diff mbox series

[5.10,114/146] KVM: X86: Fix x86_emulator slab cache leak

Message ID	20210621154918.625077118@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, Wanpeng Li <wanpengli@tencent.com>, Paolo Bonzini <pbonzini@redhat.com> Subject: [PATCH 5.10 114/146] KVM: X86: Fix x86_emulator slab cache leak Date: Mon, 21 Jun 2021 18:15:44 +0200 Message-Id: <20210621154918.625077118@linuxfoundation.org> In-Reply-To: <20210621154911.244649123@linuxfoundation.org> References: <20210621154911.244649123@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 [5.10,002/146] dmaengine: fsl-dpaa2-qdma: Fix error return code in two functions [5.10,004/146] dmaengine: ALTERA_MSGDMA depends on HAS_IOMEM [5.10,006/146] dmaengine: SF_PDMA depends on HAS_IOMEM [5.10,007/146] dmaengine: stedma40: add missing iounmap() on error in d40_probe() [5.10,008/146] afs: Fix an IS_ERR() vs NULL check [5.10,010/146] kvm: LAPIC: Restore guard to prevent illegal APIC register access [5.10,011/146] fanotify: fix copy_event_to_user() fid error clean up [5.10,013/146] mac80211: fix skb length check in ieee80211_scan_rx() [5.10,015/146] mlxsw: core: Set thermal zone polling delay argument to real value at init [5.10,017/146] net: ipv4: fix memory leak in netlbl_cipsov4_add_std [5.10,020/146] net: dsa: felix: re-enable TX flow control in ocelot_port_flush() [5.10,022/146] netfilter: nft_fib_ipv6: skip ipv6 packets from any to link-local [5.10,024/146] ice: parameterize functions responsible for Tx ring management [5.10,026/146] rtnetlink: Fix regression in bridge VLAN configuration [5.10,028/146] net/mlx5e: Remove dependency in IPsec initialization flows [5.10,031/146] net/mlx5: DR, Allow SW steering for sw_owner_v2 devices [5.10,033/146] net/mlx5e: Block offload of outer header csum for UDP tunnels [5.10,034/146] netfilter: synproxy: Fix out of bounds when parsing TCP options [5.10,037/146] mptcp: try harder to borrow memory from subflow under pressure [5.10,039/146] selftests: mptcp: enable syncookie only in absence of reorders [5.10,042/146] cxgb4: fix sleep in atomic when flashing PHY firmware [5.10,044/146] net: stmmac: dwmac1000: Fix extended MAC address registers definition [5.10,046/146] net: qualcomm: rmnet: Update rmnet device MTU based on real device [5.10,047/146] net: qualcomm: rmnet: dont over-count statistics [5.10,048/146] ethtool: strset: fix message length calculation [5.10,051/146] cxgb4: fix wrong ethtool n-tuple rule lookup [5.10,052/146] ipv4: Fix device used for dst_alloc with local routes [5.10,055/146] ptp: improve max_adj check against unreasonable values [5.10,056/146] net: cdc_ncm: switch to eth%d interface naming [5.10,059/146] net: fec_ptp: fix issue caused by refactor the fec_devtype [5.10,060/146] net: ipv4: fix memory leak in ip_mc_add1_src [5.10,062/146] net/mlx5: E-Switch, Read PF mac address [5.10,064/146] net/mlx5: Reset mkey index on creation [5.10,067/146] net: cdc_eem: fix tx fixup skb leak [5.10,068/146] cxgb4: fix wrong shift. [5.10,069/146] bnxt_en: Rediscover PHY capabilities after firmware reset [5.10,072/146] icmp: dont send out ICMP messages with a source address of 0.0.0.0 [5.10,074/146] regulator: cros-ec: Fix error code in dev_err message [5.10,076/146] regulator: bd70528: Fix off-by-one for buck123 .n_voltages setting [5.10,078/146] ASoC: rt5659: Fix the lost powers for the HDA header [5.10,079/146] phy: phy-mtk-tphy: Fix some resource leaks in mtk_phy_init() [5.10,080/146] ASoC: fsl-asoc-card: Set .owner attribute when registering card. [5.10,082/146] spi: spi-zynq-qspi: Fix some wrong goto jumps & missing error code [5.10,083/146] sched/pelt: Ensure that _sum is always synced with _avg [5.10,084/146] ASoC: tas2562: Fix TDM_CFG0_SAMPRATE values [5.10,086/146] regulator: rt4801: Fix NULL pointer dereference if priv->enable_gpios is NULL [5.10,087/146] ASoC: rt5682: Fix the fast discharge for headset unplugging in soundwire mode [5.10,089/146] drm/sun4i: dw-hdmi: Make HDMI PHY into a platform device [5.10,092/146] hwmon: (scpi-hwmon) shows the negative temperature properly [5.10,094/146] irqchip/gic-v3: Workaround inconsistent PMR setting on NMI entry [5.10,096/146] bpf: Inherit expanded/patched seen count from old aux data [5.10,098/146] can: bcm: fix infoleak in struct bcm_msg_head [5.10,099/146] can: bcm/raw/isotp: use per module netdevice notifier [5.10,101/146] can: mcba_usb: fix memory leak in mcba_usb [5.10,103/146] usb: chipidea: imx: Fix Battery Charger 1.2 CDP detection [5.10,105/146] tracing: Do not stop recording comms if the trace file is being read [5.10,107/146] PCI: Mark TI C667X to avoid bus reset [5.10,110/146] PCI: Add ACS quirk for Broadcom BCM57414 NIC [5.10,112/146] KVM: x86: Immediately reset the MMU context when the SMM flag is cleared [5.10,114/146] KVM: X86: Fix x86_emulator slab cache leak [5.10,118/146] x86/mm: Avoid truncating memblocks for SGX memory [5.10,119/146] x86/process: Check PF_KTHREAD and not current->mm for kernel threads [5.10,120/146] x86/ioremap: Map EFI-reserved memory as encrypted for SEV [5.10,121/146] x86/pkru: Write hardware init value to PKRU when xstate is init [5.10,124/146] x86/fpu: Reset state for all signal restore failures [5.10,125/146] crash_core, vmcoreinfo: append SECTION_SIZE_BITS to vmcoreinfo [5.10,128/146] cfg80211: make certificate generation more robust [5.10,129/146] cfg80211: avoid double free of PMSR request [5.10,130/146] drm/amdgpu/gfx10: enlarge CP_MEC_DOORBELL_RANGE_UPPER to cover full doorbell. [5.10,134/146] net: bridge: fix vlan tunnel dst null pointer dereference [5.10,135/146] net: bridge: fix vlan tunnel dst refcnt when egressing [5.10,137/146] mm/slub: clarify verification reporting [5.10,138/146] mm/slub: fix redzoning for small allocations [5.10,143/146] tools headers UAPI: Sync linux/in.h copy with the kernel sources [5.10,144/146] perf beauty: Update copy of linux/socket.h with the kernel sources

Commit Message

Greg KH June 21, 2021, 4:15 p.m. UTC

From: Wanpeng Li <wanpengli@tencent.com>

commit dfdc0a714d241bfbf951886c373cd1ae463fcc25 upstream.

Commit c9b8b07cded58 (KVM: x86: Dynamically allocate per-vCPU emulation context)
tries to allocate per-vCPU emulation context dynamically, however, the
x86_emulator slab cache is still exiting after the kvm module is unload
as below after destroying the VM and unloading the kvm module.

grep x86_emulator /proc/slabinfo
x86_emulator          36     36   2672   12    8 : tunables    0    0    0 : slabdata      3      3      0

This patch fixes this slab cache leak by destroying the x86_emulator slab cache
when the kvm module is unloaded.

Fixes: c9b8b07cded58 (KVM: x86: Dynamically allocate per-vCPU emulation context)
Cc: stable@vger.kernel.org
Signed-off-by: Wanpeng Li <wanpengli@tencent.com>
Message-Id: <1623387573-5969-1-git-send-email-wanpengli@tencent.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 arch/x86/kvm/x86.c |    1 +
 1 file changed, 1 insertion(+)

diff mbox series

Patch

--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -8021,6 +8021,7 @@  void kvm_arch_exit(void)
 	kvm_x86_ops.hardware_enable = NULL;
 	kvm_mmu_module_exit();
 	free_percpu(user_return_msrs);
+	kmem_cache_destroy(x86_emulator_cache);
 	kmem_cache_destroy(x86_fpu_cache);
 }