Re: [PATCH v4] sh: Restructure setup code to reserve memory regions earlier

From: John Paul Adrian Glaubitz
Date: Tue Jul 16 2024 - 06:42:45 EST


Hi Oreoluwa,

On Mon, 2024-07-15 at 17:12 -0700, Oreoluwa Babatunde wrote:
> On 7/13/2024 12:58 AM, John Paul Adrian Glaubitz wrote:
>
> Hi Adrian,
>
> > > diff --git a/arch/sh/include/asm/setup.h b/arch/sh/include/asm/setup.h
> > > index 84bb23a771f3..f8b814fb1c7f 100644
> > > --- a/arch/sh/include/asm/setup.h
> > > +++ b/arch/sh/include/asm/setup.h
> > > @@ -19,7 +19,6 @@
> > > #define COMMAND_LINE ((char *) (PARAM+0x100))
> > >
> > > void sh_mv_setup(void);
> > > -void check_for_initrd(void);
> > > void per_cpu_trap_init(void);
> > > void sh_fdt_init(phys_addr_t dt_phys);
> > >
> > > diff --git a/arch/sh/kernel/setup.c b/arch/sh/kernel/setup.c
> > > index 620e5cf8ae1e..8477491f4ffd 100644
> > > --- a/arch/sh/kernel/setup.c
> > > +++ b/arch/sh/kernel/setup.c
> > > @@ -35,6 +35,7 @@
> > > #include <asm/io.h>
> > > #include <asm/page.h>
> > > #include <asm/elf.h>
> > > +#include <asm/kexec.h>
> > > #include <asm/sections.h>
> > > #include <asm/irq.h>
> > > #include <asm/setup.h>
> > > @@ -114,7 +115,7 @@ static int __init early_parse_mem(char *p)
> > > }
> > > early_param("mem", early_parse_mem);
> > >
> > > -void __init check_for_initrd(void)
> > > +static void __init check_for_initrd(void)
> > > {
> > > #ifdef CONFIG_BLK_DEV_INITRD
> > > unsigned long start, end;
> > > @@ -172,6 +173,42 @@ void __init check_for_initrd(void)
> > > #endif
> > > }
> > Making check_for_initrd() static seems like an unrelated change to me or am
> > I missing something? If yes, it should go into a separate patch.
> ack.
> > > +static void __init early_reserve_mem(void)
> > > +{
> > > + unsigned long start_pfn;
> > > + u32 zero_base = (u32)__MEMORY_START + (u32)PHYSICAL_OFFSET;
> > > + u32 start = zero_base + (u32)CONFIG_ZERO_PAGE_OFFSET;
> > > +
> > > + /*
> > > + * Partially used pages are not usable - thus
> > > + * we are rounding upwards:
> > > + */
> > > + start_pfn = PFN_UP(__pa(_end));
> > > +
> > > + /*
> > > + * Reserve the kernel text and Reserve the bootmem bitmap. We do
> > > + * this in two steps (first step was init_bootmem()), because
> > > + * this catches the (definitely buggy) case of us accidentally
> > > + * initializing the bootmem allocator with an invalid RAM area.
> > > + */
> > > + memblock_reserve(start, (PFN_PHYS(start_pfn) + PAGE_SIZE - 1) - start);
> > > +
> > > + /*
> > > + * Reserve physical pages below CONFIG_ZERO_PAGE_OFFSET.
> > > + */
> > > + if (CONFIG_ZERO_PAGE_OFFSET != 0)
> > > + memblock_reserve(zero_base, CONFIG_ZERO_PAGE_OFFSET);
> > > +
> > > + /*
> > > + * Handle additional early reservations
> > > + */
> > > + check_for_initrd();
> > > + reserve_crashkernel();
> > > +
> > > + if (sh_mv.mv_mem_reserve)
> > > + sh_mv.mv_mem_reserve();
> > > +}
> > > +
> > > #ifndef CONFIG_GENERIC_CALIBRATE_DELAY
> > > void calibrate_delay(void)
> > > {
> > I'm not really happy with moving early_reserve_mem() from mm/init.c to
> > kernel/setup.c. Can't we just leave it where it is while still keeping
> > the changes to paging_init()?
> ack.
> >
> > > @@ -319,9 +356,14 @@ void __init setup_arch(char **cmdline_p)
> > >
> > > sh_mv_setup();
> > >
> > > + sh_mv.mv_mem_init();
> > > +
> > > /* Let earlyprintk output early console messages */
> > > sh_early_platform_driver_probe("earlyprintk", 1, 1);
> > >
> > > + /* set aside reserved memory regions */
> > > + early_reserve_mem();
> > > +
> > > #ifdef CONFIG_OF_EARLY_FLATTREE
> > > #ifdef CONFIG_USE_BUILTIN_DTB
> > > unflatten_and_copy_device_tree();
>
> I'll make adjustments based on your comments and
> resend another version.

Okay, I will wait with my pull request to Linus a few more days then.

Thanks so much for being super patient with me. It took me way too long
to test and review your patch, but I hope in the end we'll get the best
possible version merged.

Adrian

--
.''`. John Paul Adrian Glaubitz
: :' : Debian Developer
`. `' Physicist
`- GPG: 62FF 8A75 84E0 2956 9546 0006 7426 3B37 F5B5 F913