Currently kernel_page_present() function doesn't support huge page
detection causes the function to mistakenly return false to the
hibernation core.
Add huge page detection to the function to solve the problem.
Signed-off-by: Sia Jee Heng <[email protected]>
Reviewed-by: Ley Foon Tan <[email protected]>
Reviewed-by: Mason Huo <[email protected]>
---
arch/riscv/mm/pageattr.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/arch/riscv/mm/pageattr.c b/arch/riscv/mm/pageattr.c
index 86c56616e5de..73fdec8c0a72 100644
--- a/arch/riscv/mm/pageattr.c
+++ b/arch/riscv/mm/pageattr.c
@@ -221,14 +221,20 @@ bool kernel_page_present(struct page *page)
p4d = p4d_offset(pgd, addr);
if (!p4d_present(*p4d))
return false;
+ if (p4d_leaf(*pud))
+ return true;
pud = pud_offset(p4d, addr);
if (!pud_present(*pud))
return false;
+ if (pud_leaf(*pud))
+ return true;
pmd = pmd_offset(pud, addr);
if (!pmd_present(*pmd))
return false;
+ if (pmd_leaf(*pmd))
+ return true;
pte = pte_offset_kernel(pmd, addr);
return pte_present(*pte);
--
2.34.1
On Mon, Jan 09, 2023 at 02:24:06PM +0800, Sia Jee Heng wrote:
> Currently kernel_page_present() function doesn't support huge page
> detection causes the function to mistakenly return false to the
> hibernation core.
>
> Add huge page detection to the function to solve the problem.
>
> Signed-off-by: Sia Jee Heng <[email protected]>
> Reviewed-by: Ley Foon Tan <[email protected]>
> Reviewed-by: Mason Huo <[email protected]>
> ---
> arch/riscv/mm/pageattr.c | 6 ++++++
> 1 file changed, 6 insertions(+)
>
> diff --git a/arch/riscv/mm/pageattr.c b/arch/riscv/mm/pageattr.c
> index 86c56616e5de..73fdec8c0a72 100644
> --- a/arch/riscv/mm/pageattr.c
> +++ b/arch/riscv/mm/pageattr.c
> @@ -221,14 +221,20 @@ bool kernel_page_present(struct page *page)
> p4d = p4d_offset(pgd, addr);
> if (!p4d_present(*p4d))
> return false;
> + if (p4d_leaf(*pud))
^ p4d
I guess you got lucky with the stack garbage in your testing.
> + return true;
>
> pud = pud_offset(p4d, addr);
> if (!pud_present(*pud))
> return false;
> + if (pud_leaf(*pud))
> + return true;
>
> pmd = pmd_offset(pud, addr);
> if (!pmd_present(*pmd))
> return false;
> + if (pmd_leaf(*pmd))
> + return true;
>
> pte = pte_offset_kernel(pmd, addr);
> return pte_present(*pte);
> --
> 2.34.1
>
Thanks,
drew
> -----Original Message-----
> From: Andrew Jones <[email protected]>
> Sent: Monday, 9 January, 2023 10:45 PM
> To: JeeHeng Sia <[email protected]>
> Cc: [email protected]; [email protected];
> [email protected]; [email protected]; linux-
> [email protected]; Leyfoon Tan <[email protected]>;
> Mason Huo <[email protected]>
> Subject: Re: [PATCH v2 2/3] RISC-V: mm: Enable huge page support to
> kernel_page_present() function
>
> On Mon, Jan 09, 2023 at 02:24:06PM +0800, Sia Jee Heng wrote:
> > Currently kernel_page_present() function doesn't support huge page
> > detection causes the function to mistakenly return false to the
> > hibernation core.
> >
> > Add huge page detection to the function to solve the problem.
> >
> > Signed-off-by: Sia Jee Heng <[email protected]>
> > Reviewed-by: Ley Foon Tan <[email protected]>
> > Reviewed-by: Mason Huo <[email protected]>
> > ---
> > arch/riscv/mm/pageattr.c | 6 ++++++
> > 1 file changed, 6 insertions(+)
> >
> > diff --git a/arch/riscv/mm/pageattr.c b/arch/riscv/mm/pageattr.c
> > index 86c56616e5de..73fdec8c0a72 100644
> > --- a/arch/riscv/mm/pageattr.c
> > +++ b/arch/riscv/mm/pageattr.c
> > @@ -221,14 +221,20 @@ bool kernel_page_present(struct page *page)
> > p4d = p4d_offset(pgd, addr);
> > if (!p4d_present(*p4d))
> > return false;
> > + if (p4d_leaf(*pud))
> ^ p4d
>
> I guess you got lucky with the stack garbage in your testing.
You are right. But the reason it is added here is because kernel page table walk is checking for the p4d_leaf as well.
For example: walk_p4d_range() in the mm/pagewalk.c. So, I just trying to make it consistent to the Kernel page table walk mechanism.
Should I remove it?
>
> > + return true;
> >
> > pud = pud_offset(p4d, addr);
> > if (!pud_present(*pud))
> > return false;
> > + if (pud_leaf(*pud))
> > + return true;
> >
> > pmd = pmd_offset(pud, addr);
> > if (!pmd_present(*pmd))
> > return false;
> > + if (pmd_leaf(*pmd))
> > + return true;
> >
> > pte = pte_offset_kernel(pmd, addr);
> > return pte_present(*pte);
> > --
> > 2.34.1
> >
>
> Thanks,
> drew
On Tue, Jan 10, 2023 at 06:45:19AM +0000, JeeHeng Sia wrote:
>
>
> > -----Original Message-----
> > From: Andrew Jones <[email protected]>
> > Sent: Monday, 9 January, 2023 10:45 PM
> > To: JeeHeng Sia <[email protected]>
> > Cc: [email protected]; [email protected];
> > [email protected]; [email protected]; linux-
> > [email protected]; Leyfoon Tan <[email protected]>;
> > Mason Huo <[email protected]>
> > Subject: Re: [PATCH v2 2/3] RISC-V: mm: Enable huge page support to
> > kernel_page_present() function
> >
> > On Mon, Jan 09, 2023 at 02:24:06PM +0800, Sia Jee Heng wrote:
> > > Currently kernel_page_present() function doesn't support huge page
> > > detection causes the function to mistakenly return false to the
> > > hibernation core.
> > >
> > > Add huge page detection to the function to solve the problem.
> > >
> > > Signed-off-by: Sia Jee Heng <[email protected]>
> > > Reviewed-by: Ley Foon Tan <[email protected]>
> > > Reviewed-by: Mason Huo <[email protected]>
> > > ---
> > > arch/riscv/mm/pageattr.c | 6 ++++++
> > > 1 file changed, 6 insertions(+)
> > >
> > > diff --git a/arch/riscv/mm/pageattr.c b/arch/riscv/mm/pageattr.c
> > > index 86c56616e5de..73fdec8c0a72 100644
> > > --- a/arch/riscv/mm/pageattr.c
> > > +++ b/arch/riscv/mm/pageattr.c
> > > @@ -221,14 +221,20 @@ bool kernel_page_present(struct page *page)
> > > p4d = p4d_offset(pgd, addr);
> > > if (!p4d_present(*p4d))
> > > return false;
> > > + if (p4d_leaf(*pud))
> > ^ p4d
> >
> > I guess you got lucky with the stack garbage in your testing.
> You are right. But the reason it is added here is because kernel page table walk is checking for the p4d_leaf as well.
> For example: walk_p4d_range() in the mm/pagewalk.c. So, I just trying to make it consistent to the Kernel page table walk mechanism.
> Should I remove it?
The only thing I see wrong with it is the use of the uninitialized
pointer. Just change 'pud' to 'p4d'.
Thanks,
drew
> >
> > > + return true;
> > >
> > > pud = pud_offset(p4d, addr);
> > > if (!pud_present(*pud))
> > > return false;
> > > + if (pud_leaf(*pud))
> > > + return true;
> > >
> > > pmd = pmd_offset(pud, addr);
> > > if (!pmd_present(*pmd))
> > > return false;
> > > + if (pmd_leaf(*pmd))
> > > + return true;
> > >
> > > pte = pte_offset_kernel(pmd, addr);
> > > return pte_present(*pte);
> > > --
> > > 2.34.1
> > >
> >
> > Thanks,
> > drew