Signed-off-by: Xiyu Yang <xiyuyang19@fudan.edu.cn> Signed-off-by: Xin Tan <tanxin.ctf@gmail.com> Cc: <stable@vger.kernel.org> Link: https://lore.kernel.org/r/1587617711-13200-1-git-send-email-xiyuyang19@fudan.edu.cn Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The commit 1c76aa5fb48d ("ALSA: hda/realtek - Allow skipping
spec->init_amp detection") changed the way to assign spec->init_amp
field that specifies the way to initialize the amp. Along with the
change, the commit also replaced a few fixups that set spec->init_amp
in HDA_FIXUP_ACT_PROBE with HDA_FIXUP_ACT_PRE_PROBE. This was rather
aligning to the other fixups, and not supposed to change the actual
behavior.
However, this change turned out to cause a regression on FSC S7020,
which hit exactly the above. The reason was that there is still one
place that overrides spec->init_amp after HDA_FIXUP_ACT_PRE_PROBE
call, namely in alc_ssid_check().
This patch fixes the regression by adding the proper spec->init_amp
override check, i.e. verifying whether it's still ALC_INIT_UNDEFINED.
The error handling code in usX2Y_rate_set() may hit a potential NULL
dereference when an error occurs before allocating all us->urb[].
Add a proper NULL check for fixing the corner case.
Commit 7ed1c1901fe5 ("tools: fix cross-compile var clobbering") moved
the setup of the CC variable to tools/scripts/Makefile.include to make
the behavior consistent across all the tools Makefiles.
As the vm tools missed the include we end up with the wrong CC in a
cross-compiling evironment.
Fixes: 7ed1c1901fe5 (tools: fix cross-compile var clobbering) Signed-off-by: Lucas Stach <l.stach@pengutronix.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Cc: Martin Kelly <martin@martingkelly.com> Cc: <stable@vger.kernel.org> Link: http://lkml.kernel.org/r/20200416104748.25243-1-l.stach@pengutronix.de Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
find_mergeable_vma() can return NULL. In this case, it leads to a crash
when we access vm_mm(its offset is 0x40) later in write_protect_page.
And this case did happen on our server. The following call trace is
captured in kernel 4.19 with the following patch applied and KSM zero
page enabled on our server.
commit e86c59b1b12d ("mm/ksm: improve deduplication of zero pages with colouring")
[songmuchun@bytedance.com: if the vma is out of date, just exit] Link: http://lkml.kernel.org/r/20200416025034.29780-1-songmuchun@bytedance.com
[akpm@linux-foundation.org: add the conventional braces, replace /** with /*] Fixes: e86c59b1b12d ("mm/ksm: improve deduplication of zero pages with colouring") Co-developed-by: Xiongchun Duan <duanxiongchun@bytedance.com> Signed-off-by: Muchun Song <songmuchun@bytedance.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: David Hildenbrand <david@redhat.com> Reviewed-by: Kirill Tkhai <ktkhai@virtuozzo.com> Cc: Hugh Dickins <hughd@google.com> Cc: Yang Shi <yang.shi@linux.alibaba.com> Cc: Claudio Imbrenda <imbrenda@linux.vnet.ibm.com> Cc: Markus Elfring <Markus.Elfring@web.de> Cc: <stable@vger.kernel.org> Link: http://lkml.kernel.org/r/20200416025034.29780-1-songmuchun@bytedance.com Link: http://lkml.kernel.org/r/20200414132905.83819-1-songmuchun@bytedance.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
For 1G hugepages, huge_pte_offset() wants to return NULL or pudp, but it
may return a wrong 'pmdp' if there is a race. Please look at the
following code snippet:
...
pud = pud_offset(p4d, addr);
if (sz != PUD_SIZE && pud_none(*pud))
return NULL;
/* hugepage or swap? */
if (pud_huge(*pud) || !pud_present(*pud))
return (pte_t *)pud;
pmd = pmd_offset(pud, addr);
if (sz != PMD_SIZE && pmd_none(*pmd))
return NULL;
/* hugepage or swap? */
if (pmd_huge(*pmd) || !pmd_present(*pmd))
return (pte_t *)pmd;
...
The following sequence would trigger this bug:
- CPU0: sz = PUD_SIZE and *pud = 0 , continue
- CPU0: "pud_huge(*pud)" is false
- CPU1: calling hugetlb_no_page and set *pud to xxxx8e7(PRESENT)
- CPU0: "!pud_present(*pud)" is false, continue
- CPU0: pmd = pmd_offset(pud, addr) and maybe return a wrong pmdp
However, we want CPU0 to return NULL or pudp in this case.
We must make sure there is exactly one dereference of pud and pmd.
Signed-off-by: Longpeng <longpeng2@huawei.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com> Reviewed-by: Jason Gunthorpe <jgg@mellanox.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Sean Christopherson <sean.j.christopherson@intel.com> Cc: <stable@vger.kernel.org> Link: http://lkml.kernel.org/r/20200413010342.771-1-longpeng2@huawei.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
remap_vmalloc_range() has had various issues with the bounds checks it
promises to perform ("This function checks that addr is a valid
vmalloc'ed area, and that it is big enough to cover the vma") over time,
e.g.:
- not detecting pgoff<<PAGE_SHIFT overflow
- not detecting (pgoff<<PAGE_SHIFT)+usize overflow
- not checking whether addr and addr+(pgoff<<PAGE_SHIFT) are the same
vmalloc allocation
- comparing a potentially wildly out-of-bounds pointer with the end of
the vmalloc region
In particular, since commit fc9702273e2e ("bpf: Add mmap() support for
BPF_MAP_TYPE_ARRAY"), unprivileged users can cause kernel null pointer
dereferences by calling mmap() on a BPF map with a size that is bigger
than the distance from the start of the BPF map to the end of the
address space.
This could theoretically be used as a kernel ASLR bypass, by using
whether mmap() with a given offset oopses or returns an error code to
perform a binary search over the possible address range.
To allow remap_vmalloc_range_partial() to verify that addr and
addr+(pgoff<<PAGE_SHIFT) are in the same vmalloc region, pass the offset
to remap_vmalloc_range_partial() instead of adding it to the pointer in
remap_vmalloc_range().
In remap_vmalloc_range_partial(), fix the check against
get_vm_area_size() by using size comparisons instead of pointer
comparisons, and add checks for pgoff.
Fixes: 833423143c3a ("[PATCH] mm: introduce remap_vmalloc_range()") Signed-off-by: Jann Horn <jannh@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Cc: stable@vger.kernel.org Cc: Alexei Starovoitov <ast@kernel.org> Cc: Daniel Borkmann <daniel@iogearbox.net> Cc: Martin KaFai Lau <kafai@fb.com> Cc: Song Liu <songliubraving@fb.com> Cc: Yonghong Song <yhs@fb.com> Cc: Andrii Nakryiko <andriin@fb.com> Cc: John Fastabend <john.fastabend@gmail.com> Cc: KP Singh <kpsingh@chromium.org> Link: http://lkml.kernel.org/r/20200415222312.236431-1-jannh@google.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit 8099f58f1ecd ("USB: hub: Don't record a connect-change event
during reset-resume") wasn't very well conceived. The problem it
tried to fix was that if a connect-change event occurred while the
system was asleep (such as a device disconnecting itself from the bus
when it is suspended and then reconnecting when it resumes)
requiring a reset-resume during the system wakeup transition, the hub
port's change_bit entry would remain set afterward. This would cause
the hub driver to believe another connect-change event had occurred
after the reset-resume, which was wrong and would lead the driver to
send unnecessary requests to the device (which could interfere with a
firmware update).
The commit tried to fix this by not setting the change_bit during the
wakeup. But this was the wrong thing to do; it means that when a
device is unplugged while the system is asleep, the hub driver doesn't
realize anything has happened: The change_bit flag which would tell it
to handle the disconnect event is clear.
The commit needs to be reverted and the problem fixed in a different
way. Fortunately an alternative solution was noted in the commit's
Changelog: We can continue to set the change_bit entry in
hub_activate() but then clear it when a reset-resume occurs. That way
the the hub driver will see the change_bit when a device is
disconnected but won't see it when the device is still present.
That's what this patch does.
Reported-and-tested-by: Peter Chen <peter.chen@nxp.com> Signed-off-by: Alan Stern <stern@rowland.harvard.edu> Fixes: 8099f58f1ecd ("USB: hub: Don't record a connect-change event during reset-resume") Tested-by: Paul Zimmerman <pauldzim@gmail.com> CC: <stable@vger.kernel.org> Link: https://lore.kernel.org/r/Pine.LNX.4.44L0.2004221602480.11262-100000@iolanthe.rowland.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
FuzzUSB (a variant of syzkaller) found a free-while-still-in-use bug
in the USB scatter-gather library:
BUG: KASAN: use-after-free in atomic_read
include/asm-generic/atomic-instrumented.h:26 [inline]
BUG: KASAN: use-after-free in usb_hcd_unlink_urb+0x5f/0x170
drivers/usb/core/hcd.c:1607
Read of size 4 at addr ffff888065379610 by task kworker/u4:1/27
This bug occurs when cancellation of the S-G transfer races with
transfer completion. When that happens, usb_sg_cancel() may continue
to access the transfer's URBs after usb_sg_wait() has freed them.
The bug is caused by the fact that usb_sg_cancel() does not take any
sort of reference to the transfer, and so there is nothing to prevent
the URBs from being deallocated while the routine is trying to use
them. The fix is to take such a reference by incrementing the
transfer's io->count field while the cancellation is in progres and
decrementing it afterward. The transfer's URBs are not deallocated
until io->complete is triggered, which happens when io->count reaches
zero.
This fixes a bug that causes the USB3 early console to freeze after
printing a single line on AMD machines because it can't parse the
Transfer TRB properly.
The spec at
https://www.intel.com/content/dam/www/public/us/en/documents/technical-specifications/extensible-host-controler-interface-usb-xhci.pdf
says in section "4.5.1 Device Context Index" that the Context Index,
also known as Endpoint ID according to
section "1.6 Terms and Abbreviations", is normally computed as
`DCI = (Endpoint Number * 2) + Direction`, which matches the current
definitions of XDBC_EPID_OUT and XDBC_EPID_IN.
However, the numbering in a Debug Capability Context data structure is
supposed to be different:
Section "7.6.3.2 Endpoint Contexts and Transfer Rings" explains that a
Debug Capability Context data structure has the endpoints mapped to indices
0 and 1.
Change XDBC_EPID_OUT/XDBC_EPID_IN to the spec-compliant values, add
XDBC_EPID_OUT_INTEL/XDBC_EPID_IN_INTEL with Intel's incorrect values, and
let xdbc_handle_tx_event() handle both.
I have verified that with this patch applied, the USB3 early console works
on both an Intel and an AMD machine.
The Corsair K70 RGB RAPIDFIRE needs the USB_QUIRK_DELAY_INIT and
USB_QUIRK_DELAY_CTRL_MSG to function or it will randomly not
respond on boot, just like other Corsair keyboards
Change a bunch of arguments of wrapper functions which pass signed
integer to an unsigned integer which might cause undefined behaviors
when sign integer overflow.
A race condition between threads updating mountpoint reference counter
affects longterm releases 4.4.220, 4.9.220, 4.14.177 and 4.19.118.
The mountpoint reference counter corruption may occur when:
* one thread increments m_count member of struct mountpoint
[under namespace_sem, but not holding mount_lock]
pivot_root()
* another thread simultaneously decrements the same m_count
[under mount_lock, but not holding namespace_sem]
put_mountpoint()
unhash_mnt()
umount_mnt()
mntput_no_expire()
To fix this race condition, grab mount_lock before updating m_count in
pivot_root().
Reference: CVE-2020-12114 Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Piotr Krysiuk <piotras@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The XADC supports a samplerate of up to 1MSPS. Unfortunately the hardware
does not have a FIFO, which means it generates an interrupt for each
conversion sequence. At one 1MSPS this creates an interrupt storm that
causes the system to soft-lock.
For this reason the driver limits the maximum samplerate to 150kSPS.
Currently this check is only done when setting a new samplerate. But it is
also possible that the initial samplerate configured in the FPGA bitstream
exceeds the limit.
In this case when starting to capture data without first changing the
samplerate the system can overload.
To prevent this check the currently configured samplerate in the probe
function and reduce it to the maximum if necessary.
The XADC has two internal ADCs. Depending on the mode it is operating in
either one or both of them are used. The device manual calls this
continuous (one ADC) and simultaneous (both ADCs) mode.
The meaning of the sequencing register for the aux channels changes
depending on the mode.
In continuous mode each bit corresponds to one of the 16 aux channels. And
the single ADC will convert them one by one in order.
In simultaneous mode the aux channels are split into two groups the first 8
channels are assigned to the first ADC and the other 8 channels to the
second ADC. The upper 8 bits of the sequencing register are unused and the
lower 8 bits control both ADCs. This means a bit needs to be set if either
the corresponding channel from the first group or the second group (or
both) are set.
Currently the driver does not have the special handling required for
simultaneous mode. Add it.
When enabling the trigger and unmasking the end-of-sequence (EOS) interrupt
the EOS interrupt should be cleared from the status register. Otherwise it
is possible that it was still set from a previous capture. If that is the
case the interrupt would fire immediately even though no conversion has
been done yet and stale data is being read from the device.
The old code only clears the interrupt if the interrupt was previously
unmasked. Which does not make much sense since the interrupt is always
masked at this point and in addition masking the interrupt does not clear
the interrupt from the status register. So the clearing needs to be done
unconditionally.
The check for shutting down the second ADC is inverted. This causes it to
be powered down when it should be enabled. As a result channels that are
supposed to be handled by the second ADC return invalid conversion results.
Indeed, relying on addr being not 0 cannot work because some device have
their register to set odr at address 0. As a matter of fact, if the odr
can be set, then there is a mask.
Sensors with ODR register at address 0 are: lsm303dlh, lsm303dlhc, lsm303dlm
Fixes: 7d245172675a ("iio: common: st_sensors: check odr address value in st_sensors_set_odr()") Signed-off-by: Lary Gibaud <yarl-baudig@mailoo.org> Cc: <Stable@vger.kernel.org> Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This change removes the semi-colon from the devm_iio_device_register()
macro which seems to have been added by accident.
Fixes: 63b19547cc3d9 ("iio: Use macro magic to avoid manual assign of driver_module") Signed-off-by: Lars Engebretsen <lars@engebretsen.ch> Cc: <Stable@vger.kernel.org> Reviewed-by: Alexandru Ardelean <alexandru.ardelean@analog.com> Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
It turned out that ALC1220-VB USB-audio device gives the interrupt
event to some PCM terminals while those don't allow the connector
state request but only the actual I/O terminals return the request.
The recent commit 7dc3c5a0172e ("ALSA: usb-audio: Don't create jack
controls for PCM terminals") excluded those phantom terminals, so
those events are ignored, too.
My first thought was that this could be easily deduced from the
associated terminals, but some of them have even no associate terminal
ID, hence it's not too trivial to figure out.
Since the number of such terminals are small and limited, this patch
implements another quirk table for the simple mapping of the
connectors. It's not really scalable, but let's hope that there will
be not many such funky devices in future.
TRX40 mobos from MSI and others with ALC1220-VB USB-audio device need
yet more quirks for the proper control names.
This patch provides the mapping table for those boards, correcting the
FU names for volume and mute controls as well as the terminal names
for jack controls. It also improves build_connector_control() not to
add the directional suffix blindly if the string is given from the
mapping table.
With this patch applied, the new UCM profiles will be effective.
The commit 3c6fd1f07ed0 ("ALSA: hda: Add driver blacklist") added a
new blacklist for the devices that are known to have empty codecs, and
one of the entries was ASUS ROG Zenith II (PCI SSID 1043:874f).
However, it turned out that the very same PCI SSID is used for the
previous model that does have the valid HD-audio codecs and the change
broke the sound on it.
This patch reverts the corresponding entry as a temporary solution.
Although Zenith II and co will see get the empty HD-audio bus again,
it'd be merely resource wastes and won't affect the functionality,
so it's no end of the world. We'll need to address this later,
e.g. by either switching to DMI string matching or using PCI ID &
SSID pairs.
By allocating a kernel buffer with a user-supplied buffer length, it
is possible that a false positive ENOMEM error may be returned because
the user-supplied length is just too large even if the system do have
enough memory to hold the actual key data.
Moreover, if the buffer length is larger than the maximum amount of
memory that can be returned by kmalloc() (2^(MAX_ORDER-1) number of
pages), a warning message will also be printed.
To reduce this possibility, we set a threshold (PAGE_SIZE) over which we
do check the actual key length first before allocating a buffer of the
right size to hold it. The threshold is arbitrary, it is just used to
trigger a buffer length check. It does not limit the actual key length
as long as there is enough memory to satisfy the memory request.
To further avoid large buffer allocation failure due to page
fragmentation, kvmalloc() is used to allocate the buffer so that vmapped
pages can be used when there is not a large enough contiguous set of
pages available for allocation.
In the extremely unlikely scenario that the key keeps on being changed
and made longer (still <= buflen) in between 2 __keyctl_read_key()
calls, the __keyctl_read_key() calling loop in keyctl_read_key() may
have to be iterated a large number of times, but definitely not infinite.
Signed-off-by: Waiman Long <longman@redhat.com> Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The mlxsw_sp_acl_rulei_create() function is supposed to return an error
pointer from mlxsw_afa_block_create(). The problem is that these
functions both return NULL instead of error pointers. Half the callers
expect NULL and half expect error pointers so it could lead to a NULL
dereference on failure.
This patch changes both of them to return error pointers and changes all
the callers which checked for NULL to check for IS_ERR() instead.
Fixes: 4cda7d8d7098 ("mlxsw: core: Introduce flexible actions support") Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Reviewed-by: Ido Schimmel <idosch@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To avoid a loop with qdiscs and xfrms, check if the skb has already gone
through the qdisc attached to the VRF device and then to the xfrm layer.
If so, no need for a second redirect.
Fixes: 193125dbd8eb ("net: Introduce VRF device driver") Reported-by: Trev Larock <trev@larock.ca> Signed-off-by: David Ahern <dsahern@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
IPSKB_XFRM_TRANSFORMED and IP6SKB_XFRM_TRANSFORMED are skb flags set by
xfrm code to tell other skb handlers that the packet has been passed
through the xfrm output functions. Simplify the code and just always
set them rather than conditionally based on netfilter enabled thus
making the flag available for other users.
Signed-off-by: David Ahern <dsahern@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When asking the ARL to read a MAC address, we will get a number of bins
returned in a single read. Out of those bins, there can essentially be 3
states:
- all bins are full, we have no space left, and we can either replace an
existing address or return that full condition
- the MAC address was found, then we need to return its bin index and
modify that one, and only that one
- the MAC address was not found and we have a least one bin free, we use
that bin index location then
The code would unfortunately fail on all counts.
Fixes: 1da6df85c6fb ("net: dsa: b53: Implement ARL add/del/dump operations") Signed-off-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The ARL {MAC,VID} tuple and the forward entry were off by 0x10 bytes,
which means that when we read/wrote from/to ARL bin index 0, we were
actually accessing the ARLA_RWCTRL register.
Fixes: 1da6df85c6fb ("net: dsa: b53: Implement ARL add/del/dump operations") Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: Florian Fainelli <f.fainelli@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When a qdisc is attached to the VRF device, the packet goes down the ndo
xmit function which is setup to send the packet back to the VRF driver
which does a lookup to send the packet out. The lookup in the VRF driver
is not considering xfrm policies. Change it to use ip6_dst_lookup_flow
rather than ip6_route_output.
Fixes: 35402e313663 ("net: Add IPv6 support to VRF device") Signed-off-by: David Ahern <dsahern@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When team mode is changed or set, the team_mode_get() is called to check
whether the mode module is inserted or not. If the mode module is not
inserted, it calls the request_module().
In the request_module(), it creates a child process, which is
the "modprobe" process and waits for the done of the child process.
At this point, the following locks were used.
down_read(&cb_lock()); by genl_rcv()
genl_lock(); by genl_rcv_msc()
rtnl_lock(); by team_nl_cmd_options_set()
mutex_lock(&team->lock); by team_nl_team_get()
Concurrently, the team module could be removed by rmmod or "modprobe -r"
The __exit function of team module is team_module_exit(), which calls
team_nl_fini() and it tries to acquire following locks.
down_write(&cb_lock);
genl_lock();
Because of the genl_lock() and cb_lock, this process can't be finished
earlier than request_module() routine.
The problem secenario.
CPU0 CPU1
team_mode_get
request_module()
modprobe -r team_mode_roundrobin
team <--(B)
modprobe team <--(A)
team_mode_roundrobin
By request_module(), the "modprobe team_mode_roundrobin" command
will be executed. At this point, the modprobe process will decide
that the team module should be inserted before team_mode_roundrobin.
Because the team module is being removed.
By the module infrastructure, the same module insert/remove operations
can't be executed concurrently.
So, (A) waits for (B) but (B) also waits for (A) because of locks.
So that the hang occurs at this point.
Test commands:
while :
do
teamd -d &
killall teamd &
modprobe -rv team_mode_roundrobin &
done
The approach of this patch is to hold the reference count of the team
module if the team module is compiled as a module. If the reference count
of the team module is not zero while request_module() is being called,
the team module will not be removed at that moment.
So that the above scenario could not occur.
Fixes: 3d249d4ca7d0 ("net: introduce ethernet teaming device") Signed-off-by: Taehee Yoo <ap420073@gmail.com> Reviewed-by: Jiri Pirko <jiri@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
skb->sk does not always point to a full blown socket,
we need to use sk_fullsock() before accessing fields which
only make sense on full socket.
BUG: KASAN: use-after-free in report_sock_error+0x286/0x300 net/sched/sch_etf.c:141
Read of size 1 at addr ffff88805eb9b245 by task syz-executor.5/9630
Fixes: 4b15c7075352 ("net/sched: Make etf report drops on error_queue") Fixes: 25db26a91364 ("net/sched: Introduce the ETF Qdisc") Signed-off-by: Eric Dumazet <edumazet@google.com> Reported-by: syzbot <syzkaller@googlegroups.com> Cc: Vinicius Costa Gomes <vinicius.gomes@intel.com> Reviewed-by: Vinicius Costa Gomes <vinicius.gomes@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
x25_lapb_receive_frame() invokes x25_get_neigh(), which returns a
reference of the specified x25_neigh object to "nb" with increased
refcnt.
When x25_lapb_receive_frame() returns, local variable "nb" becomes
invalid, so the refcount should be decreased to keep refcount balanced.
The reference counting issue happens in one path of
x25_lapb_receive_frame(). When pskb_may_pull() returns false, the
function forgets to decrease the refcnt increased by x25_get_neigh(),
causing a refcnt leak.
Fix this issue by calling x25_neigh_put() when pskb_may_pull() returns
false.
Fixes: cb101ed2c3c7 ("x25: Handle undersized/fragmented skbs") Signed-off-by: Xiyu Yang <xiyuyang19@fudan.edu.cn> Signed-off-by: Xin Tan <tanxin.ctf@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The buggy address belongs to the variable:
div_table.63646+0x34/0xfffffffffffffa40 [dwmac_meson8b]
Memory state around the buggy address: ffffa00009061480: fa fa fa fa 00 00 00 01 fa fa fa fa 00 00 00 00 ffffa00009061500: 05 fa fa fa fa fa fa fa 00 04 fa fa fa fa fa fa
>ffffa00009061580: 00 03 fa fa fa fa fa fa 00 00 00 00 00 00 fa fa
^ ffffa00009061600: fa fa fa fa 00 01 fa fa fa fa fa fa 01 fa fa fa ffffa00009061680: fa fa fa fa 00 01 fa fa fa fa fa fa 04 fa fa fa
==================================================================
Digging into this indeed shows that the clock divider array is
lacking a final fence, and that the clock subsystems goes in the
weeds. Oh well.
Let's add the empty structure that indicates the end of the array.
Fixes: bd6f48546b9c ("net: stmmac: dwmac-meson8b: Fix the RGMII TX delay on Meson8b/8m2 SoCs") Signed-off-by: Marc Zyngier <maz@kernel.org> Cc: Martin Blumenstingl <martin.blumenstingl@googlemail.com> Reviewed-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
nr_add_node() invokes nr_neigh_get_dev(), which returns a local
reference of the nr_neigh object to "nr_neigh" with increased refcnt.
When nr_add_node() returns, "nr_neigh" becomes invalid, so the refcount
should be decreased to keep refcount balanced.
The issue happens in one normal path of nr_add_node(), which forgets to
decrease the refcnt increased by nr_neigh_get_dev() and causes a refcnt
leak. It should decrease the refcnt before the function returns like
other normal paths do.
Fix this issue by calling nr_neigh_put() before the nr_add_node()
returns.
Signed-off-by: Xiyu Yang <xiyuyang19@fudan.edu.cn> Signed-off-by: Xin Tan <tanxin.ctf@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In the macvlan_device_event(), the list_first_entry_or_null() is used.
This function could return null pointer if there is no node.
But, the macvlan module doesn't check the null pointer.
So, null-ptr-deref would occur.
The problem scenario.
If dummy1 is removed,
1. ->dellink() of dummy1 is called.
2. NETDEV_UNREGISTER of dummy1 notification is sent to macvlan module.
3. ->dellink() of macvlan1 is called.
4. NETDEV_UNREGISTER of macvlan1 notification is sent to bond module.
5. __bond_release_one() is called and it internally calls
dev_set_mac_address().
6. dev_set_mac_address() calls the ->ndo_set_mac_address() of macvlan1,
which is macvlan_set_mac_address().
7. macvlan_set_mac_address() calls the dev_set_mac_address() with dummy1.
8. NETDEV_CHANGEADDR of dummy1 is sent to macvlan module.
9. In the macvlan_device_event(), it calls list_first_entry_or_null().
At this point, dummy1 and macvlan1 were removed.
So, list_first_entry_or_null() will return NULL.
Test commands:
ip netns add nst
ip netns exec nst ip link add bond0 type bond
for i in {0..10}
do
ip netns exec nst ip link add dummy$i type dummy
ip netns exec nst ip link add macvlan$i link dummy$i \
type macvlan mode passthru
ip netns exec nst ip link set macvlan$i master bond0
done
ip netns del nst
Fixes: e289fd28176b ("macvlan: fix the problem when mac address changes for passthru mode") Reported-by: syzbot+5035b1f9dc7ea4558d5a@syzkaller.appspotmail.com Signed-off-by: Taehee Yoo <ap420073@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When a macsec interface is created, the mtu is calculated with the lower
interface's mtu value.
If the mtu of lower interface is lower than the length, which is needed
by macsec interface, macsec's mtu value will be overflowed.
So, if the lower interface's mtu is too low, macsec interface's mtu
should be set to 0.
Test commands:
ip link add dummy0 mtu 10 type dummy
ip link add macsec0 link dummy0 type macsec
ip link show macsec0
Before:
11: macsec0@dummy0: <BROADCAST,MULTICAST,M-DOWN> mtu 4294967274
After:
11: macsec0@dummy0: <BROADCAST,MULTICAST,M-DOWN> mtu 0
Fixes: c09440f7dcb3 ("macsec: introduce IEEE 802.1AE driver") Signed-off-by: Taehee Yoo <ap420073@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit b6f6118901d1 ("ipv6: restrict IPV6_ADDRFORM operation") fixed a
problem found by syzbot an unfortunate logic error meant that it
also broke IPV6_ADDRFORM.
Rearrange the checks so that the earlier test is just one of the series
of checks made before moving the socket from IPv6 to IPv4.
Fixes: b6f6118901d1 ("ipv6: restrict IPV6_ADDRFORM operation") Signed-off-by: John Haxby <john.haxby@oracle.com> Cc: stable@vger.kernel.org Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fetching PTP sync information from mailbox is slow and can take
up to 10 milliseconds. Reduce this unnecessary delay by directly
reading the information from the corresponding registers.
Fixes: 9c33e4208bce ("cxgb4: Add PTP Hardware Clock (PHC) support") Signed-off-by: Manoj Malviya <manojmalviya@chelsio.com> Signed-off-by: Rahul Lakkireddy <rahul.lakkireddy@chelsio.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In the absence of MC1, the size calculation function
cudbg_mem_region_size() was returing wrong MC size and
resulted in adapter crash. This patch adds new argument
to cudbg_mem_region_size() which will have actual size
and returns error to caller in the absence of MC1.
Fixes: a1c69520f785 ("cxgb4: collect MC memory dump") Signed-off-by: Vishal Kulkarni <vishal@chelsio.com>" Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Now that we are mapping kvm_steal_time from the guest directly we
don't need keep a copy of it in kvm_vcpu_arch.st. The same is true
for the stime field.
This is part of CVE-2019-3016.
Signed-off-by: Boris Ostrovsky <boris.ostrovsky@oracle.com> Reviewed-by: Joao Martins <joao.m.martins@oracle.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Ben Hutchings <ben.hutchings@codethink.co.uk> Signed-off-by: Sasha Levin <sashal@kernel.org>
There is a potential race in record_steal_time() between setting
host-local vcpu->arch.st.steal.preempted to zero (i.e. clearing
KVM_VCPU_PREEMPTED) and propagating this value to the guest with
kvm_write_guest_cached(). Between those two events the guest may
still see KVM_VCPU_PREEMPTED in its copy of kvm_steal_time, set
KVM_VCPU_FLUSH_TLB and assume that hypervisor will do the right
thing. Which it won't.
Instad of copying, we should map kvm_steal_time and that will
guarantee atomicity of accesses to @preempted.
This is part of CVE-2019-3016.
Signed-off-by: Boris Ostrovsky <boris.ostrovsky@oracle.com> Reviewed-by: Joao Martins <joao.m.martins@oracle.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
[bwh: Backported to 4.19: No tracepoint in record_steal_time().] Signed-off-by: Ben Hutchings <ben.hutchings@codethink.co.uk> Signed-off-by: Sasha Levin <sashal@kernel.org>
kvm_vcpu_(un)map operates on gfns from any current address space.
In certain cases we want to make sure we are not mapping SMRAM
and for that we can use kvm_(un)map_gfn() that we are introducing
in this patch.
This is part of CVE-2019-3016.
Signed-off-by: Boris Ostrovsky <boris.ostrovsky@oracle.com> Reviewed-by: Joao Martins <joao.m.martins@oracle.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Ben Hutchings <ben.hutchings@codethink.co.uk> Signed-off-by: Sasha Levin <sashal@kernel.org>
The field "page" is initialized to KVM_UNMAPPED_PAGE when it is not used
(i.e. when the memory lives outside kernel control). So this check will
always end up using kunmap even for memremap regions.
Fixes: e45adf665a53 ("KVM: Introduce a new guest mapping API") Signed-off-by: KarimAllah Ahmed <karahmed@amazon.de> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Ben Hutchings <ben.hutchings@codethink.co.uk> Signed-off-by: Sasha Levin <sashal@kernel.org>
Commit e45adf665a53 ("KVM: Introduce a new guest mapping API", 2019-01-31)
introduced a build failure on aarch64 defconfig:
$ make -j$(nproc) ARCH=arm64 CROSS_COMPILE=aarch64-linux-gnu- O=out defconfig \
Image.gz
...
../arch/arm64/kvm/../../../virt/kvm/kvm_main.c:
In function '__kvm_map_gfn':
../arch/arm64/kvm/../../../virt/kvm/kvm_main.c:1763:9: error:
implicit declaration of function 'memremap'; did you mean 'memset_p'?
../arch/arm64/kvm/../../../virt/kvm/kvm_main.c:1763:46: error:
'MEMREMAP_WB' undeclared (first use in this function)
../arch/arm64/kvm/../../../virt/kvm/kvm_main.c:
In function 'kvm_vcpu_unmap':
../arch/arm64/kvm/../../../virt/kvm/kvm_main.c:1795:3: error:
implicit declaration of function 'memunmap'; did you mean 'vm_munmap'?
because these functions are declared in <linux/io.h> rather than <asm/io.h>,
and the former was being pulled in already on x86 but not on aarch64.
Reported-by: Nathan Chancellor <natechancellor@gmail.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
[bwh: Backported to 4.19: adjust context] Signed-off-by: Ben Hutchings <ben.hutchings@codethink.co.uk> Signed-off-by: Sasha Levin <sashal@kernel.org>
In addition to all this unnecessarily noise in the code due to boiler plate
code, most of the time the mapping function does not properly handle memory
that is not backed by "struct page". This new guest mapping API encapsulate
most of this boiler plate code and also handles guest memory that is not
backed by "struct page".
The current implementation of this API is using memremap for memory that is
not backed by a "struct page" which would lead to a huge slow-down if it
was used for high-frequency mapping operations. The API does not have any
effect on current setups where guest memory is backed by a "struct page".
Further patches are going to also introduce a pfn-cache which would
significantly improve the performance of the memremap case.
Signed-off-by: KarimAllah Ahmed <karahmed@amazon.de> Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
[bwh: Backported to 4.19 as dependency of commit 1eff70a9abd4
"x86/kvm: Introduce kvm_(un)map_gfn()"] Signed-off-by: Ben Hutchings <ben.hutchings@codethink.co.uk> Signed-off-by: Sasha Levin <sashal@kernel.org>
If L1 does not set VM_ENTRY_LOAD_BNDCFGS, then L1's BNDCFGS value must
be propagated to vmcs02 since KVM always runs with VM_ENTRY_LOAD_BNDCFGS
when MPX is supported. Because the value effectively comes from vmcs01,
vmcs02 must be updated even if vmcs12 is clean.
Fixes: 62cf9bd8118c4 ("KVM: nVMX: Fix emulation of VM_ENTRY_LOAD_BNDCFGS") Cc: Liran Alon <liran.alon@oracle.com> Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
[bwh: Backported to 4.19: adjust filename, context] Signed-off-by: Ben Hutchings <ben.hutchings@codethink.co.uk> Signed-off-by: Sasha Levin <sashal@kernel.org>
...except RSP, which is restored by hardware as part of VM-Exit.
Paolo theorized that restoring registers from the stack after a VM-Exit
in lieu of zeroing them could lead to speculative execution with the
guest's values, e.g. if the stack accesses miss the L1 cache[1].
Zeroing XORs are dirt cheap, so just be ultra-paranoid.
Note that the scratch register (currently RCX) used to save/restore the
guest state is also zeroed as its host-defined value is loaded via the
stack, just with a MOV instead of a POP.
In f2fs_listxattr, there is no boundary check before
memcpy e_name to buffer.
If the e_name_len is corrupted,
unexpected memory contents may be returned to the buffer.
Signed-off-by: Randall Huang <huangrandall@google.com> Reviewed-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
[bwh: Backported to 4.19: Use f2fs_msg() instead of f2fs_err()] Signed-off-by: Ben Hutchings <ben.hutchings@codethink.co.uk> Signed-off-by: Sasha Levin <sashal@kernel.org>
There was a recent change in blktrace.c that added a RCU protection to
`q->blk_trace` in order to fix a use-after-free issue during access.
However the change missed an edge case that can lead to dereferencing of
`bt` pointer even when it's NULL:
Coverity static analyzer marked this as a FORWARD_NULL issue with CID 1460458.
```
/kernel/trace/blktrace.c: 1904 in sysfs_blk_trace_attr_store()
1898 ret = 0;
1899 if (bt == NULL)
1900 ret = blk_trace_setup_queue(q, bdev);
1901
1902 if (ret == 0) {
1903 if (attr == &dev_attr_act_mask)
>>> CID 1460458: Null pointer dereferences (FORWARD_NULL)
>>> Dereferencing null pointer "bt".
1904 bt->act_mask = value;
1905 else if (attr == &dev_attr_pid)
1906 bt->pid = value;
1907 else if (attr == &dev_attr_start_lba)
1908 bt->start_lba = value;
1909 else if (attr == &dev_attr_end_lba)
```
Added a reassignment with RCU annotation to fix the issue.
Fixes: c780e86dd48 ("blktrace: Protect q->blk_trace with RCU") Reviewed-by: Ming Lei <ming.lei@redhat.com> Reviewed-by: Bob Liu <bob.liu@oracle.com> Reviewed-by: Steven Rostedt (VMware) <rostedt@goodmis.org> Signed-off-by: Cengiz Can <cengiz@kernel.wtf> Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Ben Hutchings <ben.hutchings@codethink.co.uk> Signed-off-by: Sasha Levin <sashal@kernel.org>
KASAN is reporting that __blk_add_trace() has a use-after-free issue
when accessing q->blk_trace. Indeed the switching of block tracing (and
thus eventual freeing of q->blk_trace) is completely unsynchronized with
the currently running tracing and thus it can happen that the blk_trace
structure is being freed just while __blk_add_trace() works on it.
Protect accesses to q->blk_trace by RCU during tracing and make sure we
wait for the end of RCU grace period when shutting down tracing. Luckily
that is rare enough event that we can afford that. Note that postponing
the freeing of blk_trace to an RCU callback should better be avoided as
it could have unexpected user visible side-effects as debugfs files
would be still existing for a short while block tracing has been shut
down.
Link: https://bugzilla.kernel.org/show_bug.cgi?id=205711 CC: stable@vger.kernel.org Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com> Reviewed-by: Ming Lei <ming.lei@redhat.com> Tested-by: Ming Lei <ming.lei@redhat.com> Reviewed-by: Bart Van Assche <bvanassche@acm.org> Reported-by: Tristan Madani <tristmd@gmail.com> Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Jens Axboe <axboe@kernel.dk>
[bwh: Backported to 4.19: adjust context] Signed-off-by: Ben Hutchings <ben.hutchings@codethink.co.uk> Signed-off-by: Sasha Levin <sashal@kernel.org>
ipv6_stub uses the ip6_dst_lookup function to allow other modules to
perform IPv6 lookups. However, this function skips the XFRM layer
entirely.
All users of ipv6_stub->ip6_dst_lookup use ip_route_output_flow (via the
ip_route_output_key and ip_route_output helpers) for their IPv4 lookups,
which calls xfrm_lookup_route(). This patch fixes this inconsistent
behavior by switching the stub to ip6_dst_lookup_flow, which also calls
xfrm_lookup_route().
This requires some changes in all the callers, as these two functions
take different arguments and have different return types.
Fixes: 5f81bd2e5d80 ("ipv6: export a stub for IPv6 symbols used by vxlan") Reported-by: Xiumei Mu <xmu@redhat.com> Signed-off-by: Sabrina Dubroca <sd@queasysnail.net> Signed-off-by: David S. Miller <davem@davemloft.net>
[bwh: Backported to 4.19:
- Drop change in lwt_bpf.c
- Delete now-unused "ret" in mlx5e_route_lookup_ipv6()
- Initialise "out_dev" in mlx5e_create_encap_header_ipv6() to avoid
introducing a spurious "may be used uninitialised" warning
- Adjust filenames, context, indentation] Signed-off-by: Ben Hutchings <ben.hutchings@codethink.co.uk> Signed-off-by: Sasha Levin <sashal@kernel.org>
This will be used in the conversion of ipv6_stub to ip6_dst_lookup_flow,
as some modules currently pass a net argument without a socket to
ip6_dst_lookup. This is equivalent to commit 343d60aada5a ("ipv6: change
ipv6_stub_impl.ipv6_dst_lookup to take net argument").
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net> Signed-off-by: David S. Miller <davem@davemloft.net>
[bwh: Backported to 4.19: adjust context] Signed-off-by: Ben Hutchings <ben.hutchings@codethink.co.uk> Signed-off-by: Sasha Levin <sashal@kernel.org>
Previously Clock PM could not be re-enabled after being disabled by
pci_disable_link_state() because clkpm_capable was reset. Change this by
adding a clkpm_disable field similar to aspm_disable.
Use sas_phy_delete rather than sas_phy_free which, according to
comments, should not be called for PHYs that have been set up
successfully.
Link: https://lore.kernel.org/r/157048748876.11757.17773443136670011786.stgit@brunhilda Reviewed-by: Scott Benesh <scott.benesh@microsemi.com> Reviewed-by: Scott Teel <scott.teel@microsemi.com> Reviewed-by: Kevin Barnett <kevin.barnett@microsemi.com> Signed-off-by: Murthy Bhat <Murthy.Bhat@microsemi.com> Signed-off-by: Don Brace <don.brace@microsemi.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Let's change the mapping between virtqueue_add errors to BLK_STS
statuses, so that -ENOSPC, which indicates virtqueue full is still
mapped to BLK_STS_DEV_RESOURCE, but -ENOMEM which indicates non-device
specific resource outage is mapped to BLK_STS_RESOURCE.
Signed-off-by: Halil Pasic <pasic@linux.ibm.com> Link: https://lore.kernel.org/r/20200213123728.61216-3-pasic@linux.ibm.com Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
As the ftrace selftests can run for a long period of time, disable the
timeout that the general selftests have. If a selftest hangs, then it
probably means the machine will hang too.
Link: https://lore.kernel.org/r/alpine.LSU.2.21.1911131604170.18679@pobox.suse.cz Suggested-by: Miroslav Benes <mbenes@suse.cz> Tested-by: Miroslav Benes <mbenes@suse.cz> Reviewed-by: Miroslav Benes <mbenes@suse.cz> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org> Signed-off-by: Sasha Levin <sashal@kernel.org>