David Thompson [Wed, 15 Sep 2021 18:08:48 +0000 (14:08 -0400)]
mlxbf_gige: clear valid_polarity upon open
The network interface managed by the mlxbf_gige driver can
get into a problem state where traffic does not flow.
In this state, the interface will be up and enabled, but
will stop processing received packets. This problem state
will happen if three specific conditions occur:
1) driver has received more than (N * RxRingSize) packets but
less than (N+1 * RxRingSize) packets, where N is an odd number
Note: the command "ethtool -g <interface>" will display the
current receive ring size, which currently defaults to 128
2) the driver's interface was disabled via "ifconfig oob_net0 down"
during the window described in #1.
3) the driver's interface is re-enabled via "ifconfig oob_net0 up"
This patch ensures that the driver's "valid_polarity" field is
cleared during the open() method so that it always matches the
receive polarity used by hardware. Without this fix, the driver
needs to be unloaded and reloaded to correct this problem state.
Fixes: f92e1869d74e ("Add Mellanox BlueField Gigabit Ethernet driver") Reviewed-by: Asmaa Mnebhi <[email protected]> Signed-off-by: David Thompson <[email protected]> Signed-off-by: David S. Miller <[email protected]>
Paolo Abeni [Wed, 15 Sep 2021 17:19:07 +0000 (10:19 -0700)]
igc: fix tunnel offloading
Checking tunnel offloading, it turns out that offloading doesn't work
as expected. The following script allows to reproduce the issue.
Call it as `testscript DEVICE LOCALIP REMOTEIP NETMASK'
=== SNIP ===
if [ $# -ne 4 ]
then
echo "Usage $0 DEVICE LOCALIP REMOTEIP NETMASK"
exit 1
fi
DEVICE="$1"
LOCAL_ADDRESS="$2"
REMOTE_ADDRESS="$3"
NWMASK="$4"
echo "Driver: $(ethtool -i ${DEVICE} | awk '/^driver:/{print $2}') "
ethtool -k "${DEVICE}" | grep tx-udp
echo
echo "Set up NIC and tunnel..."
ip addr add "${LOCAL_ADDRESS}/${NWMASK}" dev "${DEVICE}"
ip link set "${DEVICE}" up
sleep 2
ip link add vxlan1 type vxlan id 42 \
remote "${REMOTE_ADDRESS}" \
local "${LOCAL_ADDRESS}" \
dstport 0 \
dev "${DEVICE}"
ip addr add fc00::1/64 dev vxlan1
ip link set vxlan1 up
sleep 2
rm -f vxlan.pcap
echo "Running tcpdump and iperf3..."
( nohup tcpdump -i any -w vxlan.pcap >/dev/null 2>&1 ) &
sleep 2
iperf3 -c fc00::2 >/dev/null
pkill tcpdump
echo
echo -n "Max. Paket Size: "
tcpdump -r vxlan.pcap -nnle 2>/dev/null \
| grep "${LOCAL_ADDRESS}.*> ${REMOTE_ADDRESS}.*OTV" \
| awk '{print $8}' | awk -F ':' '{print $1}' \
| sort -n | tail -1
echo
ip link del vxlan1
ip addr del ${LOCAL_ADDRESS}/${NWMASK} dev "${DEVICE}"
=== SNAP ===
The expected outcome is
Max. Paket Size: 64904
This is what you see on igb, the code igc has been taken from.
However, on igc the output is
Max. Paket Size: 1516
so the GSO aggregate packets are segmented by the kernel before calling
igc_xmit_frame. Inside the subsequent call to igc_tso, the check for
skb_is_gso(skb) fails and the function returns prematurely.
It turns out that this occurs because the feature flags aren't set
entirely correctly in igc_probe. In contrast to the original code
from igb_probe, igc_probe neglects to set the flags required to allow
tunnel offloading.
Setting the same flags as igb fixes the issue on igc.
Remove the assert from the callback priv lookup function since it does
not require RTNL lock and is already protected by flow_indr_block_lock.
This will avoid warnings from being emitted to dmesg if the driver
registers its callback after an ingress qdisc was created for a
netdevice.
The warnings started after the following patch was merged:
commit 74fc4f828769 ("net: Fix offloading indirect devices dependency on qdisc order creation")
Lucas Stach [Fri, 20 Aug 2021 20:18:30 +0000 (22:18 +0200)]
drm/etnaviv: add missing MMU context put when reaping MMU mapping
When we forcefully evict a mapping from the the address space and thus the
MMU context, the MMU context is leaked, as the mapping no longer points to
it, so it doesn't get freed when the GEM object is destroyed. Add the
mssing context put to fix the leak.
Lucas Stach [Fri, 20 Aug 2021 20:18:29 +0000 (22:18 +0200)]
drm/etnaviv: reference MMU context when setting up hardware state
Move the refcount manipulation of the MMU context to the point where the
hardware state is programmed. At that point it is also known if a previous
MMU state is still there, or the state needs to be reprogrammed with a
potentially different context.
Lucas Stach [Fri, 20 Aug 2021 20:18:28 +0000 (22:18 +0200)]
drm/etnaviv: fix MMU context leak on GPU reset
After a reset the GPU is no longer using the MMU context and may be
restarted with a different context. While the mmu_state proeprly was
cleared, the context wasn't unreferenced, leading to a memory leak.
Lucas Stach [Fri, 20 Aug 2021 20:18:27 +0000 (22:18 +0200)]
drm/etnaviv: exec and MMU state is lost when resetting the GPU
When the GPU is reset both the current exec state, as well as all MMU
state is lost. Move the driver side state tracking into the reset function
to keep hardware and software state from diverging.
Lucas Stach [Fri, 20 Aug 2021 20:18:26 +0000 (22:18 +0200)]
drm/etnaviv: keep MMU context across runtime suspend/resume
The MMU state may be kept across a runtime suspend/resume cycle, as we
avoid a full hardware reset to keep the latency of the runtime PM small.
Don't pretend that the MMU state is lost in driver state. The MMU
context is pushed out when new HW jobs with a different context are
coming in. The only exception to this is when the GPU is unbound, in
which case we need to make sure to also free the last active context.
Lucas Stach [Fri, 20 Aug 2021 20:18:25 +0000 (22:18 +0200)]
drm/etnaviv: stop abusing mmu_context as FE running marker
While the DMA frontend can only be active when the MMU context is set, the
reverse isn't necessarily true, as the frontend can be stopped while the
MMU state is kept. Stop treating mmu_context being set as a indication that
the frontend is running and instead add a explicit property.
Lucas Stach [Fri, 20 Aug 2021 20:18:24 +0000 (22:18 +0200)]
drm/etnaviv: put submit prev MMU context when it exists
The prev context is the MMU context at the time of the job
queueing in hardware. As a job might be queued multiple times
due to recovery after a GPU hang, we need to make sure to put
the stale prev MMU context from a prior queuing, to avoid the
reference and thus the MMU context leaking.
Merge tag 'hyperv-fixes-signed-20210915' of git://git.kernel.org/pub/scm/linux/kernel/git/hyperv/linux
Pull hyperv fixes from Wei Liu:
- Fix kernel crash caused by uio driver (Vitaly Kuznetsov)
- Remove on-stack cpumask from HV APIC code (Wei Liu)
* tag 'hyperv-fixes-signed-20210915' of git://git.kernel.org/pub/scm/linux/kernel/git/hyperv/linux:
x86/hyperv: remove on-stack cpumask from hv_send_ipi_mask_allbutself
asm-generic/hyperv: provide cpumask_to_vpset_noself
Drivers: hv: vmbus: Fix kernel crash upon unbinding a device from uio_hv_generic driver
Merge tag 'rtc-5.15-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/abelloni/linux
Pull RTC fix from Alexandre Belloni:
"Fix a locking issue in the cmos rtc driver"
* tag 'rtc-5.15-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/abelloni/linux:
rtc: cmos: Disable irq around direct invocation of cmos_interrupt()
Vladimir Oltean [Tue, 14 Sep 2021 13:47:26 +0000 (16:47 +0300)]
net: dsa: flush switchdev workqueue before tearing down CPU/DSA ports
Sometimes when unbinding the mv88e6xxx driver on Turris MOX, these error
messages appear:
mv88e6085 d0032004.mdio-mii:12: port 1 failed to delete be:79:b4:9e:9e:96 vid 1 from fdb: -2
mv88e6085 d0032004.mdio-mii:12: port 1 failed to delete be:79:b4:9e:9e:96 vid 0 from fdb: -2
mv88e6085 d0032004.mdio-mii:12: port 1 failed to delete d8:58:d7:00:ca:6d vid 100 from fdb: -2
mv88e6085 d0032004.mdio-mii:12: port 1 failed to delete d8:58:d7:00:ca:6d vid 1 from fdb: -2
mv88e6085 d0032004.mdio-mii:12: port 1 failed to delete d8:58:d7:00:ca:6d vid 0 from fdb: -2
(and similarly for other ports)
What happens is that DSA has a policy "even if there are bugs, let's at
least not leak memory" and dsa_port_teardown() clears the dp->fdbs and
dp->mdbs lists, which are supposed to be empty.
But deleting that cleanup code, the warnings go away.
=> the FDB and MDB lists (used for refcounting on shared ports, aka CPU
and DSA ports) will eventually be empty, but are not empty by the time
we tear down those ports. Aka we are deleting them too soon.
The addresses that DSA complains about are host-trapped addresses: the
local addresses of the ports, and the MAC address of the bridge device.
The problem is that offloading those entries happens from a deferred
work item scheduled by the SWITCHDEV_FDB_DEL_TO_DEVICE handler, and this
races with the teardown of the CPU and DSA ports where the refcounting
is kept.
In fact, not only it races, but fundamentally speaking, if we iterate
through the port list linearly, we might end up tearing down the shared
ports even before we delete a DSA user port which has a bridge upper.
So as it turns out, we need to first tear down the user ports (and the
unused ones, for no better place of doing that), then the shared ports
(the CPU and DSA ports). In between, we need to ensure that all work
items scheduled by our switchdev handlers (which only run for user
ports, hence the reason why we tear them down first) have finished.
Vladimir Oltean [Tue, 14 Sep 2021 14:05:15 +0000 (17:05 +0300)]
Revert "net: phy: Uniform PHY driver access"
This reverts commit 3ac8eed62596387214869319379c1fcba264d8c6, which did
more than it said on the box, and not only it replaced to_phy_driver
with phydev->drv, but it also removed the "!drv" check, without actually
explaining why that is fine.
That patch in fact breaks suspend/resume on any system which has PHY
devices with no drivers bound.
The stack trace is:
Unable to handle kernel NULL pointer dereference at virtual address 00000000000000e8
pc : mdio_bus_phy_suspend+0xd8/0xec
lr : dpm_run_callback+0x38/0x90
Call trace:
mdio_bus_phy_suspend+0xd8/0xec
dpm_run_callback+0x38/0x90
__device_suspend+0x108/0x3cc
dpm_suspend+0x140/0x210
dpm_suspend_start+0x7c/0xa0
suspend_devices_and_enter+0x13c/0x540
pm_suspend+0x2a4/0x330
Examples why that assumption is not fine:
- There is an MDIO bus with a PHY device that doesn't have a specific
PHY driver loaded, because mdiobus_register() automatically creates a
PHY device for it but there is no specific PHY driver in the system.
Normally under those circumstances, the generic PHY driver will be
bound lazily to it (at phy_attach_direct time). But some Ethernet
drivers attach to their PHY at .ndo_open time. Until then it, the
to-be-driven-by-genphy PHY device will not have a driver. The blamed
patch amounts to saying "you need to open all net devices before the
system can suspend, to avoid the NULL pointer dereference".
- There is any raw MDIO device which has 'plausible' values in the PHY
ID registers 2 and 3, which is located on an MDIO bus whose driver
does not set bus->phy_mask = ~0 (which prevents auto-scanning of PHY
devices). An example could be a MAC's internal MDIO bus with PCS
devices on it, for serial links such as SGMII. PHY devices will get
created for those PCSes too, due to that MDIO bus auto-scanning, and
although those PHY devices are not used, they do not bother anybody
either. PCS devices are usually managed in Linux as raw MDIO devices.
Nonetheless, they do not have a PHY driver, nor does anybody attempt
to connect to them (because they are not a PHY), and therefore this
patch breaks that.
The goal itself of the patch is questionable, so I am going for a
straight revert. to_phy_driver does not seem to have a need to be
replaced by phydev->drv, in fact that might even trigger code paths
which were not given too deep of a thought.
For instance:
phy_probe populates phydev->drv at the beginning, but does not clean it
up on any error (including EPROBE_DEFER). So if the phydev driver
requests probe deferral, phydev->drv will remain populated despite there
being no driver bound.
If a system suspend starts in between the initial probe deferral request
and the subsequent probe retry, we will be calling the phydev->drv->suspend
method, but _before_ any phydev->drv->probe call has succeeded.
That is to say, if the phydev->drv is allocating any driver-private data
structure in ->probe, it pretty much expects that data structure to be
available in ->suspend. But it may not. That is a pretty insane
environment to present to PHY drivers.
In the code structure before the blamed patch, mdio_bus_phy_may_suspend
would just say "no, don't suspend" to any PHY device which does not have
a driver pointer _in_the_device_structure_ (not the phydev->drv). That
would essentially ensure that ->suspend will never get called for a
device that has not yet successfully completed probe. This is the code
structure the patch is returning to, via the revert.
Vladimir Oltean [Tue, 14 Sep 2021 13:43:31 +0000 (16:43 +0300)]
net: dsa: destroy the phylink instance on any error in dsa_slave_phy_setup
DSA supports connecting to a phy-handle, and has a fallback to a non-OF
based method of connecting to an internal PHY on the switch's own MDIO
bus, if no phy-handle and no fixed-link nodes were present.
The -ENODEV error code from the first attempt (phylink_of_phy_connect)
is what triggers the second attempt (phylink_connect_phy).
However, when the first attempt returns a different error code than
-ENODEV, this results in an unbalance of calls to phylink_create and
phylink_destroy by the time we exit the function. The phylink instance
has leaked.
There are many other error codes that can be returned by
phylink_of_phy_connect. For example, phylink_validate returns -EINVAL.
So this is a practical issue too.
Evan Quan [Fri, 3 Sep 2021 06:33:11 +0000 (14:33 +0800)]
PCI: Add AMD GPU multi-function power dependencies
Some AMD GPUs have built-in USB xHCI and USB Type-C UCSI controllers with
power dependencies between the GPU and the other functions as in 6d2e369f0d4c ("PCI: Add NVIDIA GPU multi-function power dependencies").
Add device link support for the AMD integrated USB xHCI and USB Type-C UCSI
controllers.
Without this, runtime power management, including GPU resume and temp and
fan sensors don't work correctly.
Commit 375553a93201 ("PCI: Setup ACPI fwnode early and at the same time
with OF") added a call to pci_set_acpi_fwnode() in pci_setup_device(),
which unconditionally clears any fwnode previously set by
pci_set_of_node().
pci_set_acpi_fwnode() looks for ACPI_COMPANION(), which only returns the
existing fwnode if it was set by ACPI_COMPANION_SET(). If it was set by
OF instead, ACPI_COMPANION() returns NULL and pci_set_acpi_fwnode()
accidentally clears the fwnode. To fix this, look for any fwnode instead
of just ACPI companions.
Fixes a virtio-iommu boot regression in v5.15-rc1.
7bac54497c3e ("PCI/VPD: Determine VPD size in pci_vpd_init()") reads VPD at
enumeration-time to find the size. But this is quite slow, and we don't
need the size until we actually need data from VPD. Dave reported a boot
slowdown of more than two minutes [1].
Defer the VPD sizing until a driver or the user (via sysfs) requests
information from VPD.
If devices are quirked because VPD is known not to work, don't bother even
looking for the VPD capability. The VPD will not be accessible at all.
The qnx4 directory entries are 64-byte blocks that have different
contents depending on the a status byte that is in the last byte of the
block.
In particular, a directory entry can be either a "link info" entry with
a 48-byte name and pointers to the real inode information, or an "inode
entry" with a smaller 16-byte name and the full inode information.
But the code was written to always just treat the directory name as if
it was part of that "inode entry", and just extend the name to the
longer case if the status byte said it was a link entry.
That work just fine and gives the right results, but now that gcc is
tracking data structure accesses much more, the code can trigger a
compiler error about using up to 48 bytes (the long name) in a structure
that only has that shorter name in it:
fs/qnx4/dir.c: In function ‘qnx4_readdir’:
fs/qnx4/dir.c:51:32: error: ‘strnlen’ specified bound 48 exceeds source size 16 [-Werror=stringop-overread]
51 | size = strnlen(de->di_fname, size);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from fs/qnx4/qnx4.h:3,
from fs/qnx4/dir.c:16:
include/uapi/linux/qnx4_fs.h:45:25: note: source object declared here
45 | char di_fname[QNX4_SHORT_NAME_MAX];
| ^~~~~~~~
which is because the source code doesn't really make this whole "one of
two different types" explicit.
Fix this by introducing a very explicit union of the two types, and
basically explaining to the compiler what is really going on.
The sparc mdesc code does pointer games with 'struct mdesc_hdr', but
didn't describe to the compiler how that header is then followed by the
data that the header describes.
As a result, gcc is now unhappy since it does stricter pointer range
tracking, and doesn't understand about how these things work. This
results in various errors like:
arch/sparc/kernel/mdesc.c: In function ‘mdesc_node_by_name’:
arch/sparc/kernel/mdesc.c:647:22: error: ‘strcmp’ reading 1 or more bytes from a region of size 0 [-Werror=stringop-overread]
647 | if (!strcmp(names + ep[ret].name_offset, name))
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
which are easily avoided by just describing 'struct mdesc_hdr' better,
and making the node_block() helper function look into that unsized
data[] that follows the header.
This makes the sparc64 build happy again at least for my cross-compiler
version (gcc version 11.2.1).
Merge branch 'absolute-pointer' (patches from Guenter)
Merge absolute_pointer macro series from Guenter Roeck:
"Kernel test builds currently fail for several architectures with error
messages such as the following.
drivers/net/ethernet/i825xx/82596.c: In function 'i82596_probe':
arch/m68k/include/asm/string.h:72:25: error:
'__builtin_memcpy' reading 6 bytes from a region of size 0
[-Werror=stringop-overread]
Such warnings may be reported by gcc 11.x for string and memory
operations on fixed addresses if gcc's builtin functions are used for
those operations.
This series introduces absolute_pointer() to fix the problem.
absolute_pointer() disassociates a pointer from its originating symbol
type and context, and thus prevents gcc from making assumptions about
pointers passed to memory operations"
* emailed patches from Guenter Roeck <[email protected]>:
alpha: Use absolute_pointer to define COMMAND_LINE
alpha: Move setup.h out of uapi
net: i825xx: Use absolute_pointer for memcpy from fixed memory location
compiler.h: Introduce absolute_pointer macro
net: i825xx: Use absolute_pointer for memcpy from fixed memory location
gcc 11.x reports the following compiler warning/error.
drivers/net/ethernet/i825xx/82596.c: In function 'i82596_probe':
arch/m68k/include/asm/string.h:72:25: error:
'__builtin_memcpy' reading 6 bytes from a region of size 0 [-Werror=stringop-overread]
Use absolute_pointer() to work around the problem.
absolute_pointer() disassociates a pointer from its originating symbol
type and context. Use it to prevent compiler warnings/errors such as
drivers/net/ethernet/i825xx/82596.c: In function 'i82596_probe':
arch/m68k/include/asm/string.h:72:25: error:
'__builtin_memcpy' reading 6 bytes from a region of size 0 [-Werror=stringop-overread]
Such warnings may be reported by gcc 11.x for string and memory
operations on fixed addresses.
Li Jinlin [Tue, 14 Sep 2021 04:26:05 +0000 (12:26 +0800)]
blk-cgroup: fix UAF by grabbing blkcg lock before destroying blkg pd
KASAN reports a use-after-free report when doing fuzz test:
[693354.104835] ==================================================================
[693354.105094] BUG: KASAN: use-after-free in bfq_io_set_weight_legacy+0xd3/0x160
[693354.105336] Read of size 4 at addr ffff888be0a35664 by task sh/1453338
[693354.108511] Memory state around the buggy address:
[693354.108671] ffff888be0a35500: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[693354.116396] ffff888be0a35580: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[693354.124473] >ffff888be0a35600: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[693354.132421] ^
[693354.140284] ffff888be0a35680: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[693354.147912] ffff888be0a35700: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[693354.155281] ==================================================================
blkgs are protected by both queue and blkcg locks and holding
either should stabilize them. However, the path of destroying
blkg policy data is only protected by queue lock in
blkcg_activate_policy()/blkcg_deactivate_policy(). Other tasks
can get the blkg policy data before the blkg policy data is
destroyed, and use it after destroyed, which will result in a
use-after-free.
Once blk_throtl_init() queue init failed, blkcg_iolatency_exit() will
not be invoked for cleanup. That leads a memory leak. Swap the
blk_throtl_init() and blk_iolatency_init() calls can solve this.
tools/bootconfig: Define memblock_free_ptr() to fix build error
The lib/bootconfig.c file is shared with the 'bootconfig' tooling, and
as a result, the changes incommit 77e02cf57b6c ("memblock: introduce
saner 'memblock_free_ptr()' interface") need to also be reflected in the
tooling header file.
So define the new memblock_free_ptr() wrapper, and remove unused __pa()
and memblock_free().
locking/rwbase: Take care of ordering guarantee for fastpath reader
Readers of rwbase can lock and unlock without taking any inner lock, if
that happens, we need the ordering provided by atomic operations to
satisfy the ordering semantics of lock/unlock. Without that, considering
the follow case:
{ X = 0 initially }
CPU 0 CPU 1
===== =====
rt_write_lock();
X = 1
rt_write_unlock():
atomic_add(READER_BIAS - WRITER_BIAS, ->readers);
// ->readers is READER_BIAS.
rt_read_lock():
if ((r = atomic_read(->readers)) < 0) // True
atomic_try_cmpxchg(->readers, r, r + 1); // succeed.
<acquire the read lock via fast path>
r1 = X; // r1 may be 0, because nothing prevent the reordering
// of "X=1" and atomic_add() on CPU 1.
Therefore audit every usage of atomic operations that may happen in a
fast path, and add necessary barriers.
events: Reuse value read using READ_ONCE instead of re-reading it
In perf_event_addr_filters_apply, the task associated with
the event (event->ctx->task) is read using READ_ONCE at the beginning
of the function, checked, and then re-read from event->ctx->task,
voiding all guarantees of the checks. Reuse the value that was read by
READ_ONCE to ensure the consistency of the task struct throughout the
function.
Pavel Begunkov [Wed, 15 Sep 2021 10:00:05 +0000 (11:00 +0100)]
io_uring: move iopoll reissue into regular IO path
230d50d448acb ("io_uring: move reissue into regular IO path")
made non-IOPOLL I/O to not retry from ki_complete handler. Follow it
steps and do the same for IOPOLL. Same problems, same implementation,
same -EAGAIN assumptions.
Get rid of the need to do re-expand and revert on an iterator when we
encounter a short IO, or failure that warrants a retry. Use the new
state save/restore helpers instead.
We keep the iov_iter_state persistent across retries, if we need to
restart the read or write operation. If there's a pending retry, the
operation will always exit with the state correctly saved.
Merge tag 'nvme-5.15-2021-09-15' of git://git.infradead.org/nvme into block-5.15
Pull NVMe fixes from Christoph:
"nvme fixes for Linux 5.15
- fix ANA state updates when a namespace is not present (Anton Eidelman)
- nvmet: fix a width vs precision bug in nvmet_subsys_attr_serial_show
(Dan Carpenter)
- avoid race in shutdown namespace removal (Daniel Wagner)
- fix io_work priority inversion in nvme-tcp (Keith Busch)
- destroy cm id before destroy qp to avoid use after free (Ruozhu Li)"
* tag 'nvme-5.15-2021-09-15' of git://git.infradead.org/nvme:
nvme-tcp: fix io_work priority inversion
nvme-rdma: destroy cm id before destroy qp to avoid use after free
nvme-multipath: fix ANA state updates when a namespace is not present
nvme: avoid race in shutdown namespace removal
nvmet: fix a width vs precision bug in nvmet_subsys_attr_serial_show()
Some PHYs pointed to by "phy-handle" will never bind to a driver until a
consumer attaches to it. And when the consumer attaches to it, they get
forcefully bound to a generic PHY driver. In such cases, parsing the
phy-handle property and creating a device link will prevent the consumer
from ever probing. We don't want that. So revert support for
"phy-handle" property until we come up with a better mechanism for
binding PHYs to generic drivers before a consumer tries to attach to it.
s390 is the only architecture which allows to set the
-mwarn-dynamicstack compile option. This however will also always
generate a warning with system call stack randomization, which uses
alloca to generate some random sized stack frame.
On the other hand Linus just enabled "-Werror" by default with commit 3fe617ccafd6 ("Enable '-Werror' by default for all kernel builds"),
which means compiles will always fail by default.
So instead of playing once again whack-a-mole for something which is
s390 specific, simply remove this option.
drivers/s390/crypto/ap_bus.c:216: warning:
bad line:
drivers/s390/crypto/ap_bus.c:444:
warning: Function parameter or member 'floating' not described in 'ap_interrupt_handler'
s390/pci_mmio: fully validate the VMA before calling follow_pte()
We should not walk/touch page tables outside of VMA boundaries when
holding only the mmap sem in read mode. Evil user space can modify the
VMA layout just before this function runs and e.g., trigger races with
page table removal code since commit dd2283f2605e ("mm: mmap: zap pages
with read mmap_sem in munmap").
find_vma() does not check if the address is >= the VMA start address;
use vma_lookup() instead.
Reviewed-by: Niklas Schnelle <[email protected]> Reviewed-by: Liam R. Howlett <[email protected]> Fixes: dd2283f2605e ("mm: mmap: zap pages with read mmap_sem in munmap") Signed-off-by: David Hildenbrand <[email protected]> Signed-off-by: Vasily Gorbik <[email protected]>
powerpc/xics: Set the IRQ chip data for the ICS native backend
The ICS native driver relies on the IRQ chip data to find the struct
'ics_native' describing the ICS controller but it was removed by commit 248af248a8f4 ("powerpc/xics: Rename the map handler in a check handler").
Revert this change to fix the Microwatt SoC platform.
Jan Beulich [Tue, 7 Sep 2021 12:07:47 +0000 (14:07 +0200)]
swiotlb-xen: drop DEFAULT_NSLABS
It was introduced by 4035b43da6da ("xen-swiotlb: remove xen_set_nslabs")
and then not removed by 2d29960af0be ("swiotlb: dynamically allocate
io_tlb_default_mem").
Jan Beulich [Tue, 7 Sep 2021 12:07:21 +0000 (14:07 +0200)]
swiotlb-xen: arrange to have buffer info logged
I consider it unhelpful that address and size of the buffer aren't put
in the log file; it makes diagnosing issues needlessly harder. The
majority of callers of swiotlb_init() also passes 1 for the "verbose"
parameter.
Jan Beulich [Tue, 7 Sep 2021 12:06:55 +0000 (14:06 +0200)]
swiotlb-xen: drop leftover __ref
Commit a98f565462f0 ("xen-swiotlb: split xen_swiotlb_init") should not
only have added __init to the split off function, but also should have
dropped __ref from the one left.
Jan Beulich [Tue, 7 Sep 2021 12:06:37 +0000 (14:06 +0200)]
swiotlb-xen: limit init retries
Due to the use of max(1024, ...) there's no point retrying (and issuing
bogus log messages) when the number of slabs is already no larger than
this minimum value.
Jan Beulich [Tue, 7 Sep 2021 12:05:54 +0000 (14:05 +0200)]
swiotlb-xen: suppress certain init retries
Only on the 2nd of the paths leading to xen_swiotlb_init()'s "error"
label it is useful to retry the allocation; the first one did already
iterate through all possible order values.
Jan Beulich [Tue, 7 Sep 2021 12:05:12 +0000 (14:05 +0200)]
swiotlb-xen: maintain slab count properly
Generic swiotlb code makes sure to keep the slab count a multiple of the
number of slabs per segment. Yet even without checking whether any such
assumption is made elsewhere, it is easy to see that xen_swiotlb_fixup()
might alter unrelated memory when calling xen_create_contiguous_region()
for the last segment, when that's not a full one - the function acts on
full order-N regions, not individual pages.
Align the slab count suitably when halving it for a retry. Add a build
time check and a runtime one. Replace the no longer useful local
variable "slabs" by an "order" one calculated just once, outside of the
loop. Re-use "order" for calculating "dma_bits", and change the type of
the latter as well as the one of "i" while touching this anyway.
Jan Beulich [Tue, 7 Sep 2021 12:04:47 +0000 (14:04 +0200)]
swiotlb-xen: fix late init retry
The commit referenced below removed the assignment of "bytes" from
xen_swiotlb_init() without - like done for xen_swiotlb_init_early() -
adding an assignment on the retry path, thus leading to excessively
sized allocations upon retries.
xen: fix usage of pmd_populate in mremap for pv guests
Commit 0881ace292b662 ("mm/mremap: use pmd/pud_poplulate to update page
table entries") introduced a regression when running as Xen PV guest.
Today pmd_populate() for Xen PV assumes that the PFN inserted is
referencing a not yet used page table. In case of move_normal_pmd()
this is not true, resulting in WARN splats like:
Randy Dunlap [Mon, 13 Sep 2021 22:06:05 +0000 (15:06 -0700)]
ptp: dp83640: don't define PAGE0
Building dp83640.c on arch/parisc/ produces a build warning for
PAGE0 being redefined. Since the macro is not used in the dp83640
driver, just make it a comment for documentation purposes.
In file included from ../drivers/net/phy/dp83640.c:23:
../drivers/net/phy/dp83640_reg.h:8: warning: "PAGE0" redefined
8 | #define PAGE0 0x0000
from ../drivers/net/phy/dp83640.c:11:
../arch/parisc/include/asm/page.h:187: note: this is the location of the previous definition
187 | #define PAGE0 ((struct zeropage *)__PAGE_OFFSET)
block: flush the integrity workqueue in blk_integrity_unregister
When the integrity profile is unregistered there can still be integrity
reads queued up which could see a NULL verify_fn as shown by the race
window below:
CPU0 CPU1
process_one_work nvme_validate_ns
bio_integrity_verify_fn nvme_update_ns_info
nvme_update_disk_info
blk_integrity_unregister
---set queue->integrity as 0
bio_integrity_process
--access bi->profile->verify_fn(bi is a pointer of queue->integity)
Before calling blk_integrity_unregister in nvme_update_disk_info, we must
make sure that there is no work item in the kintegrityd_wq. Just call
blk_flush_integrity to flush the work queue so the bug can be resolved.
block: check if a profile is actually registered in blk_integrity_unregister
While clearing the profile itself is harmless, we really should not clear
the stable writes flag if it wasn't set due to a registered integrity
profile.
On sparc64, __fls() returns an "int", but the drm TTM code expected it
to be "unsigned long" as on x86. As a result, on sparc (and arc, and
m68k) you get build errors because 'min()' checks that the types match.
As suggested by Linus, it can use min_t instead of min to force the type
to be "unsigned int".
The boot-time allocation interface for memblock is a mess, with
'memblock_alloc()' returning a virtual pointer, but then you are
supposed to free it with 'memblock_free()' that takes a _physical_
address.
Not only is that all kinds of strange and illogical, but it actually
causes bugs, when people then use it like a normal allocation function,
and it fails spectacularly on a NULL pointer:
I really don't want to apply patches that treat the symptoms, when the
fundamental cause is this horribly confusing interface.
I started out looking at just automating a sane replacement sequence,
but because of this mix or virtual and physical addresses, and because
people have used the "__pa()" macro that can take either a regular
kernel pointer, or just the raw "unsigned long" address, it's all quite
messy.
So this just introduces a new saner interface for freeing a virtual
address that was allocated using 'memblock_alloc()', and that was kept
as a regular kernel pointer. And then it converts a couple of users
that are obvious and easy to test, including the 'xbc_nodes' case in
lib/bootconfig.c that caused problems.
In amdgpu_dm_atomic_check, dc_validate_global_state is called. On
failure this logs a warning to the kernel journal. However warnings
shouldn't be used for atomic test-only commit failures: user-space
might be perfoming a lot of atomic test-only commits to find the
best hardware configuration.
Downgrade the log to a regular DRM atomic message. While at it, use
the new device-aware logging infrastructure.
This fixes error messages in the kernel when running gamescope [1].
Ernst Sjöstrand [Thu, 2 Sep 2021 07:50:27 +0000 (09:50 +0200)]
drm/amd/amdgpu: Increase HWIP_MAX_INSTANCE to 10
Seems like newer cards can have even more instances now.
Found by UBSAN: array-index-out-of-bounds in
drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c:318:29
index 8 is out of range for type 'uint32_t *[8]'
ipc: remove memcg accounting for sops objects in do_semtimedop()
Linus proposes to revert an accounting for sops objects in
do_semtimedop() because it's really just a temporary buffer
for a single semtimedop() system call.
This object can consume up to 2 pages, syscall is sleeping
one, size and duration can be controlled by user, and this
allocation can be repeated by many thread at the same time.
However Shakeel Butt pointed that there are much more popular
objects with the same life time and similar memory
consumption, the accounting of which was decided to be
rejected for performance reasons.
Considering at least 2 pages for task_struct and 2 pages for
the kernel stack, a back of the envelope calculation gives a
footprint amplification of <1.5 so this temporal buffer can be
safely ignored.
The factor would IMO be interesting if it was >> 2 (from the
PoV of excessive (ab)use, fine-grained accounting seems to be
currently unfeasible due to performance impact).
io_uring: allow retry for O_NONBLOCK if async is supported
A common complaint is that using O_NONBLOCK files with io_uring can be a
bit of a pain. Be a bit nicer and allow normal retry IFF the file does
support async behavior. This makes it possible to use io_uring more
reliably with O_NONBLOCK files, for use cases where it either isn't
possible or feasible to modify the file flags.
James Morse [Tue, 14 Sep 2021 16:56:23 +0000 (16:56 +0000)]
cpufreq: schedutil: Destroy mutex before kobject_put() frees the memory
Since commit e5c6b312ce3c ("cpufreq: schedutil: Use kobject release()
method to free sugov_tunables") kobject_put() has kfree()d the
attr_set before gov_attr_set_put() returns.
kobject_put() isn't the last user of attr_set in gov_attr_set_put(),
the subsequent mutex_destroy() triggers a use-after-free:
| BUG: KASAN: use-after-free in mutex_is_locked+0x20/0x60
| Read of size 8 at addr ffff000800ca4250 by task cpuhp/2/20
|
| CPU: 2 PID: 20 Comm: cpuhp/2 Not tainted 5.15.0-rc1 #12369
| Hardware name: ARM LTD ARM Juno Development Platform/ARM Juno Development
| Platform, BIOS EDK II Jul 30 2018
| Call trace:
| dump_backtrace+0x0/0x380
| show_stack+0x1c/0x30
| dump_stack_lvl+0x8c/0xb8
| print_address_description.constprop.0+0x74/0x2b8
| kasan_report+0x1f4/0x210
| kasan_check_range+0xfc/0x1a4
| __kasan_check_read+0x38/0x60
| mutex_is_locked+0x20/0x60
| mutex_destroy+0x80/0x100
| gov_attr_set_put+0xfc/0x150
| sugov_exit+0x78/0x190
| cpufreq_offline.isra.0+0x2c0/0x660
| cpuhp_cpufreq_offline+0x14/0x24
| cpuhp_invoke_callback+0x430/0x6d0
| cpuhp_thread_fun+0x1b0/0x624
| smpboot_thread_fn+0x5e0/0xa6c
| kthread+0x3a0/0x450
| ret_from_fork+0x10/0x20
Swap the order of the calls.
Fixes: e5c6b312ce3c ("cpufreq: schedutil: Use kobject release() method to free sugov_tunables") Cc: 4.7+ <[email protected]> # 4.7+ Signed-off-by: James Morse <[email protected]> Signed-off-by: Rafael J. Wysocki <[email protected]>
This warning helps catch uninitialized variables. It should have been
enabled at the same time as commit b2423184ac33 ("drm/i915: Enable
-Wuninitialized") but I did not realize they were disabled separately.
Enable it now that i915 is clean so that it stays that way.
drm/i915/selftests: Always initialize err in igt_dmabuf_import_same_driver_lmem()
Clang warns:
drivers/gpu/drm/i915/gem/selftests/i915_gem_dmabuf.c:127:13: warning:
variable 'err' is used uninitialized whenever 'if' condition is false
[-Wsometimes-uninitialized]
} else if (PTR_ERR(import) != -EOPNOTSUPP) {
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/gpu/drm/i915/gem/selftests/i915_gem_dmabuf.c:138:9: note:
uninitialized use occurs here
return err;
^~~
drivers/gpu/drm/i915/gem/selftests/i915_gem_dmabuf.c:127:9: note: remove
the 'if' if its condition is always true
} else if (PTR_ERR(import) != -EOPNOTSUPP) {
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/gpu/drm/i915/gem/selftests/i915_gem_dmabuf.c:95:9: note:
initialize the variable 'err' to silence this warning
int err;
^
= 0
The test is expected to pass if i915_gem_prime_import() returns
-EOPNOTSUPP so initialize err to zero in this case.
drm/i915/selftests: Do not use import_obj uninitialized
Clang warns a couple of times:
drivers/gpu/drm/i915/gem/selftests/i915_gem_dmabuf.c:63:6: warning:
variable 'import_obj' is used uninitialized whenever 'if' condition is
true [-Wsometimes-uninitialized]
if (import != &obj->base) {
^~~~~~~~~~~~~~~~~~~~
drivers/gpu/drm/i915/gem/selftests/i915_gem_dmabuf.c:80:22: note:
uninitialized use occurs here
i915_gem_object_put(import_obj);
^~~~~~~~~~
drivers/gpu/drm/i915/gem/selftests/i915_gem_dmabuf.c:63:2: note: remove
the 'if' if its condition is always false
if (import != &obj->base) {
^~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/gpu/drm/i915/gem/selftests/i915_gem_dmabuf.c:38:46: note:
initialize the variable 'import_obj' to silence this warning
struct drm_i915_gem_object *obj, *import_obj;
^
= NULL
Shuffle the import_obj initialization above these if statements so that
it is not used uninitialized.
Pavel Begunkov [Tue, 14 Sep 2021 15:12:52 +0000 (16:12 +0100)]
io_uring: auto-removal for direct open/accept
It might be inconvenient that direct open/accept deviates from the
update semantics and fails if the slot is taken instead of removing a
file sitting there. Implement this auto-removal.
Note that removal might need to allocate and so may fail. However, if an
empty slot is specified, it's guaraneed to not fail on the fd
installation side for valid userspace programs. It's needed for users
who can't tolerate such failures, e.g. accept where the other end
never retries.
Michael Ellerman [Tue, 14 Sep 2021 12:17:23 +0000 (22:17 +1000)]
powerpc/boot: Fix build failure since GCC 4.9 removal
Stephen reported that the build was broken since commit 6d2ef226f2f1 ("compiler_attributes.h: drop __has_attribute() support for
gcc4"), with errors such as:
include/linux/compiler_attributes.h:296:5: warning: "__has_attribute" is not defined, evaluates to 0 [-Wundef]
296 | #if __has_attribute(__warning__)
| ^~~~~~~~~~~~~~~
make[2]: *** [arch/powerpc/boot/Makefile:225: arch/powerpc/boot/crt0.o] Error 1
But we expect __has_attribute() to always be defined now that we've
stopped using GCC 4.
Linus debugged it to the point of reading the GCC sources, and noticing
that the problem is that __has_attribute() is not defined when
preprocessing assembly files, which is what we're doing here.
Our assembly files don't include, or need, compiler_attributes.h, but
they are getting it unconditionally from the -include in BOOT_CFLAGS,
which is then added in its entirety to BOOT_AFLAGS.
That -include was added in commit 77433830ed16 ("powerpc: boot: include
compiler_attributes.h") so that we'd have "fallthrough" and other
attributes defined for the C files in arch/powerpc/boot. But it's not
needed for assembly files.
The minimal fix is to move the addition to BOOT_CFLAGS of -include
compiler_attributes.h until after we've copied BOOT_CFLAGS into
BOOT_AFLAGS. That avoids including compiler_attributes.h for asm files,
but makes no other change to BOOT_CFLAGS or BOOT_AFLAGS.
In an ideal world, when someone is passed an iov_iter and returns X bytes,
then X bytes would have been consumed/advanced from the iov_iter. But we
have use cases that always consume the entire iterator, a few examples
of that are iomap and bdev O_DIRECT. This means we cannot rely on the
state of the iov_iter once we've called ->read_iter() or ->write_iter().
This would be easier if we didn't always have to deal with truncate of
the iov_iter, as rewinding would be trivial without that. We recently
added a commit to track the truncate state, but that grew the iov_iter
by 8 bytes and wasn't the best solution.
Implement a helper to save enough of the iov_iter state to sanely restore
it after we've called the read/write iterator helpers. This currently
only works for IOVEC/BVEC/KVEC as that's all we need, support for other
iterator types are left as an exercise for the reader.
As mentioned in https://lkml.org/lkml/2021/9/13/1819
5 years old commit 919483096bfe ("ipv4: fix memory leaks in ip_cmsg_send() callers")
was a correct fix.
ip_cmsg_send() can loop over multiple cmsghdr()
If IP_RETOPTS has been successful, but following cmsghdr generates an error,
we do not free ipc.ok
If IP_RETOPTS is not successful, we have freed the allocated temporary space,
not the one currently in ipc.opt.
Sure, code could be refactored, but let's not bring back old bugs.
Fixes: d7807a9adf48 ("Revert "ipv4: fix memory leaks in ip_cmsg_send() callers"") Signed-off-by: Eric Dumazet <[email protected]> Cc: Yajun Deng <[email protected]> Signed-off-by: David S. Miller <[email protected]>
tcp: fix tp->undo_retrans accounting in tcp_sacktag_one()
Commit 10d3be569243 ("tcp-tso: do not split TSO packets at retransmit
time") may directly retrans a multiple segments TSO/GSO packet without
split, Since this commit, we can no longer assume that a retransmitted
packet is a single segment.
This patch fixes the tp->undo_retrans accounting in tcp_sacktag_one()
that use the actual segments(pcount) of the retransmitted packet.
Before that commit (10d3be569243), the assumption underlying the
tp->undo_retrans-- seems correct.
Andreas Larsson [Wed, 8 Sep 2021 07:48:22 +0000 (09:48 +0200)]
sparc32: page align size in arch_dma_alloc
Commit 53b7670e5735 ("sparc: factor the dma coherent mapping into
helper") lost the page align for the calls to dma_make_coherent and
srmmu_unmapiorange. The latter cannot handle a non page aligned len
argument.
The following pull-request contains BPF updates for your *net* tree.
We've added 7 non-merge commits during the last 13 day(s) which contain
a total of 18 files changed, 334 insertions(+), 193 deletions(-).
The main changes are:
1) Fix mmap_lock lockdep splat in BPF stack map's build_id lookup, from Yonghong Song.
2) Fix BPF cgroup v2 program bypass upon net_cls/prio activation, from Daniel Borkmann.
3) Fix kvcalloc() BTF line info splat on oversized allocation attempts, from Bixuan Cui.
4) Fix BPF selftest build of task_pt_regs test for arm64/s390, from Jean-Philippe Brucker.
5) Fix BPF's disasm.{c,h} to dual-license so that it is aligned with bpftool given the former
is a build dependency for the latter, from Daniel Borkmann with ACKs from contributors.
====================
Fixes: cc36a070b590 ("net-caif: add CAIF netdevice") Signed-off-by: Eric Dumazet <[email protected]> Signed-off-by: David S. Miller <[email protected]>
Keith Busch [Thu, 9 Sep 2021 15:54:52 +0000 (08:54 -0700)]
nvme-tcp: fix io_work priority inversion
Dispatching requests inline with the .queue_rq() call may block while
holding the send_mutex. If the tcp io_work also happens to schedule, it
may see the req_list is non-empty, leaving "pending" true and remaining
in TASK_RUNNING. Since io_work is of higher scheduling priority, the
.queue_rq task may not get a chance to run, blocking forward progress
and leading to io timeouts.
Instead of checking for pending requests within io_work, let the queueing
restart io_work outside the send_mutex lock if there is more work to be
done.
Ruozhu Li [Mon, 6 Sep 2021 03:51:34 +0000 (11:51 +0800)]
nvme-rdma: destroy cm id before destroy qp to avoid use after free
We should always destroy cm_id before destroy qp to avoid to get cma
event after qp was destroyed, which may lead to use after free.
In RDMA connection establishment error flow, don't destroy qp in cm
event handler.Just report cm_error to upper level, qp will be destroy
in nvme_rdma_alloc_queue() after destroy cm id.
Anton Eidelman [Sun, 12 Sep 2021 18:54:57 +0000 (12:54 -0600)]
nvme-multipath: fix ANA state updates when a namespace is not present
nvme_update_ana_state() has a deficiency that results in a failure to
properly update the ana state for a namespace in the following case:
NSIDs in ctrl->namespaces: 1, 3, 4
NSIDs in desc->nsids: 1, 2, 3, 4
Loop iteration 0:
ns index = 0, n = 0, ns->head->ns_id = 1, nsid = 1, MATCH.
Loop iteration 1:
ns index = 1, n = 1, ns->head->ns_id = 3, nsid = 2, NO MATCH.
Loop iteration 2:
ns index = 2, n = 2, ns->head->ns_id = 4, nsid = 4, MATCH.
Where the update to the ANA state of NSID 3 is missed. To fix this
increment n and retry the update with the same ns when ns->head->ns_id is
higher than nsid,
Tony Luck [Mon, 13 Sep 2021 21:52:39 +0000 (14:52 -0700)]
x86/mce: Avoid infinite loop for copy from user recovery
There are two cases for machine check recovery:
1) The machine check was triggered by ring3 (application) code.
This is the simpler case. The machine check handler simply queues
work to be executed on return to user. That code unmaps the page
from all users and arranges to send a SIGBUS to the task that
triggered the poison.
2) The machine check was triggered in kernel code that is covered by
an exception table entry. In this case the machine check handler
still queues a work entry to unmap the page, etc. but this will
not be called right away because the #MC handler returns to the
fix up code address in the exception table entry.
Problems occur if the kernel triggers another machine check before the
return to user processes the first queued work item.
Specifically, the work is queued using the ->mce_kill_me callback
structure in the task struct for the current thread. Attempting to queue
a second work item using this same callback results in a loop in the
linked list of work functions to call. So when the kernel does return to
user, it enters an infinite loop processing the same entry for ever.
There are some legitimate scenarios where the kernel may take a second
machine check before returning to the user.
1) Some code (e.g. futex) first tries a get_user() with page faults
disabled. If this fails, the code retries with page faults enabled
expecting that this will resolve the page fault.
2) Copy from user code retries a copy in byte-at-time mode to check
whether any additional bytes can be copied.
On the other side of the fence are some bad drivers that do not check
the return value from individual get_user() calls and may access
multiple user addresses without noticing that some/all calls have
failed.
Fix by adding a counter (current->mce_count) to keep track of repeated
machine checks before task_work() is called. First machine check saves
the address information and calls task_work_add(). Subsequent machine
checks before that task_work call back is executed check that the address
is in the same page as the first machine check (since the callback will
offline exactly one page).
Expected worst case is four machine checks before moving on (e.g. one
user access with page faults disabled, then a repeat to the same address
with page faults enabled ... repeat in copy tail bytes). Just in case
there is some code that loops forever enforce a limit of 10.
Daniel Vetter [Thu, 2 Sep 2021 14:20:48 +0000 (16:20 +0200)]
drm/i915: Release ctx->syncobj on final put, not on ctx close
gem context refcounting is another exercise in least locking design it
seems, where most things get destroyed upon context closure (which can
race with anything really). Only the actual memory allocation and the
locks survive while holding a reference.
This tripped up Jason when reimplementing the single timeline feature
in
drm/i915: Implement SINGLE_TIMELINE with a syncobj (v4)
We could fix the bug by holding ctx->mutex in execbuf and clear the
pointer (again while holding the mutex) context_close, but it's
cleaner to just make the context object actually invariant over its
_entire_ lifetime. This way any other ioctl that's potentially racing,
but holding a full reference, can still rely on ctx->syncobj being
an immutable pointer. Which without this change, is not the case.
Thomas Hellström [Tue, 31 Aug 2021 12:29:31 +0000 (14:29 +0200)]
drm/i915/gem: Fix the mman selftest
Using the I915_MMAP_TYPE_FIXED mmap type requires the TTM backend, so
for that mmap type, use __i915_gem_object_create_user() instead of
i915_gem_object_create_internal(), as we really want to tests objects
mmap-able by user-space.
This also means that the out-of-space error happens at object creation
and returns -ENXIO rather than -ENOSPC, so fix the code up to expect
that on out-of-offset-space errors.
Finally only use I915_MMAP_TYPE_FIXED for LMEM and SMEM for now if
testing on LMEM-capable devices. For stolen LMEM, we still take the
same path as for integrated, as that haven't been moved over to TTM yet,
and user-space should not be able to create out of stolen LMEM anyway.
v2:
- Check the presence of the obj->ops->mmap_offset callback rather than
hardcoding the supported mmap regions in can_mmap() (Maarten Lankhorst)
The function is only used from within GEM_BUG_ON(), which is causing
warnings with Wunneeded-internal-declaration in some builds. Since the
function is a simple wrapper around a CT function, we can just call the
CT function directly instead.
Kai-Heng Feng [Fri, 20 Aug 2021 07:52:59 +0000 (15:52 +0800)]
drm/i915/dp: Use max params for panels < eDP 1.4
Users reported that after commit 2bbd6dba84d4 ("drm/i915: Try to use
fast+narrow link on eDP again and fall back to the old max strategy on
failure"), the screen starts to have wobbly effect.
Commit a5c936add6a2 ("drm/i915/dp: Use slow and wide link training for
everything") doesn't help either, that means the affected eDP 1.2 panels
only work with max params.
So use max params for panels < eDP 1.4 as Windows does to solve the
issue.
v3:
- Do the eDP rev check in intel_edp_init_dpcd()
v2:
- Check eDP 1.4 instead of DPCD 1.1 to apply max params
Lee Shawn C [Tue, 6 Jul 2021 15:25:41 +0000 (23:25 +0800)]
drm/i915/dp: return proper DPRX link training result
After DPRX link training, intel_dp_link_train_phy() did not
return the training result properly. If link training failed,
i915 driver would not run into link train fallback function.
And no hotplug uevent would be received by user space application.
projects / linux.git / log