Re: [PATCH] x86/boot/64: Initialize 5-level paging variables earlier

2020-10-16 Thread Borislav Petkov
On Thu, Oct 15, 2020 at 04:52:09PM +0300, Kirill A. Shutemov wrote: > Yes, the patch helps to fix 5-level paging boot: > > Tested-by: Kirill A. Shutemov Thx! -- Regards/Gruss, Boris. https://people.kernel.org/tglx/notes-about-netiquette

Re: [PATCH] x86/boot/64: Initialize 5-level paging variables earlier

2020-10-15 Thread Kirill A. Shutemov
On Tue, Oct 13, 2020 at 10:33:58AM +0200, Borislav Petkov wrote: > On Tue, Oct 13, 2020 at 11:20:47AM +0300, Kirill A. Shutemov wrote: > > With TCG or KVM? I use -machine "type=q35,accel=tcg". > > Thx, that triggered it - it was KVM before. > > Btw, are 5level boxes shipping already or not yet?

Re: [PATCH] x86/boot/64: Initialize 5-level paging variables earlier

2020-10-13 Thread Borislav Petkov
On Tue, Oct 13, 2020 at 12:12:28PM +0300, Kirill A. Shutemov wrote: > Not yet. Oh good, that makes this fix a lot less critical then. > Okay, will do. Thx! -- Regards/Gruss, Boris. https://people.kernel.org/tglx/notes-about-netiquette

Re: [PATCH] x86/boot/64: Initialize 5-level paging variables earlier

2020-10-13 Thread Kirill A. Shutemov
On Tue, Oct 13, 2020 at 10:33:58AM +0200, Borislav Petkov wrote: > On Tue, Oct 13, 2020 at 11:20:47AM +0300, Kirill A. Shutemov wrote: > > With TCG or KVM? I use -machine "type=q35,accel=tcg". > > Thx, that triggered it - it was KVM before. > > Btw, are 5level boxes shipping already or not yet?

Re: [PATCH] x86/boot/64: Initialize 5-level paging variables earlier

2020-10-13 Thread Joerg Roedel
Hi Arvind, On Sat, Oct 10, 2020 at 03:11:10PM -0400, Arvind Sankar wrote: > Commit > ca0e22d4f011 ("x86/boot/compressed/64: Always switch to own page table") > started using a new set of pagetables even without KASLR. > > After that commit, initialize_identity_maps() is called before the >

Re: [PATCH] x86/boot/64: Initialize 5-level paging variables earlier

2020-10-13 Thread Borislav Petkov
On Tue, Oct 13, 2020 at 11:20:47AM +0300, Kirill A. Shutemov wrote: > With TCG or KVM? I use -machine "type=q35,accel=tcg". Thx, that triggered it - it was KVM before. Btw, are 5level boxes shipping already or not yet? Because I've not seen one yet. If you have access to the hw, I'd appreciate

Re: [PATCH] x86/boot/64: Initialize 5-level paging variables earlier

2020-10-13 Thread Kirill A. Shutemov
On Tue, Oct 13, 2020 at 10:11:17AM +0200, Borislav Petkov wrote: > On Mon, Oct 12, 2020 at 11:35:01AM -0400, Arvind Sankar wrote: > > > qemu supports it. -cpu "qemu64,+la57" > > > > Thanks! On QEMU, it does crash without this patch. > > Works fine here. I gave this to qemu: > > -cpu

Re: [PATCH] x86/boot/64: Initialize 5-level paging variables earlier

2020-10-13 Thread Borislav Petkov
On Mon, Oct 12, 2020 at 11:35:01AM -0400, Arvind Sankar wrote: > > qemu supports it. -cpu "qemu64,+la57" > > Thanks! On QEMU, it does crash without this patch. Works fine here. I gave this to qemu: -cpu qemu64,+la57,vendor=GenuineIntel it said: qemu-system-x86_64: warning: host doesn't support

Re: [PATCH] x86/boot/64: Initialize 5-level paging variables earlier

2020-10-12 Thread Arvind Sankar
On Mon, Oct 12, 2020 at 05:08:30PM +0300, Kirill A. Shutemov wrote: > On Sat, Oct 10, 2020 at 03:26:24PM -0400, Arvind Sankar wrote: > > On Sat, Oct 10, 2020 at 03:11:10PM -0400, Arvind Sankar wrote: > > > Commit > > > ca0e22d4f011 ("x86/boot/compressed/64: Always switch to own page table") > >

Re: [PATCH] x86/boot/64: Initialize 5-level paging variables earlier

2020-10-12 Thread Kirill A. Shutemov
On Sat, Oct 10, 2020 at 03:26:24PM -0400, Arvind Sankar wrote: > On Sat, Oct 10, 2020 at 03:11:10PM -0400, Arvind Sankar wrote: > > Commit > > ca0e22d4f011 ("x86/boot/compressed/64: Always switch to own page table") > > started using a new set of pagetables even without KASLR. > > > > After

Re: [PATCH] x86/boot/64: Initialize 5-level paging variables earlier

2020-10-10 Thread Arvind Sankar
On Sat, Oct 10, 2020 at 03:11:10PM -0400, Arvind Sankar wrote: > Commit > ca0e22d4f011 ("x86/boot/compressed/64: Always switch to own page table") > started using a new set of pagetables even without KASLR. > > After that commit, initialize_identity_maps() is called before the > 5-level paging

[PATCH] x86/boot/64: Initialize 5-level paging variables earlier

2020-10-10 Thread Arvind Sankar
Commit ca0e22d4f011 ("x86/boot/compressed/64: Always switch to own page table") started using a new set of pagetables even without KASLR. After that commit, initialize_identity_maps() is called before the 5-level paging variables are setup in choose_random_location(), which will not work if