linux-kernel - Re: [PATCH v2 10/10] RISC-V: provide a Kconfig option to disable parsing "riscv,isa"

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20230630-urgency-emission-882cd52a7a27@wendy>
Date:   Fri, 30 Jun 2023 14:19:46 +0100
From:   Conor Dooley <conor.dooley@...rochip.com>
To:     Andrew Jones <ajones@...tanamicro.com>
CC:     Conor Dooley <conor@...nel.org>,
        Palmer Dabbelt <palmer@...osinc.com>, <robh+dt@...nel.org>,
        <krzysztof.kozlowski+dt@...aro.org>,
        Paul Walmsley <paul.walmsley@...ive.com>,
        <aou@...s.berkeley.edu>, <heiko.stuebner@...ll.eu>,
        Evan Green <evan@...osinc.com>, <sunilvl@...tanamicro.com>,
        <linux-riscv@...ts.infradead.org>, <devicetree@...r.kernel.org>,
        <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2 10/10] RISC-V: provide a Kconfig option to disable
 parsing "riscv,isa"

On Fri, Jun 30, 2023 at 09:46:48AM +0200, Andrew Jones wrote:
> On Thu, Jun 29, 2023 at 10:44:18PM +0100, Conor Dooley wrote:
> > On Thu, Jun 29, 2023 at 02:16:49PM -0700, Palmer Dabbelt wrote:
> > > On Thu, 29 Jun 2023 13:20:55 PDT (-0700), Conor Dooley wrote:
> ...
> > > > +bool __initdata riscv_isa_fallback_cmdline = false;
> > > > +static int __init riscv_isa_fallback_setup(char *__unused)
> > > 
> > > Maybe it's better to support =true and =false here?  Not sure it matters,
> > > we're already down a rabbit hole ;)
> > 
> > Dunno, not implemented a cmdline param before. Seemed "cleaner" to check
> > for presence, don't really care so I'll adapt to w/e.
> >
> 
> I don't have a strong preference here, but to throw in more food for
> thought, I see this DT-v1 vs. DT-v2 choice to be a bit analogous to the
> DT vs. ACPI choice. The 'acpi' command line parameter, for RISC-V, can
> be 'off', 'on', and 'force', where
> 
>   off -- disable ACPI if default was on
>   on -- enable ACPI but allow fallback to DT
>   force -- enable ACPI if default was off
> 
> So, if the default of the isa fallback command line option will depend on
> Kconfig, then we may also want a 'force'.

I'm not sure that I understand what "force" would give us.
There's 4 cases:
- CONFIG_RISCV_ISA_FALLBACK is enabled, cmdline option is present:
  cmdline option is ignored, fallback is taken if needed.
  crash if neither are present.

- CONFIG_RISCV_ISA_FALLBACK is enabled, cmdline option is not present:
  cmdline option is ignored, fallback is taken if needed.
  crash if neither are present.

- CONFIG_RISCV_ISA_FALLBACK is disabled, cmdline option is present:
  cmdline option takes priority, fallback is taken if needed.
  crash if neither are present.

- CONFIG_RISCV_ISA_FALLBACK is disabled, cmdline option is not present:
  fallback is never taken
  crash if new properties aren't present.

I don't really see the value in having an equivalent to acpi=off,
because the order of precedence is, to use your naming, "DT-v2" falling
back to "DT-v1" & the default value concerns the use of the fallback.
For ACPI, it is the other way around & the option controls the use of
"DT-v2"'s analogue. Trying to slot in that logic:

- CONFIG_RISCV_ISA_FALLBACK is enabled, cmdline option "=on":
  cmdline option is ignored, fallback is taken if needed.
  crash if neither are present.

- CONFIG_RISCV_ISA_FALLBACK is enabled, cmdline option "=off":
  cmdline option is prioritised, fallback is taken if needed.
  crash if new properties aren't present.

- CONFIG_RISCV_ISA_FALLBACK is disabled, cmdline option "=on":
  cmdline option is prioritised, fallback is taken if needed.
  crash if neither are present.

- CONFIG_RISCV_ISA_FALLBACK is disabled, cmdline option "=off":
  fallback is never taken
  crash if new properties aren't present.

I think I prefer the behaviour of what I currently have & I don't really
get where the "force" option is supposed to fit in either?

Cheers,
Conor.

Download attachment "signature.asc" of type "application/pgp-signature" (229 bytes)