[5.10,167/221] drm/msm: Fix suspend/resume on i.MX5

Message ID	20210329075634.718810265@linuxfoundation.org
State	New
Headers	show Return-Path: <stable-owner@kernel.org> From: Greg Kroah-Hartman <gregkh@linuxfoundation.org> To: linux-kernel@vger.kernel.org Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>, stable@vger.kernel.org, Fabio Estevam <festevam@gmail.com>, Rob Clark <robdclark@chromium.org>, Sasha Levin <sashal@kernel.org> Subject: [PATCH 5.10 167/221] drm/msm: Fix suspend/resume on i.MX5 Date: Mon, 29 Mar 2021 09:58:18 +0200 Message-Id: <20210329075634.718810265@linuxfoundation.org> In-Reply-To: <20210329075629.172032742@linuxfoundation.org> References: <20210329075629.172032742@linuxfoundation.org> User-Agent: quilt/0.66 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Precedence: bulk
Series	None \| expand [5.10,002/221] mm/memcg: set memcg when splitting page [5.10,003/221] mt76: fix tx skb error handling in mt76_dma_tx_queue_skb [5.10,004/221] net: stmmac: fix dma physical address of descriptor when display ring [5.10,005/221] net: fec: ptp: avoid register access when ipg clock is disabled [5.10,006/221] powerpc/4xx: Fix build errors from mfdcr() [5.10,007/221] atm: eni: dont release is never initialized [5.10,008/221] atm: lanai: dont run lanai_dev_close if not open [5.10,009/221] Revert "r8152: adjust the settings about MAC clock speed down for RTL8153" [5.10,010/221] ALSA: hda: ignore invalid NHLT table [5.10,011/221] ixgbe: Fix memleak in ixgbe_configure_clsu32 [5.10,012/221] scsi: ufs: ufs-qcom: Disable interrupt in reset path [5.10,013/221] blk-cgroup: Fix the recursive blkg rwstat [5.10,014/221] net: tehuti: fix error return code in bdx_probe() [5.10,015/221] net: intel: iavf: fix error return code of iavf_init_get_resources() [5.10,016/221] sun/niu: fix wrong RXMAC_BC_FRM_CNT_COUNT count [5.10,017/221] gianfar: fix jumbo packets+napi+rx overrun crash [5.10,018/221] cifs: ask for more credit on async read/write code paths [5.10,019/221] gfs2: fix use-after-free in trans_drain [5.10,020/221] cpufreq: blacklist Arm Vexpress platforms in cpufreq-dt-platdev [5.10,021/221] gpiolib: acpi: Add missing IRQF_ONESHOT [5.10,022/221] nfs: fix PNFS_FLEXFILE_LAYOUT Kconfig default [5.10,023/221] NFS: Correct size calculation for create reply length [5.10,024/221] net: hisilicon: hns: fix error return code of hns_nic_clear_all_rx_fetch() [5.10,025/221] net: wan: fix error return code of uhdlc_init() [5.10,026/221] net: davicom: Use platform_get_irq_optional() [5.10,027/221] net: enetc: set MAC RX FIFO to recommended value [5.10,028/221] atm: uPD98402: fix incorrect allocation [5.10,029/221] atm: idt77252: fix null-ptr-dereference [5.10,030/221] cifs: change noisy error message to FYI [5.10,031/221] irqchip/ingenic: Add support for the JZ4760 [5.10,032/221] kbuild: add image_name to no-sync-config-targets [5.10,033/221] kbuild: dummy-tools: fix inverted tests for gcc [5.10,034/221] umem: fix error return code in mm_pci_probe() [5.10,035/221] sparc64: Fix opcode filtering in handling of no fault loads [5.10,036/221] habanalabs: Call put_pid() when releasing control device [5.10,037/221] staging: rtl8192e: fix kconfig dependency on CRYPTO [5.10,038/221] u64_stats,lockdep: Fix u64_stats_init() vs lockdep [5.10,039/221] kselftest: arm64: Fix exit code of sve-ptrace [5.10,040/221] regulator: qcom-rpmh: Correct the pmic5_hfsmps515 buck [5.10,041/221] block: Fix REQ_OP_ZONE_RESET_ALL handling [5.10,042/221] drm/amd/display: Revert dram_clock_change_latency for DCN2.1 [5.10,043/221] drm/amdgpu: fb BO should be ttm_bo_type_device [5.10,044/221] drm/radeon: fix AGP dependency [5.10,045/221] nvme: simplify error logic in nvme_validate_ns() [5.10,046/221] nvme: add NVME_REQ_CANCELLED flag in nvme_cancel_request() [5.10,047/221] nvme-fc: set NVME_REQ_CANCELLED in nvme_fc_terminate_exchange() [5.10,048/221] nvme-fc: return NVME_SC_HOST_ABORTED_CMD when a command has been aborted [5.10,049/221] nvme-core: check ctrl css before setting up zns [5.10,050/221] nvme-rdma: Fix a use after free in nvmet_rdma_write_data_done [5.10,051/221] nvme-pci: add the DISABLE_WRITE_ZEROES quirk for a Samsung PM1725a [5.10,052/221] nfs: we dont support removing system.nfs4_acl [5.10,053/221] block: Suppress uevent for hidden device when removed [5.10,054/221] mm/fork: clear PASID for new mm [5.10,055/221] ia64: fix ia64_syscall_get_set_arguments() for break-based syscalls [5.10,056/221] ia64: fix ptrace(PTRACE_SYSCALL_INFO_EXIT) sign [5.10,057/221] static_call: Pull some static_call declarations to the type headers [5.10,058/221] static_call: Allow module use without exposing static_call_key [5.10,059/221] static_call: Fix the module key fixup [5.10,060/221] static_call: Fix static_call_set_init() [5.10,061/221] KVM: x86: Protect userspace MSR filter with SRCU, and set atomically-ish [5.10,062/221] btrfs: fix sleep while in non-sleep context during qgroup removal [5.10,063/221] selinux: dont log MAC_POLICY_LOAD record on failed policy load [5.10,064/221] selinux: fix variable scope issue in live sidtab conversion [5.10,065/221] netsec: restore phy power state after controller reset [5.10,066/221] platform/x86: intel-vbtn: Stop reporting SW_DOCK events [5.10,067/221] psample: Fix user API breakage [5.10,068/221] z3fold: prevent reclaim/free race for headless pages [5.10,069/221] squashfs: fix inode lookup sanity checks [5.10,070/221] squashfs: fix xattr id and id lookup sanity checks [5.10,071/221] hugetlb_cgroup: fix imbalanced css_get and css_put pair for shared mappings [5.10,072/221] kasan: fix per-page tags for non-page_alloc pages [5.10,073/221] gcov: fix clang-11+ support [5.10,074/221] ACPI: video: Add missing callback back for Sony VPCEH3U1E [5.10,075/221] ACPICA: Always create namespace nodes using acpi_ns_create_node() [5.10,076/221] arm64: stacktrace: dont trace arch_stack_walk() [5.10,077/221] arm64: dts: ls1046a: mark crypto engine dma coherent [5.10,078/221] arm64: dts: ls1012a: mark crypto engine dma coherent [5.10,079/221] arm64: dts: ls1043a: mark crypto engine dma coherent [5.10,080/221] ARM: dts: at91: sam9x60: fix mux-mask for PA7 so it can be set to A, B and C [5.10,081/221] ARM: dts: at91: sam9x60: fix mux-mask to match products datasheet [5.10,082/221] ARM: dts: at91-sama5d27_som1: fix phy address to 7 [5.10,083/221] integrity: double check iint_cache was initialized [5.10,084/221] drm/etnaviv: Use FOLL_FORCE for userptr [5.10,085/221] drm/amd/pm: workaround for audio noise issue [5.10,086/221] drm/amdgpu/display: restore AUX_DPHY_TX_CONTROL for DCN2.x [5.10,087/221] drm/amdgpu: Add additional Sienna Cichlid PCI ID [5.10,088/221] drm/i915: Fix the GT fence revocation runtime PM logic [5.10,089/221] dm verity: fix DM_VERITY_OPTS_MAX value [5.10,090/221] dm ioctl: fix out of bounds array access when no devices [5.10,091/221] bus: omap_l3_noc: mark l3 irqs as IRQF_NO_THREAD [5.10,092/221] ARM: OMAP2+: Fix smartreflex init regression after dropping legacy data [5.10,093/221] soc: ti: omap-prm: Fix occasional abort on reset deassert for dra7 iva [5.10,094/221] veth: Store queue_mapping independently of XDP prog presence [5.10,095/221] bpf: Change inode_storages lookup_elem return value from NULL to -EBADF [5.10,096/221] libbpf: Fix INSTALL flag order [5.10,097/221] net/mlx5e: RX, Mind the MPWQE gaps when calculating offsets [5.10,098/221] net/mlx5e: When changing XDP program without reset, take refs for XSK RQs [5.10,099/221] net/mlx5e: Dont match on Geneve options in case option masks are all zero [5.10,100/221] ipv6: fix suspecious RCU usage warning [5.10,101/221] drop_monitor: Perform cleanup upon probe registration failure [5.10,102/221] macvlan: macvlan_count_rx() needs to be aware of preemption [5.10,103/221] net: sched: validate stab values [5.10,104/221] net: dsa: bcm_sf2: Qualify phydev->dev_flags based on port [5.10,105/221] igc: reinit_locked() should be called with rtnl_lock [5.10,106/221] igc: Fix Pause Frame Advertising [5.10,107/221] igc: Fix Supported Pause Frame Link Setting [5.10,108/221] igc: Fix igc_ptp_rx_pktstamp() [5.10,109/221] e1000e: add rtnl_lock() to e1000_reset_task [5.10,110/221] e1000e: Fix error handling in e1000_set_d0_lplu_state_82571 [5.10,111/221] net/qlcnic: Fix a use after free in qlcnic_83xx_get_minidump_template [5.10,112/221] net: phy: broadcom: Add power down exit reset state delay [5.10,113/221] ftgmac100: Restart MAC HW once [5.10,114/221] clk: qcom: gcc-sc7180: Use floor ops for the correct sdcc1 clk [5.10,115/221] net: ipa: terminate message handler arrays [5.10,116/221] net: qrtr: fix a kernel-infoleak in qrtr_recvmsg() [5.10,117/221] flow_dissector: fix byteorder of dissected ICMP ID [5.10,118/221] selftests/bpf: Set gopt opt_class to 0 if get tunnel opt failed [5.10,119/221] netfilter: ctnetlink: fix dump of the expect mask attribute [5.10,120/221] net: hdlc_x25: Prevent racing between "x25_close" and "x25_xmit"/"x25_rx" [5.10,121/221] net: phylink: Fix phylink_err() function name error in phylink_major_config [5.10,122/221] tipc: better validate user input in tipc_nl_retrieve_key() [5.10,123/221] tcp: relookup sock for RST+ACK packets handled by obsolete req sock [5.10,124/221] can: isotp: isotp_setsockopt(): only allow to set low level TX flags for CAN-FD [5.10,125/221] can: isotp: TX-path: ensure that CAN frame flags are initialized [5.10,126/221] can: peak_usb: add forgotten supported devices [5.10,127/221] can: flexcan: flexcan_chip_freeze(): fix chip freeze for missing bitrate [5.10,128/221] can: kvaser_pciefd: Always disable bus load reporting [5.10,129/221] can: c_can_pci: c_can_pci_remove(): fix use-after-free [5.10,130/221] can: c_can: move runtime PM enable/disable to c_can_platform [5.10,131/221] can: m_can: m_can_do_rx_poll(): fix extraneous msg loss warning [5.10,132/221] can: m_can: m_can_rx_peripheral(): fix RX being blocked by errors [5.10,133/221] mac80211: fix rate mask reset [5.10,134/221] mac80211: Allow HE operation to be longer than expected. [5.10,135/221] selftests/net: fix warnings on reuseaddr_ports_exhausted [5.10,136/221] nfp: flower: fix unsupported pre_tunnel flows [5.10,137/221] nfp: flower: add ipv6 bit to pre_tunnel control message [5.10,138/221] nfp: flower: fix pre_tun mask id allocation [5.10,139/221] ftrace: Fix modify_ftrace_direct. [5.10,140/221] drm/msm/dsi: fix check-before-set in the 7nm dsi_pll code [5.10,141/221] ionic: linearize tso skb with too many frags [5.10,142/221] net/sched: cls_flower: fix only mask bit check in the validate_ct_state [5.10,143/221] netfilter: nftables: report EOPNOTSUPP on unsupported flowtable flags [5.10,144/221] netfilter: nftables: allow to update flowtable flags [5.10,145/221] netfilter: flowtable: Make sure GC works periodically in idle system [5.10,146/221] libbpf: Fix error path in bpf_object__elf_init() [5.10,147/221] libbpf: Use SOCK_CLOEXEC when opening the netlink socket [5.10,148/221] ARM: dts: imx6ull: fix ubi filesystem mount failed [5.10,149/221] ipv6: weaken the v4mapped source check [5.10,150/221] octeontx2-af: Formatting debugfs entry rsrc_alloc. [5.10,151/221] octeontx2-af: Modify default KEX profile to extract TX packet fields [5.10,152/221] octeontx2-af: Remove TOS field from MKEX TX [5.10,153/221] octeontx2-af: Fix irq free in rvu teardown [5.10,154/221] octeontx2-pf: Clear RSS enable flag on interace down [5.10,155/221] octeontx2-af: fix infinite loop in unmapping NPC counter [5.10,156/221] net: check all name nodes in __dev_alloc_name [5.10,157/221] net: cdc-phonet: fix data-interface release on probe failure [5.10,158/221] igb: check timestamp validity [5.10,159/221] r8152: limit the RX buffer size of RTL8153A for USB 2.0 [5.10,160/221] net: stmmac: dwmac-sun8i: Provide TX and RX fifo sizes [5.10,161/221] selinux: vsock: Set SID for socket returned by accept() [5.10,162/221] selftests: forwarding: vxlan_bridge_1d: Fix vxlan ecn decapsulate value [5.10,163/221] libbpf: Fix BTF dump of pointer-to-array-of-struct [5.10,164/221] bpf: Fix umd memory leak in copy_process() [5.10,165/221] can: isotp: tx-path: zero initialize outgoing CAN frames [5.10,166/221] drm/msm: fix shutdown hook in case GPU components failed to bind [5.10,167/221] drm/msm: Fix suspend/resume on i.MX5 [5.10,168/221] arm64: kdump: update ppos when reading elfcorehdr [5.10,169/221] PM: runtime: Defer suspending suppliers [5.10,170/221] net/mlx5: Add back multicast stats for uplink representor [5.10,171/221] net/mlx5e: Allow to match on MPLS parameters only for MPLS over UDP [5.10,172/221] net/mlx5e: Offload tuple rewrite for non-CT flows [5.10,173/221] net/mlx5e: Fix error path for ethtool set-priv-flag [5.10,174/221] PM: EM: postpone creating the debugfs dir till fs_initcall [5.10,175/221] net: bridge: dont notify switchdev for local FDB addresses [5.10,176/221] octeontx2-af: Fix memory leak of object buf [5.10,177/221] xen/x86: make XEN_BALLOON_MEMORY_HOTPLUG_LIMIT depend on MEMORY_HOTPLUG [5.10,178/221] RDMA/cxgb4: Fix adapter LE hash errors while destroying ipv6 listening server [5.10,179/221] bpf: Dont do bpf_cgroup_storage_set() for kuprobe/tp programs [5.10,180/221] net: Consolidate common blackhole dst ops [5.10,181/221] net, bpf: Fix ip6ip6 crash with collect_md populated skbs [5.10,182/221] igb: avoid premature Rx buffer reuse [5.10,183/221] net: axienet: Properly handle PCS/PMA PHY for 1000BaseX mode [5.10,184/221] net: axienet: Fix probe error cleanup [5.10,185/221] net: phy: introduce phydev->port [5.10,186/221] net: phy: broadcom: Avoid forward for bcm54xx_config_clock_delay() [5.10,187/221] net: phy: broadcom: Set proper 1000BaseX/SGMII interface mode for BCM54616S [5.10,188/221] net: phy: broadcom: Fix RGMII delays for BCM50160 and BCM50610M [5.10,189/221] Revert "netfilter: x_tables: Switch synchronization to RCU" [5.10,190/221] netfilter: x_tables: Use correct memory barriers. [5.10,191/221] arm64/mm: define arch_get_mappable_range() [5.10,192/221] arm64: mm: correct the inside linear map range during hotplug check [5.10,193/221] dm table: Fix zoned model check and zone sectors check [5.10,194/221] mm/mmu_notifiers: ensure range_end() is paired with range_start() [5.10,195/221] Revert "netfilter: x_tables: Update remaining dereference to RCU" [5.10,196/221] ACPI: scan: Rearrange memory allocation in acpi_device_add() [5.10,197/221] ACPI: scan: Use unique number for instance_no [5.10,198/221] perf auxtrace: Fix auxtrace queue conflict [5.10,199/221] perf synthetic events: Avoid write of uninitialized memory when generating PERF_RE... [5.10,200/221] io_uring: fix provide_buffers sign extension [5.10,201/221] block: recalculate segment count for multi-segment discards correctly [5.10,202/221] scsi: Revert "qla2xxx: Make sure that aborted commands are freed" [5.10,203/221] scsi: qedi: Fix error return code of qedi_alloc_global_queues() [5.10,204/221] scsi: mpt3sas: Fix error return code of mpt3sas_base_attach() [5.10,205/221] smb3: fix cached file size problems in duplicate extents (reflink) [5.10,206/221] cifs: Adjust key sizes and key generation routines for AES256 encryption [5.10,207/221] locking/mutex: Fix non debug version of mutex_lock_io_nested() [5.10,208/221] x86/mem_encrypt: Correct physical address calculation in __set_clr_pte_enc() [5.10,209/221] mm/memcg: fix 5.10 backport of splitting page memcg [5.10,210/221] fs/cachefiles: Remove wait_bit_key layout dependency [5.10,211/221] ch_ktls: fix enum-conversion warning [5.10,212/221] can: dev: Move device back to init netns on owning netns delete [5.10,213/221] r8169: fix DMA being used after buffer free if WoL is enabled [5.10,214/221] net: dsa: b53: VLAN filtering is global to all users [5.10,215/221] mac80211: fix double free in ibss_leave [5.10,216/221] ext4: add reclaim checks to xattr code [5.10,217/221] fs/ext4: fix integer overflow in s_log_groups_per_flex [5.10,218/221] Revert "xen: fix p2m size in dom0 for disabled memory hotplug case" [5.10,219/221] Revert "net: bonding: fix error return code of bond_neigh_init()" [5.10,220/221] nvme: fix the nsid value to print in nvme_validate_or_alloc_ns [5.10,221/221] can: peak_usb: Revert "can: peak_usb: add forgotten supported devices"

Message ID

20210329075634.718810265@linuxfoundation.org

State

New

Headers

From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	stable@vger.kernel.org, Fabio Estevam <festevam@gmail.com>,
	Rob Clark <robdclark@chromium.org>, Sasha Levin <sashal@kernel.org>
Subject: [PATCH 5.10 167/221] drm/msm: Fix suspend/resume on i.MX5
Date: Mon, 29 Mar 2021 09:58:18 +0200
Message-Id: <20210329075634.718810265@linuxfoundation.org>
In-Reply-To: <20210329075629.172032742@linuxfoundation.org>
References: <20210329075629.172032742@linuxfoundation.org>
User-Agent: quilt/0.66
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Precedence: bulk

Series

None | expand

Commit Message

Greg Kroah-Hartman March 29, 2021, 7:58 a.m. UTC

From: Fabio Estevam <festevam@gmail.com>

[ Upstream commit a9748134ea4aad989e52a6a91479e0acfd306e5b ]

When putting iMX5 into suspend, the following flow is
observed:

[   70.023427] [<c07755f0>] (msm_atomic_commit_tail) from [<c06e7218>]
(commit_tail+0x9c/0x18c)
[   70.031890] [<c06e7218>] (commit_tail) from [<c0e2920c>]
(drm_atomic_helper_commit+0x1a0/0x1d4)
[   70.040627] [<c0e2920c>] (drm_atomic_helper_commit) from
[<c06e74d4>] (drm_atomic_helper_disable_all+0x1c4/0x1d4)
[   70.050913] [<c06e74d4>] (drm_atomic_helper_disable_all) from
[<c0e2943c>] (drm_atomic_helper_suspend+0xb8/0x170)
[   70.061198] [<c0e2943c>] (drm_atomic_helper_suspend) from
[<c06e84bc>] (drm_mode_config_helper_suspend+0x24/0x58)

In the i.MX5 case, priv->kms is not populated (as i.MX5 does not use any
of the Qualcomm display controllers), causing a NULL pointer
dereference in msm_atomic_commit_tail():

[   24.268964] 8<--- cut here ---
[   24.274602] Unable to handle kernel NULL pointer dereference at
virtual address 00000000
[   24.283434] pgd = (ptrval)
[   24.286387] [00000000] *pgd=ca212831
[   24.290788] Internal error: Oops: 17 [#1] SMP ARM
[   24.295609] Modules linked in:
[   24.298777] CPU: 0 PID: 197 Comm: init Not tainted 5.11.0-rc2-next-20210111 #333
[   24.306276] Hardware name: Freescale i.MX53 (Device Tree Support)
[   24.312442] PC is at msm_atomic_commit_tail+0x54/0xb9c
[   24.317743] LR is at commit_tail+0xa4/0x1b0

Fix the problem by calling drm_mode_config_helper_suspend/resume()
only when priv->kms is available.

Fixes: ca8199f13498 ("drm/msm/dpu: ensure device suspend happens during PM sleep")
Signed-off-by: Fabio Estevam <festevam@gmail.com>
Signed-off-by: Rob Clark <robdclark@chromium.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/msm/msm_drv.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/drivers/gpu/drm/msm/msm_drv.c b/drivers/gpu/drm/msm/msm_drv.c
index 45e325c982c2..b38ebccad42f 100644
--- a/drivers/gpu/drm/msm/msm_drv.c
+++ b/drivers/gpu/drm/msm/msm_drv.c
@@ -1079,6 +1079,10 @@  static int __maybe_unused msm_pm_resume(struct device *dev)
 static int __maybe_unused msm_pm_prepare(struct device *dev)
 {
 	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct msm_drm_private *priv = ddev ? ddev->dev_private : NULL;
+
+	if (!priv || !priv->kms)
+		return 0;
 
 	return drm_mode_config_helper_suspend(ddev);
 }
@@ -1086,6 +1090,10 @@  static int __maybe_unused msm_pm_prepare(struct device *dev)
 static void __maybe_unused msm_pm_complete(struct device *dev)
 {
 	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct msm_drm_private *priv = ddev ? ddev->dev_private : NULL;
+
+	if (!priv || !priv->kms)
+		return;
 
 	drm_mode_config_helper_resume(ddev);
 }

[5.10,167/221] drm/msm: Fix suspend/resume on i.MX5

Commit Message

Patch