On Tue, Jul 14, 2020 at 10:05:28PM +0000, Suraj Jitindar Singh wrote: > This patch is to fix the backport of the upstream patch: > cc51e5428ea5 x86/speculation/l1tf: Increase l1tf memory limit for Nehalem+ > > When this was backported to the 4.9 and 4.14 stable branches the line > + c->x86_cache_bits = c->x86_phys_bits; > was applied in the wrong place, being added to the > identify_cpu_without_cpuid() function instead of the get_cpu_cap() > function which it was correctly applied to in the 4.4 backport. > > This means that x86_cache_bits is not set correctly resulting in the > following warning due to the cache bits being left uninitalised (zero). > > WARNING: CPU: 0 PID: 7566 at arch/x86/kvm/mmu.c:284 kvm_mmu_set_mmio_spte_mask+0x4e/0x60 [kvm > Modules linked in: kvm_intel(+) kvm irqbypass ipv6 crc_ccitt binfmt_misc evdev lpc_ich mfd_core ioatdma pcc_cpufreq dca ena acpi_power_meter hwmon acpi_pad button ext4 crc16 mbcache jbd2 fscrypto nvme nvme_core dm_mirror dm_region_hash dm_log dm_mod dax > Hardware name: Amazon EC2 i3.metal/Not Specified, BIOS 1.0 10/16/2017 > task: ffff88ff77704c00 task.stack: ffffc9000edac000 > RIP: 0010:kvm_mmu_set_mmio_spte_mask+0x4e/0x60 [kvm > RSP: 0018:ffffc9000edafc60 EFLAGS: 00010206 > RAX: 0000000000000000 RBX: 0000000000000000 RCX: 00000000ffffff45 > RDX: 000000000000002e RSI: 0008000000000001 RDI: 0008000000000001 > RBP: ffffffffa036f000 R08: ffffffffffffff80 R09: ffffe8ffffccb3c0 > R10: 0000000000000038 R11: 0000000000000000 R12: 0000000000005b80 > R13: ffffffffa0370e40 R14: 0000000000000001 R15: ffff88bf7c0927e0 > FS: 00007fa316f24740(0000) GS:ffff88bf7f600000(0000) knlGS:0000000000000000 > CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 > CR2: 00007fa316ea0000 CR3: 0000003f7e986004 CR4: 00000000003606f0 > DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 > DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 > Call Trace: > kvm_mmu_module_init+0x166/0x230 [kvm > kvm_arch_init+0x5d/0x150 [kvm > kvm_init+0x1c/0x2d0 [kvm > ? hardware_setup+0x4a6/0x4a6 [kvm_intel > vmx_init+0x23/0x6aa [kvm_intel > ? hardware_setup+0x4a6/0x4a6 [kvm_intel > do_one_initcall+0x3e/0x15d > do_init_module+0x5b/0x1e5 > load_module+0x19e6/0x1dc0 > ? SYSC_init_module+0x13b/0x170 > SYSC_init_module+0x13b/0x170 > do_syscall_64+0x67/0x110 > entry_SYSCALL_64_after_hwframe+0x41/0xa6 > RIP: 0033:0x7fa316828f3a > RSP: 002b:00007ffc9d65c1f8 EFLAGS: 00000246 ORIG_RAX: 00000000000000af > RAX: ffffffffffffffda RBX: 00007fa316b08849 RCX: 00007fa316828f3a > RDX: 00007fa316b08849 RSI: 0000000000071328 RDI: 00007fa316e37000 > RBP: 0000000000b47e80 R08: 0000000000000003 R09: 0000000000000000 > R10: 00007fa316822dba R11: 0000000000000246 R12: 0000000000b46340 > R13: 0000000000b464c0 R14: 0000000000000000 R15: 0000000000040000 > Code: e9 65 06 00 75 25 48 b8 00 00 00 00 00 00 00 40 48 09 c6 48 09 c7 48 89 35 f8 65 06 00 48 89 3d f9 65 06 00 c3 0f 0b 0f 0b eb d2 <0f> 0b eb d7 0f 1f 40 00 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 > > Fixes: 4.9.x ef3d45c95764 x86/speculation/l1tf: Increase l1tf memory limit for Nehalem+ > Fixes: 4.14.x ec4034835eaf x86/speculation/l1tf: Increase l1tf memory limit for Nehalem+ > Cc: stable@xxxxxxxxxxxxxxx # 4.9.x-4.14.x > Signed-off-by: Suraj Jitindar Singh <surajjs@xxxxxxxxxx> > Reviewed-by: Samuel Mendoza-Jonas <samjonas@xxxxxxxxxx> > Reviewed-by: Frank van der Linden <fllinden@xxxxxxxxxx> > --- > arch/x86/kernel/cpu/common.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) Wonderful, thanks for this, now queued up. greg k-h
