]> git.ipfire.org Git - thirdparty/kernel/stable-queue.git/commitdiff
drop arm64-kasan-avoid-bad-virt_to_pfn.patch
authorGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sat, 7 Oct 2017 08:42:08 +0000 (10:42 +0200)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sat, 7 Oct 2017 08:42:08 +0000 (10:42 +0200)
queue-4.9/arm64-kasan-avoid-bad-virt_to_pfn.patch [deleted file]
queue-4.9/series

diff --git a/queue-4.9/arm64-kasan-avoid-bad-virt_to_pfn.patch b/queue-4.9/arm64-kasan-avoid-bad-virt_to_pfn.patch
deleted file mode 100644 (file)
index 67391e0..0000000
+++ /dev/null
@@ -1,57 +0,0 @@
-From foo@baz Thu Oct  5 10:28:31 CEST 2017
-From: Mark Rutland <mark.rutland@arm.com>
-Date: Mon, 6 Mar 2017 19:06:40 +0000
-Subject: arm64: kasan: avoid bad virt_to_pfn()
-
-From: Mark Rutland <mark.rutland@arm.com>
-
-
-[ Upstream commit b0de0ccc8b9edd8846828e0ecdc35deacdf186b0 ]
-
-Booting a v4.11-rc1 kernel with DEBUG_VIRTUAL and KASAN enabled produces
-the following splat (trimmed for brevity):
-
-[    0.000000] virt_to_phys used for non-linear address: ffff200008080000 (0xffff200008080000)
-[    0.000000] WARNING: CPU: 0 PID: 0 at arch/arm64/mm/physaddr.c:14 __virt_to_phys+0x48/0x70
-[    0.000000] PC is at __virt_to_phys+0x48/0x70
-[    0.000000] LR is at __virt_to_phys+0x48/0x70
-[    0.000000] Call trace:
-[    0.000000] [<ffff2000080b1ac0>] __virt_to_phys+0x48/0x70
-[    0.000000] [<ffff20000a03b86c>] kasan_init+0x1c0/0x498
-[    0.000000] [<ffff20000a034018>] setup_arch+0x2fc/0x948
-[    0.000000] [<ffff20000a030c68>] start_kernel+0xb8/0x570
-[    0.000000] [<ffff20000a0301e8>] __primary_switched+0x6c/0x74
-
-This is because we use virt_to_pfn() on a kernel image address when
-trying to figure out its nid, so that we can allocate its shadow from
-the same node.
-
-As with other recent changes, this patch uses lm_alias() to solve this.
-
-We could instead use NUMA_NO_NODE, as x86 does for all shadow
-allocations, though we'll likely want the "real" memory shadow to be
-backed from its corresponding nid anyway, so we may as well be
-consistent and find the nid for the image shadow.
-
-Cc: Catalin Marinas <catalin.marinas@arm.com>
-Cc: Will Deacon <will.deacon@arm.com>
-Acked-by: Laura Abbott <labbott@redhat.com>
-Signed-off-by: Mark Rutland <mark.rutland@arm.com>
-Signed-off-by: Will Deacon <will.deacon@arm.com>
-Signed-off-by: Sasha Levin <alexander.levin@verizon.com>
-Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
----
- arch/arm64/mm/kasan_init.c |    2 +-
- 1 file changed, 1 insertion(+), 1 deletion(-)
-
---- a/arch/arm64/mm/kasan_init.c
-+++ b/arch/arm64/mm/kasan_init.c
-@@ -153,7 +153,7 @@ void __init kasan_init(void)
-       clear_pgds(KASAN_SHADOW_START, KASAN_SHADOW_END);
-       vmemmap_populate(kimg_shadow_start, kimg_shadow_end,
--                       pfn_to_nid(virt_to_pfn(_text)));
-+                       pfn_to_nid(virt_to_pfn(lm_alias(_text))));
-       /*
-        * vmemmap_populate() has populated the shadow region that covers the
index 43491524d9204c8a81d89a458224801cdb8cc6e2..0c2ebd70bdf6e32bb6707022b45d2fb1c0def7ee 100644 (file)
@@ -81,7 +81,6 @@ ibmvnic-free-tx-rx-scrq-pointer-array-when-releasing-sub-crqs.patch
 cpufreq-intel_pstate-update-pid_params.sample_rate_ns-in-pid_param_set.patch
 x86-acpi-restore-the-order-of-cpu-ids.patch
 iommu-io-pgtable-arm-check-for-leaf-entry-before-dereferencing-it.patch
-arm64-kasan-avoid-bad-virt_to_pfn.patch
 mm-cgroup-avoid-panic-when-init-with-low-memory.patch
 rds-ib-add-error-handle.patch
 md-raid10-submit-bio-directly-to-replacement-disk.patch