Re: [PATCH v5 2/2] x86_64: fix KASan shadow region page tables

2015-06-12 Thread Andrey Ryabinin
2015-06-09 12:42 GMT+03:00 Alexander Popov : > Physical addresses in KASan shadow region page tables need fixup similarly > to the other page tables. Current code doesn't do it which causes > kernel halt if phys_base is not zero. > So let's initialize KASan shadow region page tables in

Re: [PATCH v5 2/2] x86_64: fix KASan shadow region page tables

2015-06-12 Thread Andrey Ryabinin
2015-06-09 12:42 GMT+03:00 Alexander Popov alpo...@ptsecurity.com: Physical addresses in KASan shadow region page tables need fixup similarly to the other page tables. Current code doesn't do it which causes kernel halt if phys_base is not zero. So let's initialize KASan shadow region page

[PATCH v5 2/2] x86_64: fix KASan shadow region page tables

2015-06-09 Thread Alexander Popov
Physical addresses in KASan shadow region page tables need fixup similarly to the other page tables. Current code doesn't do it which causes kernel halt if phys_base is not zero. So let's initialize KASan shadow region page tables in kasan_early_init() using __pa_nodebug() which considers

[PATCH v5 2/2] x86_64: fix KASan shadow region page tables

2015-06-09 Thread Alexander Popov
Physical addresses in KASan shadow region page tables need fixup similarly to the other page tables. Current code doesn't do it which causes kernel halt if phys_base is not zero. So let's initialize KASan shadow region page tables in kasan_early_init() using __pa_nodebug() which considers