Re: [PATCH] x86/kexec: Only write through identity mapping of control page

2024-12-12 Thread Ning, Hongyu
On 2024/12/13 4:11, David Woodhouse wrote: From: David Woodhouse The virtual mapping of the control page may have been _PAGE_GLOBAL and thus its PTE might not have been flushed on the %cr3 switch and it might effectively still be read-only. Move the writes to it down into the identity_mapped

Re: [PATCH] x86/kexec: Only write through identity mapping of control page

2024-12-12 Thread David Woodhouse
On 12 December 2024 21:43:57 GMT, Dave Hansen wrote: >On 12/12/24 13:32, David Woodhouse wrote: >> On 12 December 2024 21:18:10 GMT, Dave Hansen wrote: >>> On 12/12/24 12:11, David Woodhouse wrote: From: David Woodhouse The virtual mapping of the control page may have been _PAGE_G

Re: [PATCH] x86/kexec: Only write through identity mapping of control page

2024-12-12 Thread Dave Hansen
On 12/12/24 13:32, David Woodhouse wrote: > On 12 December 2024 21:18:10 GMT, Dave Hansen wrote: >> On 12/12/24 12:11, David Woodhouse wrote: >>> From: David Woodhouse >>> >>> The virtual mapping of the control page may have been _PAGE_GLOBAL and >>> thus its PTE might not have been flushed on th

Re: [PATCH] x86/kexec: Only write through identity mapping of control page

2024-12-12 Thread David Woodhouse
On 12 December 2024 21:18:10 GMT, Dave Hansen wrote: >On 12/12/24 12:11, David Woodhouse wrote: >> From: David Woodhouse >> >> The virtual mapping of the control page may have been _PAGE_GLOBAL and >> thus its PTE might not have been flushed on the %cr3 switch and it might >> effectively still b

Re: [PATCH] x86/kexec: Only write through identity mapping of control page

2024-12-12 Thread Dave Hansen
On 12/12/24 12:11, David Woodhouse wrote: > From: David Woodhouse > > The virtual mapping of the control page may have been _PAGE_GLOBAL and > thus its PTE might not have been flushed on the %cr3 switch and it might > effectively still be read-only. Move the writes to it down into the > identity_

Re: [PATCH] x86/kexec: Only write through identity mapping of control page

2024-12-12 Thread Nathan Chancellor
On Thu, Dec 12, 2024 at 08:11:19PM +, David Woodhouse wrote: > From: David Woodhouse > > The virtual mapping of the control page may have been _PAGE_GLOBAL and > thus its PTE might not have been flushed on the %cr3 switch and it might > effectively still be read-only. Move the writes to it do

[PATCH] x86/kexec: Only write through identity mapping of control page

2024-12-12 Thread David Woodhouse
From: David Woodhouse The virtual mapping of the control page may have been _PAGE_GLOBAL and thus its PTE might not have been flushed on the %cr3 switch and it might effectively still be read-only. Move the writes to it down into the identity_mapped() function where the same %rip-relative address