Re: [PATCH 4.4 05/27] x86/efi: Build our own page table structures

2017-12-14 Thread Greg Kroah-Hartman 
On Tue, Dec 12, 2017 at 05:24:32PM +, Ghannam, Yazen wrote:
> > -Original Message-
> > From: Greg Kroah-Hartman [mailto:gre...@linuxfoundation.org]
> > Sent: Sunday, December 10, 2017 4:27 PM
> > To: Ben Hutchings <ben.hutchi...@codethink.co.uk>; Ghannam, Yazen
> > <yazen.ghan...@amd.com>
> > Cc: linux-kernel@vger.kernel.org; sta...@vger.kernel.org; Matt Fleming
> > <m...@codeblueprint.co.uk>; Borislav Petkov <b...@suse.de>; Andrew
> > Morton <a...@linux-foundation.org>; Andy Lutomirski
> > <l...@amacapital.net>; Andy Lutomirski <l...@kernel.org>; Ard Biesheuvel
> > <ard.biesheu...@linaro.org>; Borislav Petkov <b...@alien8.de>; Brian Gerst
> > <brge...@gmail.com>; Dave Jones <da...@codemonkey.org.uk>; Denys
> > Vlasenko <dvlas...@redhat.com>; H. Peter Anvin <h...@zytor.com>; Linus
> > Torvalds <torva...@linux-foundation.org>; Peter Zijlstra
> > <pet...@infradead.org>; Sai Praneeth Prakhya
> > <sai.praneeth.prak...@intel.com>; Stephen Smalley <s...@tycho.nsa.gov>;
> > Thomas Gleixner <t...@linutronix.de>; Toshi Kani <toshi.k...@hp.com>;
> > linux-...@vger.kernel.org; Ingo Molnar <mi...@kernel.org>; Ghannam,
> > Yazen <yazen.ghan...@amd.com>; Alex Thorlton <athorl...@sgi.com>
> > Subject: Re: [PATCH 4.4 05/27] x86/efi: Build our own page table structures
> > 
> > On Thu, Dec 07, 2017 at 10:43:49PM +, Ben Hutchings wrote:
> > > On Mon, 2017-12-04 at 16:59 +0100, Greg Kroah-Hartman wrote:
> > > > 4.4-stable review patch.  If anyone has any objections, please let me
> > know.
> > > >
> > > > --
> > > >
> > > > From: Matt Fleming <m...@codeblueprint.co.uk>
> > > >
> > > > commit 67a9108ed4313b85a9c53406d80dc1ae3f8c3e36 upstream.
> > > [...]
> > >
> > > It looks like this will cause regressions unless you also apply:
> > >
> > > 21cdb6b56843 x86/mm: Page align the '_end' symbol to avoid pfn
> > conversion bugs
> > > b61a76f8850d x86/efi: Map RAM into the identity page table for mixed
> > mode
> > > 753b11ef8e92 x86/efi: Setup separate EFI page tables in kexec paths
> > >
> > > and possibly also:
> > >
> > > 08914f436bdd x86/platform/UV: Bring back the call to map_low_mmrs in
> > uv_system_init
> > > 683ad8092cd2 x86/efi: Fix 7-parameter efi_call()s
> > > 80e75596079f efi: Convert efi_call_virt() to efi_call_virt_pointer()
> > > d1be84a232e3 x86/uv: Update uv_bios_call() to use efi_call_virt_pointer()
> > 
> > Ugh, that's a mess.  Yazen is the one who reported this patch, and 2
> > others, were needed to resolve an issue on their hardware.
> > 
> > Yazen, any thoughts here?  Should I just revert the 3 patches, or do you
> > also feel I need all 7 of these to ensure everything works properly?
> > 
> 
> I think it's okay to revert the 3 patches.

Ugh, messy.  Ok, all now reverted.

Thanks Ben for the review, much appreciated.

greg k-h

Re: [PATCH 4.4 05/27] x86/efi: Build our own page table structures

2017-12-14 Thread Greg Kroah-Hartman 
On Tue, Dec 12, 2017 at 05:24:32PM +, Ghannam, Yazen wrote:
> > -Original Message-
> > From: Greg Kroah-Hartman [mailto:gre...@linuxfoundation.org]
> > Sent: Sunday, December 10, 2017 4:27 PM
> > To: Ben Hutchings ; Ghannam, Yazen
> > 
> > Cc: linux-kernel@vger.kernel.org; sta...@vger.kernel.org; Matt Fleming
> > ; Borislav Petkov ; Andrew
> > Morton ; Andy Lutomirski
> > ; Andy Lutomirski ; Ard Biesheuvel
> > ; Borislav Petkov ; Brian Gerst
> > ; Dave Jones ; Denys
> > Vlasenko ; H. Peter Anvin ; Linus
> > Torvalds ; Peter Zijlstra
> > ; Sai Praneeth Prakhya
> > ; Stephen Smalley ;
> > Thomas Gleixner ; Toshi Kani ;
> > linux-...@vger.kernel.org; Ingo Molnar ; Ghannam,
> > Yazen ; Alex Thorlton 
> > Subject: Re: [PATCH 4.4 05/27] x86/efi: Build our own page table structures
> > 
> > On Thu, Dec 07, 2017 at 10:43:49PM +, Ben Hutchings wrote:
> > > On Mon, 2017-12-04 at 16:59 +0100, Greg Kroah-Hartman wrote:
> > > > 4.4-stable review patch.  If anyone has any objections, please let me
> > know.
> > > >
> > > > --
> > > >
> > > > From: Matt Fleming 
> > > >
> > > > commit 67a9108ed4313b85a9c53406d80dc1ae3f8c3e36 upstream.
> > > [...]
> > >
> > > It looks like this will cause regressions unless you also apply:
> > >
> > > 21cdb6b56843 x86/mm: Page align the '_end' symbol to avoid pfn
> > conversion bugs
> > > b61a76f8850d x86/efi: Map RAM into the identity page table for mixed
> > mode
> > > 753b11ef8e92 x86/efi: Setup separate EFI page tables in kexec paths
> > >
> > > and possibly also:
> > >
> > > 08914f436bdd x86/platform/UV: Bring back the call to map_low_mmrs in
> > uv_system_init
> > > 683ad8092cd2 x86/efi: Fix 7-parameter efi_call()s
> > > 80e75596079f efi: Convert efi_call_virt() to efi_call_virt_pointer()
> > > d1be84a232e3 x86/uv: Update uv_bios_call() to use efi_call_virt_pointer()
> > 
> > Ugh, that's a mess.  Yazen is the one who reported this patch, and 2
> > others, were needed to resolve an issue on their hardware.
> > 
> > Yazen, any thoughts here?  Should I just revert the 3 patches, or do you
> > also feel I need all 7 of these to ensure everything works properly?
> > 
> 
> I think it's okay to revert the 3 patches.

Ugh, messy.  Ok, all now reverted.

Thanks Ben for the review, much appreciated.

greg k-h

RE: [PATCH 4.4 05/27] x86/efi: Build our own page table structures

2017-12-12 Thread Ghannam, Yazen 
> -Original Message-
> From: Greg Kroah-Hartman [mailto:gre...@linuxfoundation.org]
> Sent: Sunday, December 10, 2017 4:27 PM
> To: Ben Hutchings <ben.hutchi...@codethink.co.uk>; Ghannam, Yazen
> <yazen.ghan...@amd.com>
> Cc: linux-kernel@vger.kernel.org; sta...@vger.kernel.org; Matt Fleming
> <m...@codeblueprint.co.uk>; Borislav Petkov <b...@suse.de>; Andrew
> Morton <a...@linux-foundation.org>; Andy Lutomirski
> <l...@amacapital.net>; Andy Lutomirski <l...@kernel.org>; Ard Biesheuvel
> <ard.biesheu...@linaro.org>; Borislav Petkov <b...@alien8.de>; Brian Gerst
> <brge...@gmail.com>; Dave Jones <da...@codemonkey.org.uk>; Denys
> Vlasenko <dvlas...@redhat.com>; H. Peter Anvin <h...@zytor.com>; Linus
> Torvalds <torva...@linux-foundation.org>; Peter Zijlstra
> <pet...@infradead.org>; Sai Praneeth Prakhya
> <sai.praneeth.prak...@intel.com>; Stephen Smalley <s...@tycho.nsa.gov>;
> Thomas Gleixner <t...@linutronix.de>; Toshi Kani <toshi.k...@hp.com>;
> linux-...@vger.kernel.org; Ingo Molnar <mi...@kernel.org>; Ghannam,
> Yazen <yazen.ghan...@amd.com>; Alex Thorlton <athorl...@sgi.com>
> Subject: Re: [PATCH 4.4 05/27] x86/efi: Build our own page table structures
> 
> On Thu, Dec 07, 2017 at 10:43:49PM +, Ben Hutchings wrote:
> > On Mon, 2017-12-04 at 16:59 +0100, Greg Kroah-Hartman wrote:
> > > 4.4-stable review patch.  If anyone has any objections, please let me
> know.
> > >
> > > --
> > >
> > > From: Matt Fleming <m...@codeblueprint.co.uk>
> > >
> > > commit 67a9108ed4313b85a9c53406d80dc1ae3f8c3e36 upstream.
> > [...]
> >
> > It looks like this will cause regressions unless you also apply:
> >
> > 21cdb6b56843 x86/mm: Page align the '_end' symbol to avoid pfn
> conversion bugs
> > b61a76f8850d x86/efi: Map RAM into the identity page table for mixed
> mode
> > 753b11ef8e92 x86/efi: Setup separate EFI page tables in kexec paths
> >
> > and possibly also:
> >
> > 08914f436bdd x86/platform/UV: Bring back the call to map_low_mmrs in
> uv_system_init
> > 683ad8092cd2 x86/efi: Fix 7-parameter efi_call()s
> > 80e75596079f efi: Convert efi_call_virt() to efi_call_virt_pointer()
> > d1be84a232e3 x86/uv: Update uv_bios_call() to use efi_call_virt_pointer()
> 
> Ugh, that's a mess.  Yazen is the one who reported this patch, and 2
> others, were needed to resolve an issue on their hardware.
> 
> Yazen, any thoughts here?  Should I just revert the 3 patches, or do you
> also feel I need all 7 of these to ensure everything works properly?
> 

I think it's okay to revert the 3 patches.

Thanks,
Yazen

RE: [PATCH 4.4 05/27] x86/efi: Build our own page table structures

2017-12-12 Thread Ghannam, Yazen 
> -Original Message-
> From: Greg Kroah-Hartman [mailto:gre...@linuxfoundation.org]
> Sent: Sunday, December 10, 2017 4:27 PM
> To: Ben Hutchings ; Ghannam, Yazen
> 
> Cc: linux-kernel@vger.kernel.org; sta...@vger.kernel.org; Matt Fleming
> ; Borislav Petkov ; Andrew
> Morton ; Andy Lutomirski
> ; Andy Lutomirski ; Ard Biesheuvel
> ; Borislav Petkov ; Brian Gerst
> ; Dave Jones ; Denys
> Vlasenko ; H. Peter Anvin ; Linus
> Torvalds ; Peter Zijlstra
> ; Sai Praneeth Prakhya
> ; Stephen Smalley ;
> Thomas Gleixner ; Toshi Kani ;
> linux-...@vger.kernel.org; Ingo Molnar ; Ghannam,
> Yazen ; Alex Thorlton 
> Subject: Re: [PATCH 4.4 05/27] x86/efi: Build our own page table structures
> 
> On Thu, Dec 07, 2017 at 10:43:49PM +, Ben Hutchings wrote:
> > On Mon, 2017-12-04 at 16:59 +0100, Greg Kroah-Hartman wrote:
> > > 4.4-stable review patch.  If anyone has any objections, please let me
> know.
> > >
> > > --
> > >
> > > From: Matt Fleming 
> > >
> > > commit 67a9108ed4313b85a9c53406d80dc1ae3f8c3e36 upstream.
> > [...]
> >
> > It looks like this will cause regressions unless you also apply:
> >
> > 21cdb6b56843 x86/mm: Page align the '_end' symbol to avoid pfn
> conversion bugs
> > b61a76f8850d x86/efi: Map RAM into the identity page table for mixed
> mode
> > 753b11ef8e92 x86/efi: Setup separate EFI page tables in kexec paths
> >
> > and possibly also:
> >
> > 08914f436bdd x86/platform/UV: Bring back the call to map_low_mmrs in
> uv_system_init
> > 683ad8092cd2 x86/efi: Fix 7-parameter efi_call()s
> > 80e75596079f efi: Convert efi_call_virt() to efi_call_virt_pointer()
> > d1be84a232e3 x86/uv: Update uv_bios_call() to use efi_call_virt_pointer()
> 
> Ugh, that's a mess.  Yazen is the one who reported this patch, and 2
> others, were needed to resolve an issue on their hardware.
> 
> Yazen, any thoughts here?  Should I just revert the 3 patches, or do you
> also feel I need all 7 of these to ensure everything works properly?
> 

I think it's okay to revert the 3 patches.

Thanks,
Yazen

Re: [PATCH 4.4 05/27] x86/efi: Build our own page table structures

2017-12-10 Thread Greg Kroah-Hartman 
On Thu, Dec 07, 2017 at 10:43:49PM +, Ben Hutchings wrote:
> On Mon, 2017-12-04 at 16:59 +0100, Greg Kroah-Hartman wrote:
> > 4.4-stable review patch.  If anyone has any objections, please let me know.
> > 
> > --
> > 
> > From: Matt Fleming 
> > 
> > commit 67a9108ed4313b85a9c53406d80dc1ae3f8c3e36 upstream.
> [...]
> 
> It looks like this will cause regressions unless you also apply:
> 
> 21cdb6b56843 x86/mm: Page align the '_end' symbol to avoid pfn conversion bugs
> b61a76f8850d x86/efi: Map RAM into the identity page table for mixed mode
> 753b11ef8e92 x86/efi: Setup separate EFI page tables in kexec paths
> 
> and possibly also:
> 
> 08914f436bdd x86/platform/UV: Bring back the call to map_low_mmrs in 
> uv_system_init
> 683ad8092cd2 x86/efi: Fix 7-parameter efi_call()s
> 80e75596079f efi: Convert efi_call_virt() to efi_call_virt_pointer()
> d1be84a232e3 x86/uv: Update uv_bios_call() to use efi_call_virt_pointer()

Ugh, that's a mess.  Yazen is the one who reported this patch, and 2
others, were needed to resolve an issue on their hardware.

Yazen, any thoughts here?  Should I just revert the 3 patches, or do you
also feel I need all 7 of these to ensure everything works properly?

thanks,

greg k-h

Re: [PATCH 4.4 05/27] x86/efi: Build our own page table structures

2017-12-10 Thread Greg Kroah-Hartman 
On Thu, Dec 07, 2017 at 10:43:49PM +, Ben Hutchings wrote:
> On Mon, 2017-12-04 at 16:59 +0100, Greg Kroah-Hartman wrote:
> > 4.4-stable review patch.  If anyone has any objections, please let me know.
> > 
> > --
> > 
> > From: Matt Fleming 
> > 
> > commit 67a9108ed4313b85a9c53406d80dc1ae3f8c3e36 upstream.
> [...]
> 
> It looks like this will cause regressions unless you also apply:
> 
> 21cdb6b56843 x86/mm: Page align the '_end' symbol to avoid pfn conversion bugs
> b61a76f8850d x86/efi: Map RAM into the identity page table for mixed mode
> 753b11ef8e92 x86/efi: Setup separate EFI page tables in kexec paths
> 
> and possibly also:
> 
> 08914f436bdd x86/platform/UV: Bring back the call to map_low_mmrs in 
> uv_system_init
> 683ad8092cd2 x86/efi: Fix 7-parameter efi_call()s
> 80e75596079f efi: Convert efi_call_virt() to efi_call_virt_pointer()
> d1be84a232e3 x86/uv: Update uv_bios_call() to use efi_call_virt_pointer()

Ugh, that's a mess.  Yazen is the one who reported this patch, and 2
others, were needed to resolve an issue on their hardware.

Yazen, any thoughts here?  Should I just revert the 3 patches, or do you
also feel I need all 7 of these to ensure everything works properly?

thanks,

greg k-h

Re: [PATCH 4.4 05/27] x86/efi: Build our own page table structures

2017-12-07 Thread Ben Hutchings 
On Mon, 2017-12-04 at 16:59 +0100, Greg Kroah-Hartman wrote:
> 4.4-stable review patch.  If anyone has any objections, please let me know.
> 
> --
> 
> From: Matt Fleming 
> 
> commit 67a9108ed4313b85a9c53406d80dc1ae3f8c3e36 upstream.
[...]

It looks like this will cause regressions unless you also apply:

21cdb6b56843 x86/mm: Page align the '_end' symbol to avoid pfn conversion bugs
b61a76f8850d x86/efi: Map RAM into the identity page table for mixed mode
753b11ef8e92 x86/efi: Setup separate EFI page tables in kexec paths

and possibly also:

08914f436bdd x86/platform/UV: Bring back the call to map_low_mmrs in 
uv_system_init
683ad8092cd2 x86/efi: Fix 7-parameter efi_call()s
80e75596079f efi: Convert efi_call_virt() to efi_call_virt_pointer()
d1be84a232e3 x86/uv: Update uv_bios_call() to use efi_call_virt_pointer()

Ben.

-- 
Ben Hutchings
Software Developer, Codethink Ltd.

Re: [PATCH 4.4 05/27] x86/efi: Build our own page table structures

2017-12-07 Thread Ben Hutchings 
On Mon, 2017-12-04 at 16:59 +0100, Greg Kroah-Hartman wrote:
> 4.4-stable review patch.  If anyone has any objections, please let me know.
> 
> --
> 
> From: Matt Fleming 
> 
> commit 67a9108ed4313b85a9c53406d80dc1ae3f8c3e36 upstream.
[...]

It looks like this will cause regressions unless you also apply:

21cdb6b56843 x86/mm: Page align the '_end' symbol to avoid pfn conversion bugs
b61a76f8850d x86/efi: Map RAM into the identity page table for mixed mode
753b11ef8e92 x86/efi: Setup separate EFI page tables in kexec paths

and possibly also:

08914f436bdd x86/platform/UV: Bring back the call to map_low_mmrs in 
uv_system_init
683ad8092cd2 x86/efi: Fix 7-parameter efi_call()s
80e75596079f efi: Convert efi_call_virt() to efi_call_virt_pointer()
d1be84a232e3 x86/uv: Update uv_bios_call() to use efi_call_virt_pointer()

Ben.

-- 
Ben Hutchings
Software Developer, Codethink Ltd.