Christian König [Fri, 21 Feb 2020 14:10:31 +0000 (15:10 +0100)]
drm/amdgpu: stop disable the scheduler during HW fini
When we stop the HW for example for GPU reset we should not stop the
front-end scheduler. Otherwise we run into intermediate failures during
command submission.
The scheduler should only be stopped in very few cases:
1. We can't get the hardware working in ring or IB test after a GPU reset.
2. The KIQ scheduler is not used in the front-end and should be disabled during GPU reset.
3. In amdgpu_ring_fini() when the driver unloads.
Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Acked-by: Nirmoy Das <nirmoy.das@amd.com> Test-by: Dennis Li <dennis.li@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Sierra [Mon, 23 Mar 2020 19:00:43 +0000 (14:00 -0500)]
drm/amdgpu: reroute VMC and UMD to IH ring 1 for oss v5
[Why]
Due Page faults can easily overwhelm the interrupt handler.
So to make sure that we never lose valuable interrupts on the primary ring
we re-route page faults to IH ring 1.
It also facilitates the recovery page process, since it's already
running from a process context.
This is valid for Arcturus and future Navi generation GPUs.
[How]
Setting IH_CLIENT_CFG_DATA for VMC and UMD IH clients.
Signed-off-by: Alex Sierra <alex.sierra@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Sierra [Mon, 23 Mar 2020 18:53:39 +0000 (13:53 -0500)]
drm/amdgpu: call psp to program ih cntl in SR-IOV for Navi
call psp to program ih cntl in SR-IOV if supported on Navi and Arcturus.
Signed-off-by: Alex Sierra <alex.sierra@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Sierra [Mon, 23 Mar 2020 18:28:15 +0000 (13:28 -0500)]
drm/amdgpu: enable IH ring 1 and ring 2 for navi
Support added into IH to enable ring1 and ring2 for navi10_ih.
Signed-off-by: Alex Sierra <alex.sierra@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Sierra [Wed, 18 Mar 2020 23:26:19 +0000 (18:26 -0500)]
drm/amdgpu: ih doorbell size of range changed for nbio v7.4
[Why]
nbio v7.4 size of ih doorbell range is 64 bit. This requires 2 DWords per register.
[How]
Change ih doorbell size from 2 to 4. This means two Dwords per ring.
Current configuration uses two ih rings.
Signed-off-by: Alex Sierra <alex.sierra@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Sierra [Wed, 18 Mar 2020 22:57:10 +0000 (17:57 -0500)]
drm/amdgpu: infinite retries fix from UTLC1 RB SDMA
[Why]
Previously these registers were set to 0. This was causing an
infinite retry on the UTCL1 RB, preventing higher priority RB such as paging RB.
[How]
Set to one the SDMAx_UTLC1_TIMEOUT registers for all SDMAs on Vega10, Vega12,
Vega20 and Arcturus.
Signed-off-by: Alex Sierra <alex.sierra@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Evan Quan <evan.quan@amd.com> Reviewed-by: John Clements <John.Clements@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tom St Denis [Fri, 27 Mar 2020 13:30:52 +0000 (09:30 -0400)]
drm/amd/amdgpu: Include headers for PWR and SMUIO registers
Clean up the smu10, smu12, and gfx9 drivers to use headers for
registers instead of hardcoding in the C source files.
Signed-off-by: Tom St Denis <tom.stdenis@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
xinhui pan [Thu, 26 Mar 2020 00:38:29 +0000 (08:38 +0800)]
drm/amdgpu: implement more ib pools (v2)
We have three ib pools, they are normal, VM, direct pools.
Any jobs which schedule IBs without dependence on gpu scheduler should
use DIRECT pool.
Any jobs schedule direct VM update IBs should use VM pool.
Any other jobs use NORMAL pool.
v2: squash in coding style fix
Signed-off-by: xinhui pan <xinhui.pan@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Monk Liu [Tue, 10 Mar 2020 10:12:13 +0000 (18:12 +0800)]
drm/amdgpu: postpone entering fullaccess mode
if host support new handshake we only need to enter
fullaccess_mode in ip_init() part, otherwise we need
to do it before reading vbios (becuase host prepares vbios
for VF only after received REQ_GPU_INIT event under
legacy handshake)
Signed-off-by: Monk Liu <Monk.Liu@amd.com> Reviewed-by: Emily Deng <Emily.Deng@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Monk Liu [Wed, 4 Mar 2020 13:33:27 +0000 (21:33 +0800)]
drm/amdgpu: adjust sequence of ip_discovery init and timeout_setting
what:
1)move timtout setting before ip_early_init to reduce exclusive mode
cost for SRIOV
2)move ip_discovery_init() to inside of amdgpu_discovery_reg_base_init()
it is a prepare for the later upcoming patches.
why:
in later upcoming patches we would use a new mailbox event --
"req_gpu_init_data", which is a callback hooked in adev->virt.ops and
this callback send a new event "REQ_GPU_INIT_DAT" to host to notify
host to do some preparation like "IP discovery/vbios on the VF FB"
and this callback must be:
A) invoked after set_ip_block() because virt.ops is configured during
set_ip_block()
B) invoked before ip_discovery_init() becausen ip_discovery_init()
need host side prepares everything in VF FB first.
current place of ip_discovery_init() is before we can invoke callback
of adev->virt.ops, thus we must move ip_discovery_init() to a place
after the adev->virt.ops all settle done, and the perfect place is in
amdgpu_discovery_reg_base_init()
Signed-off-by: Monk Liu <Monk.Liu@amd.com> Reviewed-by: Emily Deng <Emily.Deng@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Monk Liu [Wed, 4 Mar 2020 15:46:45 +0000 (23:46 +0800)]
drm/amdgpu: use static mmio offset for NV mailbox
what:
with the new "req_init_data" handshake we need to use mailbox
before do IP discovery, so in mxgpu_nv.c file the original
SOC15_REG method won'twork because that depends on IP discovery
complete first.
how:
so the solution is to always use static MMIO offset for NV+ mailbox
registers.
HW team confirm us all MAILBOX registers will be at the same
offset for all ASICs, no IP discovery needed for those registers
Signed-off-by: Monk Liu <Monk.Liu@amd.com> Reviewed-by: Emily Deng <Emily.Deng@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tom St Denis [Wed, 25 Mar 2020 19:07:01 +0000 (15:07 -0400)]
drm/amd/amdgpu: Fix SMUIO/PWR Confusion (v2)
The PWR block was merged into the SMUIO block by revision 12 so we add
that to the smuio_12_0_0 headers.
(v2): Drop nonsensical smuio_10_0_0 header
Signed-off-by: Tom St Denis <tom.stdenis@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tom St Denis [Wed, 25 Mar 2020 17:44:54 +0000 (13:44 -0400)]
drm/amd/amdgpu: Move PWR_MISC_CNTL_STATUS to its own header
The register is part of the PWR block not the GC block. Move to
its own header.
Signed-off-by: Tom St Denis <tom.stdenis@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tom St Denis [Wed, 25 Mar 2020 13:33:29 +0000 (09:33 -0400)]
drm/amd/amdgpu: Add missing SMUIO v12 register to headers
This register is needed by umr.
Signed-off-by: Tom St Denis <tom.stdenis@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Monk Liu [Wed, 4 Mar 2020 06:02:55 +0000 (14:02 +0800)]
drm/amdgpu: cleanup all virtualization detection routine
we need to move virt detection much earlier because:
1) HW team confirms us that RCC_IOV_FUNC_IDENTIFIER will always
be at DE5 (dw) mmio offset from vega10, this way there is no
need to implement detect_hw_virt() routine in each nbio/chip file.
for VI SRIOV chip (tonga & fiji), the BIF_IOV_FUNC_IDENTIFIER is at
0x1503
2) we need to acknowledged we are SRIOV VF before we do IP discovery because
the IP discovery content will be updated by host everytime after it recieved
a new coming "REQ_GPU_INIT_DATA" request from guest (there will be patches
for this new handshake soon).
Signed-off-by: Monk Liu <Monk.Liu@amd.com> Reviewed-by: Emily Deng <Emily.Deng@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Kent Russell [Tue, 24 Mar 2020 09:29:46 +0000 (05:29 -0400)]
drm/amdgpu: Fix FRU data checking
Ensure that when we memcpy, we don't end up copying more data than
the struct supports. For now, this is 16 characters for product number
and serial number, and 32 chars for product name
Signed-off-by: Kent Russell <kent.russell@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Kent Russell [Tue, 24 Mar 2020 11:40:20 +0000 (07:40 -0400)]
drm/amdgpu: Expose TA FW version in fw_version file
Reporting the fw_version just returns 0, the actual version is kept as
ta_*_ucode_version. This is the same as the feature reported in
the amdgpu_firmware_info debugfs file.
Signed-off-by: Kent Russell <kent.russell@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Tom St Denis [Fri, 20 Mar 2020 18:21:58 +0000 (14:21 -0400)]
drm/amd/amdgpu: Add GFX9.1 PWR_MISC_CNTL_STATUS register to headers
The registers are needed for umr and not in the headers. I left them
in the gfx_v9_0.c since it includes 9.0 and 9.4 headers and including
9.1 headers would result in a lot of duplicate registers clashing.
Signed-off-by: Tom St Denis <tom.stdenis@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Kent Russell [Fri, 20 Mar 2020 13:19:01 +0000 (09:19 -0400)]
drm/amdgpu: Add documentation for unique_id
Add the amdgpu.rst tie-ins for the unique_id documentation
Signed-off-by: Kent Russell <kent.russell@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Kent Russell [Fri, 20 Mar 2020 13:17:21 +0000 (09:17 -0400)]
drm/amdgpu: Add documentation for PCIe accounting
Add the amdgpu.rst tie-ins for the pcie accounting documentation
Signed-off-by: Kent Russell <kent.russell@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Kent Russell [Thu, 19 Mar 2020 19:03:40 +0000 (15:03 -0400)]
drm/amdgpu: Add documentation for memory info
Add the amdgpu.rst tie-ins for the mem_info documentation
Signed-off-by: Kent Russell <kent.russell@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Kent Russell [Fri, 13 Mar 2020 13:21:55 +0000 (09:21 -0400)]
drm/amdgpu: Enable reading FRU chip via I2C v3
Allow for reading of information like manufacturer, product number
and serial number from the FRU chip. Report the serial number as
the new sysfs file serial_number. Note that this only works on
server cards, as consumer cards do not feature the FRU chip, which
contains this information.
v2: Add documentation to amdgpu.rst, add helper functions,
rename functions for consistency, fix bad starting offset
v3: Remove testing definitions
Signed-off-by: Kent Russell <kent.russell@amd.com> Reviewed-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
Some displays have an issue where the hdcp chips are initialized after the
display has already lit up. This means we can sometimes authentication too early
and cause authentication failures.
This happens when HDCP is enabled and the display is power cycled. Normally we
will authenticate 2 seconds after the display is lit, but some displays need a
bit more time.
[How]
Increase delay to 3 second before we start authentication.
Signed-off-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com> Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amd/display: Don't try hdcp1.4 when content_type is set to type1
[Why]
When content type property is set to 1. We should enable hdcp2.2 and if we cant
then stop. Currently the way it works in DC is that if we fail hdcp2, we will
try hdcp1 after.
[How]
Use link config to force disable hdcp1.4 when type1 is set.
Signed-off-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com> Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Evan Quan [Fri, 27 Mar 2020 07:33:00 +0000 (15:33 +0800)]
drm/amd/powerplay: move the ASIC specific nbio operation out of smu_v11_0.c
This is ASIC specific and should be placed in _ppt.c of each ASIC.
Signed-off-by: Evan Quan <evan.quan@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Evan Quan [Fri, 27 Mar 2020 07:05:09 +0000 (15:05 +0800)]
drm/amd/powerplay: drop redundant BIF doorbell interrupt operations
This is already done in soc15.c. And this is really ASIC specific
and should not be placed here.
Signed-off-by: Evan Quan <evan.quan@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Aric Cyr [Wed, 11 Mar 2020 22:10:20 +0000 (18:10 -0400)]
drm/amd/display: LFC not working on 2.0x range monitors (v2)
[Why]
Nominal pixel clock and EDID information differ in precision so although
monitor reports maximum refresh is 2x minimum, LFC was not being
enabled.
[How]
Use minimum refresh rate as nominal/2 when EDID dictates that min
refresh = max refresh/2.
v2: squash in 64 bit divide fix
Signed-off-by: Aric Cyr <aric.cyr@amd.com> Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com> Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Yongqiang Sun [Mon, 9 Mar 2020 21:13:02 +0000 (17:13 -0400)]
drm/amd/display: Not doing optimize bandwidth if flip pending.
[Why]
In some scenario like 1366x768 VSR enabled connected with a 4K monitor
and playing 4K video in clone mode, underflow will be observed due to
decrease dppclk when previouse surface scan isn't finished
[How]
In this use case, surface flip is switching between 4K and 1366x768,
1366x768 needs smaller dppclk, and when decrease the clk and previous
surface scan is for 4K and scan isn't done, underflow will happen. Not
doing optimize bandwidth in case of flip pending.
Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com> Reviewed-by: Tony Cheng <Tony.Cheng@amd.com> Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amd/display: Use double buffered DRR timing update by default
[Why]
For some monitors extreme flickering can occur while using LFC for if
we're not doing the DRR timing update for V_TOTAL_MIN / V_TOTAL_MAX at
the DP start of frame.
Hardware can default to any time in the frame which isn't the behavior
we want.
[How]
Add a new function for setting the double buffering mode for DRR timing.
Default to DP start of frame double buffering on timing generator init.
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com> Reviewed-by: Aric Cyr <Aric.Cyr@amd.com> Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Eric Bernstein [Fri, 6 Mar 2020 22:07:12 +0000 (17:07 -0500)]
drm/amd/display: Update function to get optimal number of taps
[Why]
Diagnostics scaling test failing to set required number of vertical taps
in 4:2:0 surface case
[How]
In dpp3_get_optimal_number_of_taps() need to use LB_MEMORY_CONFIG_3 for
4:2:0 surface case. In resource_build_scaling_params() make sure to also
set plane res alpha enable based on updated surface state
Signed-off-by: Eric Bernstein <eric.bernstein@amd.com> Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com> Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Kevin Wang [Wed, 25 Mar 2020 09:06:14 +0000 (17:06 +0800)]
drm/amdgpu: fix hpd bo size calculation error
the HPD bo size calculation error.
the "mem.size" can't present actual BO size all time.
Signed-off-by: Kevin Wang <kevin1.wang@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Acked-by: Christian König <Christian.Koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Dave Airlie [Tue, 31 Mar 2020 06:34:49 +0000 (16:34 +1000)]
Merge tag 'drm-msm-next-2020-03-22' of https://gitlab.freedesktop.org/drm/msm into drm-next
A bit smaller this time around.. there are still a couple uabi
additions for vulkan waiting in the wings, but I punted on them this
cycle due to running low on time. (They should be easy enough to
rebase, and if it is a problem for anyone I can push a next+uabi
branch so that tu work can proceed.)
The bigger change is refactoring dpu resource manager and moving dpu
to use atomic global state. Other than that, it is mostly cleanups
and fixes.
Dave Airlie [Mon, 30 Mar 2020 05:56:03 +0000 (15:56 +1000)]
Merge tag 'drm-intel-next-fixes-2020-03-27' of git://anongit.freedesktop.org/drm/drm-intel into drm-next
Fixes for instability on Baytrail and Haswell;
Ice Lake RPS; Sandy Bridge RC6; and few others around
GT hangchec/reset; livelock; and a null dereference.
Linus Torvalds [Sun, 29 Mar 2020 17:40:31 +0000 (10:40 -0700)]
Merge branch 'akpm' (patches from Andrew)
Merge vm fixes from Andrew Morton:
"5 fixes"
* emailed patches from Andrew Morton <akpm@linux-foundation.org>:
mm/sparse: fix kernel crash with pfn_section_valid check
mm: fork: fix kernel_stack memcg stats for various stack implementations
hugetlb_cgroup: fix illegal access to memory
drivers/base/memory.c: indicate all memory blocks as removable
mm/swapfile.c: move inode_lock out of claim_swapfile
Linus Torvalds [Sun, 29 Mar 2020 17:36:29 +0000 (10:36 -0700)]
Merge tag 'timers-urgent-2020-03-29' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull timer fix from Thomas Gleixner:
"A single fix for the Hyper-V clocksource driver to make sched clock
actually return nanoseconds and not the virtual clock value which
increments at 10e7 HZ (100ns)"
* tag 'timers-urgent-2020-03-29' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
clocksource/drivers/hyper-v: Make sched clock return nanoseconds correctly
mm/sparse: fix kernel crash with pfn_section_valid check
Fix the crash like this:
BUG: Kernel NULL pointer dereference on read at 0x00000000
Faulting instruction address: 0xc000000000c3447c
Oops: Kernel access of bad area, sig: 11 [#1]
LE PAGE_SIZE=64K MMU=Hash SMP NR_CPUS=2048 NUMA pSeries
CPU: 11 PID: 7519 Comm: lt-ndctl Not tainted 5.6.0-rc7-autotest #1
...
NIP [c000000000c3447c] vmemmap_populated+0x98/0xc0
LR [c000000000088354] vmemmap_free+0x144/0x320
Call Trace:
section_deactivate+0x220/0x240
__remove_pages+0x118/0x170
arch_remove_memory+0x3c/0x150
memunmap_pages+0x1cc/0x2f0
devm_action_release+0x30/0x50
release_nodes+0x2f8/0x3e0
device_release_driver_internal+0x168/0x270
unbind_store+0x130/0x170
drv_attr_store+0x44/0x60
sysfs_kf_write+0x68/0x80
kernfs_fop_write+0x100/0x290
__vfs_write+0x3c/0x70
vfs_write+0xcc/0x240
ksys_write+0x7c/0x140
system_call+0x5c/0x68
The crash is due to NULL dereference at
test_bit(idx, ms->usage->subsection_map);
due to ms->usage = NULL in pfn_section_valid()
With commit d41e2f3bd546 ("mm/hotplug: fix hot remove failure in
SPARSEMEM|!VMEMMAP case") section_mem_map is set to NULL after
depopulate_section_mem(). This was done so that pfn_page() can work
correctly with kernel config that disables SPARSEMEM_VMEMMAP. With that
config pfn_to_page does
Now with SPASEMEM_VMEMAP enabled, mem_section->usage->subsection_map is
used to check the pfn validity (pfn_valid()). Since section_deactivate
release mem_section->usage if a section is fully deactivated,
pfn_valid() check after a subsection_deactivate cause a kernel crash.
static inline int pfn_valid(unsigned long pfn)
{
...
return early_section(ms) || pfn_section_valid(ms, pfn);
}
where
static inline int pfn_section_valid(struct mem_section *ms, unsigned long pfn)
{
int idx = subsection_map_index(pfn);
Avoid this by clearing SECTION_HAS_MEM_MAP when mem_section->usage is
freed. For architectures like ppc64 where large pages are used for
vmmemap mapping (16MB), a specific vmemmap mapping can cover multiple
sections. Hence before a vmemmap mapping page can be freed, the kernel
needs to make sure there are no valid sections within that mapping.
Clearing the section valid bit before depopulate_section_memap enables
this.
[aneesh.kumar@linux.ibm.com: add comment] Link: http://lkml.kernel.org/r/20200326133235.343616-1-aneesh.kumar@linux.ibm.comLink: Fixes: d41e2f3bd546 ("mm/hotplug: fix hot remove failure in SPARSEMEM|!VMEMMAP case") Reported-by: Sachin Sant <sachinp@linux.vnet.ibm.com> Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Tested-by: Sachin Sant <sachinp@linux.vnet.ibm.com> Reviewed-by: Baoquan He <bhe@redhat.com> Reviewed-by: Wei Yang <richard.weiyang@gmail.com> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Pankaj Gupta <pankaj.gupta.linux@gmail.com> Cc: Michael Ellerman <mpe@ellerman.id.au> Cc: Dan Williams <dan.j.williams@intel.com> Cc: David Hildenbrand <david@redhat.com> Cc: Oscar Salvador <osalvador@suse.de> Cc: Mike Rapoport <rppt@linux.ibm.com> Cc: <stable@vger.kernel.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Roman Gushchin [Sun, 29 Mar 2020 02:17:25 +0000 (19:17 -0700)]
mm: fork: fix kernel_stack memcg stats for various stack implementations
Depending on CONFIG_VMAP_STACK and the THREAD_SIZE / PAGE_SIZE ratio the
space for task stacks can be allocated using __vmalloc_node_range(),
alloc_pages_node() and kmem_cache_alloc_node().
In the first and the second cases page->mem_cgroup pointer is set, but
in the third it's not: memcg membership of a slab page should be
determined using the memcg_from_slab_page() function, which looks at
page->slab_cache->memcg_params.memcg . In this case, using
mod_memcg_page_state() (as in account_kernel_stack()) is incorrect:
page->mem_cgroup pointer is NULL even for pages charged to a non-root
memory cgroup.
It can lead to kernel_stack per-memcg counters permanently showing 0 on
some architectures (depending on the configuration).
In order to fix it, let's introduce a mod_memcg_obj_state() helper,
which takes a pointer to a kernel object as a first argument, uses
mem_cgroup_from_obj() to get a RCU-protected memcg pointer and calls
mod_memcg_state(). It allows to handle all possible configurations
(CONFIG_VMAP_STACK and various THREAD_SIZE/PAGE_SIZE values) without
spilling any memcg/kmem specifics into fork.c .
Note: This is a special version of the patch created for stable
backports. It contains code from the following two patches:
- mm: memcg/slab: introduce mem_cgroup_from_obj()
- mm: fork: fix kernel_stack memcg stats for various stack implementations
[guro@fb.com: introduce mem_cgroup_from_obj()] Link: http://lkml.kernel.org/r/20200324004221.GA36662@carbon.dhcp.thefacebook.com Fixes: 4d96ba353075 ("mm: memcg/slab: stop setting page->mem_cgroup pointer for slab pages") Signed-off-by: Roman Gushchin <guro@fb.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Bharata B Rao <bharata@linux.ibm.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: <stable@vger.kernel.org> Link: http://lkml.kernel.org/r/20200303233550.251375-1-guro@fb.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Mina Almasry [Sun, 29 Mar 2020 02:17:22 +0000 (19:17 -0700)]
hugetlb_cgroup: fix illegal access to memory
This appears to be a mistake in commit faced7e0806cf ("mm: hugetlb
controller for cgroups v2").
Essentially that commit does a hugetlb_cgroup_from_counter assuming that
page_counter_try_charge has initialized counter.
But if that has failed then it seems will not initialize counter, so
hugetlb_cgroup_from_counter(counter) ends up pointing to random memory,
causing kasan to complain.
The solution is to simply use 'h_cg', instead of
hugetlb_cgroup_from_counter(counter), since that is a reference to the
hugetlb_cgroup anyway. After this change kasan ceases to complain.
Fixes: faced7e0806cf ("mm: hugetlb controller for cgroups v2") Reported-by: syzbot+cac0c4e204952cf449b1@syzkaller.appspotmail.com Signed-off-by: Mina Almasry <almasrymina@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Giuseppe Scrivano <gscrivan@redhat.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: Mike Kravetz <mike.kravetz@oracle.com> Cc: David Rientjes <rientjes@google.com> Link: http://lkml.kernel.org/r/20200313223920.124230-1-almasrymina@google.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
drivers/base/memory.c: indicate all memory blocks as removable
We see multiple issues with the implementation/interface to compute
whether a memory block can be offlined (exposed via
/sys/devices/system/memory/memoryX/removable) and would like to simplify
it (remove the implementation).
1. It runs basically lockless. While this might be good for performance,
we see possible races with memory offlining that will require at
least some sort of locking to fix.
2. Nowadays, more false positives are possible. No arch-specific checks
are performed that validate if memory offlining will not be denied
right away (and such check will require locking). For example, arm64
won't allow to offline any memory block that was added during boot -
which will imply a very high error rate. Other archs have other
constraints.
3. The interface is inherently racy. E.g., if a memory block is detected
to be removable (and was not a false positive at that time), there is
still no guarantee that offlining will actually succeed. So any
caller already has to deal with false positives.
4. It is unclear which performance benefit this interface actually
provides. The introducing commit 5c755e9fd813 ("memory-hotplug: add
sysfs removable attribute for hotplug memory remove") mentioned
"A user-level agent must be able to identify which sections
of memory are likely to be removable before attempting the
potentially expensive operation."
However, no actual performance comparison was included.
Known users:
- lsmem: Will group memory blocks based on the "removable" property. [1]
- chmem: Indirect user. It has a RANGE mode where one can specify
removable ranges identified via lsmem to be offlined. However,
it also has a "SIZE" mode, which allows a sysadmin to skip the
manual "identify removable blocks" step. [2]
- powerpc-utils: Uses the "removable" attribute to skip some memory
blocks right away when trying to find some to offline+remove.
However, with ballooning enabled, it already skips this
information completely (because it once resulted in many false
negatives). Therefore, the implementation can deal with false
positives properly already. [3]
According to Nathan Fontenot, DLPAR on powerpc is nowadays no longer
driven from userspace via the drmgr command (powerpc-utils). Nowadays
it's managed in the kernel - including onlining/offlining of memory
blocks - triggered by drmgr writing to /sys/kernel/dlpar. So the
affected legacy userspace handling is only active on old kernels. Only
very old versions of drmgr on a new kernel (unlikely) might execute
slower - totally acceptable.
With CONFIG_MEMORY_HOTREMOVE, always indicating "removable" should not
break any user space tool. We implement a very bad heuristic now.
Without CONFIG_MEMORY_HOTREMOVE we cannot offline anything, so report
"not removable" as before.
Original discussion can be found in [4] ("[PATCH RFC v1] mm:
is_mem_section_removable() overhaul").
Other users of is_mem_section_removable() will be removed next, so that
we can remove is_mem_section_removable() completely.
Naohiro Aota [Sun, 29 Mar 2020 02:17:15 +0000 (19:17 -0700)]
mm/swapfile.c: move inode_lock out of claim_swapfile
claim_swapfile() currently keeps the inode locked when it is successful,
or the file is already swapfile (with -EBUSY). And, on the other error
cases, it does not lock the inode.
This inconsistency of the lock state and return value is quite confusing
and actually causing a bad unlock balance as below in the "bad_swap"
section of __do_sys_swapon().
This commit fixes this issue by moving the inode_lock() and IS_SWAPFILE
check out of claim_swapfile(). The inode is unlocked in
"bad_swap_unlock_inode" section, so that the inode is ensured to be
unlocked at "bad_swap". Thus, error handling codes after the locking now
jumps to "bad_swap_unlock_inode" instead of "bad_swap".
=====================================
WARNING: bad unlock balance detected!
5.5.0-rc7+ #176 Not tainted
-------------------------------------
swapon/4294 is trying to release lock (&sb->s_type->i_mutex_key) at: __do_sys_swapon+0x94b/0x3550
but there are no more locks to release!
other info that might help us debug this:
no locks held by swapon/4294.
1) Fix memory leak in vti6, from Torsten Hilbrich.
2) Fix double free in xfrm_policy_timer, from YueHaibing.
3) NL80211_ATTR_CHANNEL_WIDTH attribute is put with wrong type, from
Johannes Berg.
4) Wrong allocation failure check in qlcnic driver, from Xu Wang.
5) Get ks8851-ml IO operations right, for real this time, from Marek
Vasut.
* git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (22 commits)
r8169: fix PHY driver check on platforms w/o module softdeps
net: ks8851-ml: Fix IO operations, again
mlxsw: spectrum_mr: Fix list iteration in error path
qlcnic: Fix bad kzalloc null test
mac80211: set IEEE80211_TX_CTRL_PORT_CTRL_PROTO for nl80211 TX
mac80211: mark station unauthorized before key removal
mac80211: Check port authorization in the ieee80211_tx_dequeue() case
cfg80211: Do not warn on same channel at the end of CSA
mac80211: drop data frames without key on encrypted links
ieee80211: fix HE SPR size calculation
nl80211: fix NL80211_ATTR_CHANNEL_WIDTH attribute type
xfrm: policy: Fix doulbe free in xfrm_policy_timer
bpf: Explicitly memset some bpf info structures declared on the stack
bpf: Explicitly memset the bpf_attr structure
bpf: Sanitize the bpf_struct_ops tcp-cc name
vti6: Fix memory leak of skb if input policy check fails
esp: remove the skb from the chain when it's enqueued in cryptd_wq
ipv6: xfrm6_tunnel.c: Use built-in RCU list checking
xfrm: add the missing verify_sec_ctx_len check in xfrm_add_acquire
xfrm: fix uctx len check in verify_sec_ctx_len
...
Linus Torvalds [Sat, 28 Mar 2020 16:14:16 +0000 (09:14 -0700)]
Merge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Pull SCSI fixes from James Bottomley:
"Two small fixes: one in drivers (qla2xxx), and one in the core (sd) to
try to cope with USB enclosures that silently change reported
parameters"
* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
scsi: sd: Fix optimal I/O size for devices that change reported values
scsi: qla2xxx: Fix I/Os being passed down when FC device is being deleted
The following pull-request contains BPF updates for your *net* tree.
We've added 3 non-merge commits during the last 4 day(s) which contain
a total of 4 files changed, 25 insertions(+), 20 deletions(-).
The main changes are:
1) Explicitly memset the bpf_attr structure on bpf() syscall to avoid
having to rely on compiler to do so. Issues have been noticed on
some compilers with padding and other oddities where the request was
then unexpectedly rejected, from Greg Kroah-Hartman.
2) Sanitize the bpf_struct_ops TCP congestion control name in order to
avoid problematic characters such as whitespaces, from Martin KaFai Lau.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Heiner Kallweit [Fri, 27 Mar 2020 16:33:32 +0000 (17:33 +0100)]
r8169: fix PHY driver check on platforms w/o module softdeps
On Android/x86 the module loading infrastructure can't deal with
softdeps. Therefore the check for presence of the Realtek PHY driver
module fails. mdiobus_register() will try to load the PHY driver
module, therefore move the check to after this call and explicitly
check that a dedicated PHY driver is bound to the PHY device.
Fixes: f32593773549 ("r8169: check that Realtek PHY driver module is loaded") Reported-by: Chih-Wei Huang <cwhuang@android-x86.org> Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
1) Handle NETDEV_UNREGISTER for xfrm device to handle asynchronous
unregister events cleanly. From Raed Salem.
2) Fix vti6 tunnel inter address family TX through bpf_redirect().
From Nicolas Dichtel.
3) Fix lenght check in verify_sec_ctx_len() to avoid a
slab-out-of-bounds. From Xin Long.
4) Add a missing verify_sec_ctx_len check in xfrm_add_acquire
to avoid a possible out-of-bounds to access. From Xin Long.
5) Use built-in RCU list checking of hlist_for_each_entry_rcu
to silence false lockdep warning in __xfrm6_tunnel_spi_lookup
when CONFIG_PROVE_RCU_LIST is enabled. From Madhuparna Bhowmik.
6) Fix a panic on esp offload when crypto is done asynchronously.
From Xin Long.
7) Fix a skb memory leak in an error path of vti6_rcv.
From Torsten Hilbrich.
8) Fix a race that can lead to a doulbe free in xfrm_policy_timer.
From Xin Long.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Linus Torvalds [Fri, 27 Mar 2020 20:52:32 +0000 (13:52 -0700)]
Merge tag 'arm-soc-fixes-5.6' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc
Pull ARM DT and driver fixes from Arnd Bergmann:
"For the devicetree files, there are a total of 20 patches, almost
entirely for 32-bit machines:
- The Allwinner/sun9i r40 SoC dtsi file contains a number of issues,
both for correctness and for style that are addressed in separate
patches. This causes most of the changed lines of the DT updates
this time.
- More Allwinner updates fixing the identification of the security
system on sun8i/A33, a recent regression of the A83t ethernet, and
a few board specific issues on the TBS-A711 macine.
- Several bug fixes for OMAP dts files, most notably fixing the
timings for the NAND flash on the Nokia N900 that regressed a while
ago after the move to configuring them from DT. Some other OMAPs
now set the correct dma limits on the L3 bus, and a regression fix
addresses lost Ethernet on dm814x
- One incorrect setting in the newly added Raspberry Pi Zero W that
may cause issues with the SD card controller.
- A missing property on the bcm2835 firmware node caused incorrect
DMA settings.
- An old bug on the oxnas platform causing spurious interrupts is
finally addressed.
- A regression on the Exynos Midas board broke the OLED panel power
supply.
- The i.MX6 phycore SoM specified the wrong voltage for the SoC, this
is now set to the values from the datasheet.
- Some 64-bit machines use a deprecated string to identify the PSCI
firmware.
There are also several small code fixes addressing mostly serious
issues:
- Fix the sunxi rsb bus access to no longer return incorrect data
when mixing 8 and 16 bit I/O.
- Fix a suspend/resume regression on the OMAP2+ lcdc from a missing
quirk in the ti-sysc driver
- Fix a NULL pointer access from a race in the fsl dpio driver
- Fix a v5.5 regression in the exynos-chipid driver that caused an
invalid error code probing the device on non-exynos platforms
- Fix an out-of-bounds access in the AMD TEE driver"
* tag 'arm-soc-fixes-5.6' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc: (24 commits)
soc: samsung: chipid: Fix return value on non-Exynos platforms
arm64: dts: Fix leftover entry-methods for PSCI
ARM: dts: exynos: Fix regulator node aliasing on Midas-based boards
ARM: dts: oxnas: Fix clear-mask property
ARM: dts: bcm283x: Fix vc4's firmware bus DMA limitations
ARM: dts: omap5: Add bus_dma_limit for L3 bus
ARM: dts: omap4-droid4: Fix lost touchscreen interrupts
ARM: dts: dra7: Add bus_dma_limit for L3 bus
ARM: bcm2835-rpi-zero-w: Add missing pinctrl name
ARM: dts: sun8i: a33: add the new SS compatible
dt-bindings: crypto: add new compatible for A33 SS
ARM: dts: sun8i: r40: Move SPI device nodes based on address order
ARM: dts: sun8i: r40: Fix register base address for SPI2 and SPI3
ARM: dts: sun8i: r40: Move AHCI device node based on address order
ARM: dts: imx6: phycore-som: fix arm and soc minimum voltage
soc: fsl: dpio: register dpio irq handlers after dpio create
tee: amdtee: out of bounds read in find_session()
ARM: dts: N900: fix onenand timings
bus: ti-sysc: Fix quirk flags for lcdc on am335x
ARM: dts: Fix dm814x Ethernet by changing to use rgmii-id mode
...
Linus Torvalds [Fri, 27 Mar 2020 18:06:10 +0000 (11:06 -0700)]
Merge tag 'riscv-for-linus-5.6' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux
Pull RISC-V fixes from Palmer Dabbelt:
"Sorry for the last minute patches, but a few things fell through the
cracks recently. I was on the fence about sending a late pull request
just for the M-mode fixes, as we don't really have any users, but the
last patch fixes the build for Fedora which I consider pretty
important.
Given that the M-mode fixes should be very low risk, I figured it's
worth sending them along as well.
Thhis passes my standard 'boot in QEMU' test"
* tag 'riscv-for-linus-5.6' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux:
RISC-V: Move all address space definition macros to one place
RISC-V: Only select essential drivers for SOC_VIRT config
riscv: fix the IPI missing issue in nommu mode
riscv: uaccess should be used in nommu mode
Linus Torvalds [Fri, 27 Mar 2020 16:33:48 +0000 (09:33 -0700)]
Merge tag 'clk-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux
Pull clk fixes from Stephen Boyd:
"A handful of clk driver fixes.
Mostly they're around the i.MX drivers fixing the parents of a few
clks and making KASAN happy with how the message passing code works.
Besides that we have a TI driver fix for the RTC parent and a fix for
the basic gate type registration functions introduced this release
where they didn't actually pass the arguments in the right places to
the multiplexer function down below"
* tag 'clk-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux:
clk: imx: Align imx sc clock parent msg structs to 4
clk: imx: Align imx sc clock msg structs to 4
clk: Pass correct arguments to __clk_hw_register_gate()
clk: ti: am43xx: Fix clock parent for RTC clock
clk: imx8mp: Correct the enet_qos parent clock
clk: imx8mp: Correct IMX8MP_CLK_HDMI_AXI clock parent
Linus Torvalds [Fri, 27 Mar 2020 16:21:52 +0000 (09:21 -0700)]
Merge tag 'drm-fixes-2020-03-27' of git://anongit.freedesktop.org/drm/drm
Pull drm fixes from Dave Airlie:
"Pretty quiet: some minor sg mapping fixes for 3 drivers, and a single
oops fix for the scheduler. I'm hoping nobody tries to send me a fixes
pull today but I'll keep an eye out of the weekend.
radeon/amdgpu/dma-buf:
- sg list fixes
scheduler:
- oops fix"
* tag 'drm-fixes-2020-03-27' of git://anongit.freedesktop.org/drm/drm:
drm/scheduler: fix rare NULL ptr race
drm/radeon: fix scatter-gather mapping with user pages
drm/amdgpu: fix scatter-gather mapping with user pages
drm/prime: use dma length macro when mapping sg
Dirk Mueller [Tue, 14 Jan 2020 17:53:41 +0000 (18:53 +0100)]
scripts/dtc: Remove redundant YYLOC global declaration
gcc 10 will default to -fno-common, which causes this error at link
time:
(.text+0x0): multiple definition of `yylloc'; dtc-lexer.lex.o (symbol from plugin):(.text+0x0): first defined here
This is because both dtc-lexer as well as dtc-parser define the same
global symbol yyloc. Before with -fcommon those were merged into one
defintion. The proper solution would be to to mark this as "extern",
however that leads to:
dtc-lexer.l:26:16: error: redundant redeclaration of 'yylloc' [-Werror=redundant-decls]
26 | extern YYLTYPE yylloc;
| ^~~~~~
In file included from dtc-lexer.l:24:
dtc-parser.tab.h:127:16: note: previous declaration of 'yylloc' was here
127 | extern YYLTYPE yylloc;
| ^~~~~~
cc1: all warnings being treated as errors
which means the declaration is completely redundant and can just be
dropped.
Signed-off-by: Dirk Mueller <dmueller@suse.com> Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
[robh: cherry-pick from upstream] Cc: stable@vger.kernel.org Signed-off-by: Rob Herring <robh@kernel.org>