Date:   Sun, 17 Jan 2021 19:55:20 +0100
From:   Geert Uytterhoeven <geert@...ux-m68k.org>
To:     Atish Patra <atishp@...shpatra.org>
Cc:     Palmer Dabbelt <palmer@...belt.com>,
        Atish Patra <Atish.Patra@....com>,
        Albert Ou <aou@...s.berkeley.edu>,
        Anup Patel <Anup.Patel@....com>,
        "linux-kernel@...r.kernel.org List" <linux-kernel@...r.kernel.org>,
        linux-riscv <linux-riscv@...ts.infradead.org>,
        Paul Walmsley <paul.walmsley@...ive.com>,
        Nick Kossifidis <mick@....forth.gr>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Ard Biesheuvel <ardb@...nel.org>,
        Mike Rapoport <rppt@...nel.org>
Subject: Re: [PATCH 3/4] RISC-V: Fix L1_CACHE_BYTES for RV32

Hi Atish,

On Sat, Jan 16, 2021 at 2:39 AM Atish Patra <atishp@...shpatra.org> wrote:
> On Thu, Jan 14, 2021 at 11:59 PM Geert Uytterhoeven
> <geert@...ux-m68k.org> wrote:
> > On Thu, Jan 14, 2021 at 10:11 PM Atish Patra <atishp@...shpatra.org> wrote:
> > > On Thu, Jan 14, 2021 at 11:46 AM Palmer Dabbelt <palmer@...belt.com> wrote:
> > > > On Thu, 14 Jan 2021 10:33:01 PST (-0800), atishp@...shpatra.org wrote:
> > > > > On Wed, Jan 13, 2021 at 9:10 PM Palmer Dabbelt <palmer@...belt.com> wrote:
> > > > >>
> > > > >> On Thu, 07 Jan 2021 01:26:51 PST (-0800), Atish Patra wrote:
> > > > >> > SMP_CACHE_BYTES/L1_CACHE_BYTES should be defined as 32 instead of
> > > > >> > 64 for RV32. Otherwise, there will be hole of 32 bytes with each memblock
> > > > >> > allocation if it is requested to be aligned with SMP_CACHE_BYTES.
> > > > >> >
> > > > >> > Signed-off-by: Atish Patra <atish.patra@....com>
> > > > >> > ---
> > > > >> >  arch/riscv/include/asm/cache.h | 4 ++++
> > > > >> >  1 file changed, 4 insertions(+)
> > > > >> >
> > > > >> > diff --git a/arch/riscv/include/asm/cache.h b/arch/riscv/include/asm/cache.h
> > > > >> > index 9b58b104559e..c9c669ea2fe6 100644
> > > > >> > --- a/arch/riscv/include/asm/cache.h
> > > > >> > +++ b/arch/riscv/include/asm/cache.h
> > > > >> > @@ -7,7 +7,11 @@
> > > > >> >  #ifndef _ASM_RISCV_CACHE_H
> > > > >> >  #define _ASM_RISCV_CACHE_H
> > > > >> >
> > > > >> > +#ifdef CONFIG_64BIT
> > > > >> >  #define L1_CACHE_SHIFT               6
> > > > >> > +#else
> > > > >> > +#define L1_CACHE_SHIFT               5
> > > > >> > +#endif
> > > > >> >
> > > > >> >  #define L1_CACHE_BYTES               (1 << L1_CACHE_SHIFT)
> > > > >>
> > > > >> Should we not instead just
> > > > >>
> > > > >> #define SMP_CACHE_BYTES L1_CACHE_BYTES
> > > > >>
> > > > >> like a handful of architectures do?
> > > > >>
> > > > >
> > > > > The generic code already defines it that way in include/linux/cache.h
> > > > >
> > > > >> The cache size is sort of fake here, as we don't have any non-coherent
> > > > >> mechanisms, but IIRC we wrote somewhere that it's recommended to have 64-byte
> > > > >> cache lines in RISC-V implementations as software may assume that for
> > > > >> performance reasons.  Not really a strong reason, but I'd prefer to just make
> > > > >> these match.
> > > > >>
> > > > >
> > > > > If it is documented somewhere in the kernel, we should update that. I
> > > > > think SMP_CACHE_BYTES being 64
> > > > > actually degrades the performance as there will be a fragmented memory
> > > > > blocks with 32 bit bytes gap wherever
> > > > > SMP_CACHE_BYTES is used as an alignment requirement.
> > > >
> > > > I don't buy that: if you're trying to align to the cache size then the gaps are
> > > > the whole point.  IIUC the 64-byte cache lines come from DDR, not XLEN, so
> > > > there's really no reason for these to be different between the base ISAs.
> > > >
> > >
> > > Got your point. I noticed this when fixing the resource tree issue
> > > where the SMP_CACHE_BYTES
> > > alignment was not intentional but causing the issue. The real issue
> > > was solved via another patch in this series though.
> > >
> > > Just to clarify, if the allocation function intends to allocate
> > > consecutive memory, it should use 32 instead of SMP_CACHE_BYTES.
> > > This will lead to a #ifdef macro in the code.
> > >
> > > > > In addition to that, Geert Uytterhoeven mentioned some panic on vex32
> > > > > without this patch.
> > > > > I didn't see anything in Qemu though.
> > > >
> > > > Something like that is probably only going to show up on real hardware, QEMU
> > > > doesn't really do anything with the cache line size.  That said, as there's
> > > > nothing in our kernel now related to non-coherent memory there really should
> > > > only be performance issue (at least until we have non-coherent systems).
> > > >
> > > > I'd bet that the change is just masking some other bug, either in the software
> > > > or the hardware.  I'd prefer to root cause this rather than just working around
> > > > it, as it'll probably come back later and in a more difficult way to find.
> > > >
> > >
> > > Agreed. @Geert Uytterhoeven Can you do a further analysis of the panic
> > > you were saying ?
> > > We may need to change an alignment requirement to 32 for RV32 manually
> > > at some place in code.
> >
> > My findings were in
> > https://lore.kernel.org/linux-riscv/CAMuHMdWf6K-5y02+WJ6Khu1cD6P0n5x1wYQikrECkuNtAA1pgg@mail.gmail.com/
> >
> > Note that when the memblock.reserved list kept increasing, it kept on
> > adding the same entry to the list.  But that was fixed by "[PATCH 1/4]
> > RISC-V: Do not allocate memblock while iterating reserved memblocks".
> >
> > After that, only the (reproducible) "Unable to handle kernel paging
> > request at virtual address 61636473" was left, always at the same place.
> > No idea where the actual corruption happened.
> >
>
> Yes. I was asking about this panic. I don't have the litex fpga to
> reproduce this as well.
> Can you take a look at the epc & ra to figure out where exactly is the fault ?
>
> That will help to understand the real cause for this panic.

[...]
Freeing initrd memory: 8192K
workingset: timestamp_bits=30 max_order=15 bucket_order=0
Block layer SCSI generic (bsg) driver version 0.4 loaded (major 253)
io scheduler mq-deadline registered
io scheduler kyber registered
Unable to handle kernel paging request at virtual address 61636473
Oops [#1]
CPU: 0 PID: 1 Comm: swapper/0 Not tainted
5.11.0-rc3-orangecrab-00068-g267ecb2e2e9d-dirty #37
epc: c000f8e4 ra : c00110e8 sp : c082bc70
 gp : c0665948 tp : c0830000 t0 : c08ba500
 t1 : 00000002 t2 : 00000000 s0 : c082bc80
 s1 : 00000000 a0 : c05e2dec a1 : c08ba4e0
 a2 : c0665d38 a3 : 61636473 a4 : f0004003
 a5 : f0004000 a6 : c7efffb8 a7 : c08ba4e0
 s2 : 01001f00 s3 : c0666000 s4 : c05e2e0c
 s5 : c0666000 s6 : 80000000 s7 : 00000006
 s8 : c05a4000 s9 : c08ba4e0 s10: c05e2dec
 s11: 00000000 t3 : c08ba500 t4 : 00000001
 t5 : 00076400 t6 : c08bbb5e
status: 00000120 badaddr: 61636473 cause: 0000000d
---[ end trace 50524759df172195 ]---
Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
---[ end Kernel panic - not syncing: Attempted to kill init!
exitcode=0x0000000b ]---

Looking up epc and ra in System.map, closest symbols are:

    c000f8b0 t __request_resource
    c0010ff4 T __request_region

The above is with a kernel built from my own config, but using
litex_vexriscv_defconfig with https://github.com/geertu/linux branch
litex-v5.11 and commit 718aaf7d1c351035 ("RISC-V: Fix L1_CACHE_BYTES for
RV32") reverted gives the exact same results.

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@...ux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Powered by blists - more mailing lists