[PATCH v6 08/14] riscv: move boot alternatives to a slightly earlier position

[Heiko Stübner]

Am Freitag, 11. Februar 2022, 02:57:19 CET schrieb Atish Patra:
> On Thu, Feb 10, 2022 at 5:11 PM Heiko Stübner <heiko at sntech.de> wrote:
> >
> > Hi Atish,
> >
> > Am Donnerstag, 10. Februar 2022, 23:42:35 CET schrieb Atish Patra:
> > > On Wed, Feb 9, 2022 at 4:39 AM Heiko Stuebner <heiko at sntech.de> wrote:
> > > >
> > > > Move the application of boot alternatives to soc_early_init().
> > > > This allows to catch more generic cases of code needing patches
> > > > than doing it in smp_prepare_boot_cpu() and also makes it actually
> > > > work if CONFIG_SMP is disabled for whatever reason.
> > > >
> > > > The position is chosen mainly as it is before the actual soc early
> > > > init runs but also already allows accessing the devicetree
> > > > via fdt_* functions.
> > > >
> > > > Signed-off-by: Heiko Stuebner <heiko at sntech.de>
> > > > ---
> > > >  arch/riscv/kernel/head.S    | 2 ++
> > > >  arch/riscv/kernel/smpboot.c | 2 --
> > > >  2 files changed, 2 insertions(+), 2 deletions(-)
> > > >
> > > > diff --git a/arch/riscv/kernel/head.S b/arch/riscv/kernel/head.S
> > > > index 2363b43312fc..0e1bb97f9749 100644
> > > > --- a/arch/riscv/kernel/head.S
> > > > +++ b/arch/riscv/kernel/head.S
> > > > @@ -10,6 +10,7 @@
> > > >  #include <asm/thread_info.h>
> > > >  #include <asm/page.h>
> > > >  #include <asm/pgtable.h>
> > > > +#include <asm/alternative.h>
> > > >  #include <asm/csr.h>
> > > >  #include <asm/cpu_ops_sbi.h>
> > > >  #include <asm/hwcap.h>
> > > > @@ -341,6 +342,7 @@ clear_bss_done:
> > > >         call kasan_early_init
> > > >  #endif
> > > >         /* Start the kernel */
> > > > +       call apply_boot_alternatives
> > >
> > > Do you really need this early ?
> > > if non-smp configuration is the only option, Can you do it in
> > > setup_arch() after riscv_fill_hwcap() is called ?
> >
> > The issue I see is, we we have the soc_early_init [0] running
> > directly after this and the one user I see [1] already wants to
> > ioremap io-memory at this stage.
> >
> 
> Kendryte is always a special case. IIRC, ioremap is done so early
> so that it can use all 8MB of SRAM.
> 
> > So judging by the fact that more early-inits will get added
> > in the future I do guess we should've set up the io-memory
> > page-type by this point?
> >
> 
> I hope there won't be :). For any normal mmu capable SoC, I don't
> see why that would be required.

ok, so we'll assume there won't be another special-special case SoC
forthcoming and hope for the best - works for me :-D

Thanks
Heiko


> > [0] https://elixir.bootlin.com/linux/latest/source/arch/riscv/kernel/soc.c#L14
> > [1] https://elixir.bootlin.com/linux/latest/source/drivers/soc/canaan/k210-sysctl.c#L66
> >
> > > By doing that, we can unify the cpu feature probing and you don't need
> > > a separate DT parsing just for svpbmt.
> > >
> > > >         call soc_early_init
> > > >         tail start_kernel
> > > >
> > > > diff --git a/arch/riscv/kernel/smpboot.c b/arch/riscv/kernel/smpboot.c
> > > > index a6d13dca1403..f1e4948a4b52 100644
> > > > --- a/arch/riscv/kernel/smpboot.c
> > > > +++ b/arch/riscv/kernel/smpboot.c
> > > > @@ -32,7 +32,6 @@
> > > >  #include <asm/sections.h>
> > > >  #include <asm/sbi.h>
> > > >  #include <asm/smp.h>
> > > > -#include <asm/alternative.h>
> > > >
> > > >  #include "head.h"
> > > >
> > > > @@ -41,7 +40,6 @@ static DECLARE_COMPLETION(cpu_running);
> > > >  void __init smp_prepare_boot_cpu(void)
> > > >  {
> > > >         init_cpu_topology();
> > > > -       apply_boot_alternatives();
> > > >  }
> > > >
> > > >  void __init smp_prepare_cpus(unsigned int max_cpus)
> > > > --
> > > > 2.30.2
> > > >
> > > >
> > > > _______________________________________________
> > > > linux-riscv mailing list
> > > > linux-riscv at lists.infradead.org
> > > > http://lists.infradead.org/mailman/listinfo/linux-riscv
> > >
> > >
> > >
> > >
> >
> >
> >
> >
> 
> 
>