The Linux RISC-V 32bit kernel is broken after we moved setup_vm() from
kernel/setup.c to mm/init.c because Linux RISC-V 32bit kernel by default
uses cmodel=medlow which results in a non-position-independent setup_vm().
This patch fixes Linux RISC-V 32bit kernel booting by:
1. Forcing cmodel=medany for mm/init.c
2. Moving remaing MM-related stuff va_pa_offset, pfn_base and
empty_zero_page from kernel/setup.c to mm/init.c
Fixes: 6f1e9e946f0b ("RISC-V: Move setup_vm() to mm/init.c")
Signed-off-by: Christoph Hellwig <[email protected]>
Signed-off-by: Mike Rapoport <[email protected]>
Signed-off-by: Anup Patel <[email protected]>
---
arch/riscv/kernel/setup.c | 8 --------
arch/riscv/mm/Makefile | 2 ++
arch/riscv/mm/init.c | 9 +++++++++
3 files changed, 11 insertions(+), 8 deletions(-)
diff --git a/arch/riscv/kernel/setup.c b/arch/riscv/kernel/setup.c
index ecb654f6a79e..540a331d1376 100644
--- a/arch/riscv/kernel/setup.c
+++ b/arch/riscv/kernel/setup.c
@@ -48,14 +48,6 @@ struct screen_info screen_info = {
};
#endif
-unsigned long va_pa_offset;
-EXPORT_SYMBOL(va_pa_offset);
-unsigned long pfn_base;
-EXPORT_SYMBOL(pfn_base);
-
-unsigned long empty_zero_page[PAGE_SIZE / sizeof(unsigned long)] __page_aligned_bss;
-EXPORT_SYMBOL(empty_zero_page);
-
/* The lucky hart to first increment this variable will boot the other cores */
atomic_t hart_lottery;
unsigned long boot_cpu_hartid;
diff --git a/arch/riscv/mm/Makefile b/arch/riscv/mm/Makefile
index eb22ab49b3e0..7307609d405b 100644
--- a/arch/riscv/mm/Makefile
+++ b/arch/riscv/mm/Makefile
@@ -3,3 +3,5 @@ obj-y += fault.o
obj-y += extable.o
obj-y += ioremap.o
obj-y += cacheflush.o
+
+CFLAGS_init.o := -mcmodel=medany
diff --git a/arch/riscv/mm/init.c b/arch/riscv/mm/init.c
index b379a75ac6a6..1e7a106fb583 100644
--- a/arch/riscv/mm/init.c
+++ b/arch/riscv/mm/init.c
@@ -25,6 +25,15 @@
#include <asm/pgtable.h>
#include <asm/io.h>
+unsigned long va_pa_offset;
+EXPORT_SYMBOL(va_pa_offset);
+unsigned long pfn_base;
+EXPORT_SYMBOL(pfn_base);
+
+unsigned long empty_zero_page[PAGE_SIZE / sizeof(unsigned long)]
+ __page_aligned_bss;
+EXPORT_SYMBOL(empty_zero_page);
+
static void __init zone_sizes_init(void)
{
unsigned long max_zone_pfns[MAX_NR_ZONES] = { 0, };
--
2.17.1
On Sun, Mar 24, 2019 at 11:08:29AM +0000, Anup Patel wrote:
> Signed-off-by: Christoph Hellwig <[email protected]>
> Signed-off-by: Mike Rapoport <[email protected]>
Huh? I don't think this is from Mike and me.
> +unsigned long va_pa_offset;
> +EXPORT_SYMBOL(va_pa_offset);
> +unsigned long pfn_base;
> +EXPORT_SYMBOL(pfn_base);
Can you move va_pa_offset and pfn_base next to the swapper_pg_dir
declaration? Thats where I move them in my current nommu patchset,
as that allows to just have a single CONFIG_NOMMU in mm/init.c.
Otherwise this looks fine to me.
On Sun, Mar 24, 2019 at 03:57:08PM +0100, Christoph Hellwig wrote:
> On Sun, Mar 24, 2019 at 11:08:29AM +0000, Anup Patel wrote:
> > Signed-off-by: Christoph Hellwig <[email protected]>
> > Signed-off-by: Mike Rapoport <[email protected]>
>
> Huh? I don't think this is from Mike and me.
Suggested-by: would me more appropriate indeed :)
> > +unsigned long va_pa_offset;
> > +EXPORT_SYMBOL(va_pa_offset);
> > +unsigned long pfn_base;
> > +EXPORT_SYMBOL(pfn_base);
>
> Can you move va_pa_offset and pfn_base next to the swapper_pg_dir
> declaration? Thats where I move them in my current nommu patchset,
> as that allows to just have a single CONFIG_NOMMU in mm/init.c.
>
> Otherwise this looks fine to me.
>
--
Sincerely yours,
Mike.
> -----Original Message-----
> From: linux-riscv <[email protected]> On Behalf Of Anup
> Patel
> Sent: Sunday, March 24, 2019 11:08
> To: Palmer Dabbelt <[email protected]>; Albert Ou <[email protected]>
> Cc: Anup Patel <[email protected]>; [email protected]; Mike
> Rapoport <[email protected]>; Christoph Hellwig <[email protected]>; Atish
> Patra <[email protected]>; Paul Walmsley <[email protected]>;
> [email protected]; Christoph Hellwig <[email protected]>
> Subject: [PATCH] RISC-V: Always compile mm/init.c with cmodel=medany
>
> diff --git a/arch/riscv/mm/Makefile b/arch/riscv/mm/Makefile
> index eb22ab49b3e0..7307609d405b 100644
> --- a/arch/riscv/mm/Makefile
> +++ b/arch/riscv/mm/Makefile
> @@ -3,3 +3,5 @@ obj-y += fault.o
> obj-y += extable.o
> obj-y += ioremap.o
> obj-y += cacheflush.o
> +
> +CFLAGS_init.o := -mcmodel=medany
I think you should also remove CFLAGS_setup.o from kernel/Makefile, as I think it was there solely for the correctness of setup_vm. As setup_vm is now in mm/init.o, the old CFLAGS probably isn't needed and should be removed in the same patch. Disclaimer: I haven't tested it.
On Sun, Mar 24, 2019 at 10:13 PM Mike Rapoport <[email protected]> wrote:
>
> On Sun, Mar 24, 2019 at 03:57:08PM +0100, Christoph Hellwig wrote:
> > On Sun, Mar 24, 2019 at 11:08:29AM +0000, Anup Patel wrote:
> > > Signed-off-by: Christoph Hellwig <[email protected]>
> > > Signed-off-by: Mike Rapoport <[email protected]>
> >
> > Huh? I don't think this is from Mike and me.
>
> Suggested-by: would me more appropriate indeed :)
Sure, I will use Suggested-by: here.
Regards,
Anup
On Sun, Mar 24, 2019 at 8:27 PM Christoph Hellwig <[email protected]> wrote:
>
> On Sun, Mar 24, 2019 at 11:08:29AM +0000, Anup Patel wrote:
> > Signed-off-by: Christoph Hellwig <[email protected]>
> > Signed-off-by: Mike Rapoport <[email protected]>
>
> Huh? I don't think this is from Mike and me.
>
> > +unsigned long va_pa_offset;
> > +EXPORT_SYMBOL(va_pa_offset);
> > +unsigned long pfn_base;
> > +EXPORT_SYMBOL(pfn_base);
>
> Can you move va_pa_offset and pfn_base next to the swapper_pg_dir
> declaration? Thats where I move them in my current nommu patchset,
> as that allows to just have a single CONFIG_NOMMU in mm/init.c.
>
> Otherwise this looks fine to me.
Sure, no problem. I will send update and send v2.
Regards,
Anup
On Mon, Mar 25, 2019 at 12:44 AM Gary Guo <[email protected]> wrote:
>
>
>
> > -----Original Message-----
> > From: linux-riscv <[email protected]> On Behalf Of Anup
> > Patel
> > Sent: Sunday, March 24, 2019 11:08
> > To: Palmer Dabbelt <[email protected]>; Albert Ou <[email protected]>
> > Cc: Anup Patel <[email protected]>; [email protected]; Mike
> > Rapoport <[email protected]>; Christoph Hellwig <[email protected]>; Atish
> > Patra <[email protected]>; Paul Walmsley <[email protected]>;
> > [email protected]; Christoph Hellwig <[email protected]>
> > Subject: [PATCH] RISC-V: Always compile mm/init.c with cmodel=medany
> >
> > diff --git a/arch/riscv/mm/Makefile b/arch/riscv/mm/Makefile
> > index eb22ab49b3e0..7307609d405b 100644
> > --- a/arch/riscv/mm/Makefile
> > +++ b/arch/riscv/mm/Makefile
> > @@ -3,3 +3,5 @@ obj-y += fault.o
> > obj-y += extable.o
> > obj-y += ioremap.o
> > obj-y += cacheflush.o
> > +
> > +CFLAGS_init.o := -mcmodel=medany
>
> I think you should also remove CFLAGS_setup.o from kernel/Makefile, as I think it was there solely for the correctness of setup_vm. As setup_vm is now in mm/init.o, the old CFLAGS probably isn't needed and should be removed in the same patch. Disclaimer: I haven't tested it.
Sure, I will remove it in v2 patch. Thanks for catching.
Regards,
Anup