linux-kernel - Re: [PATCH] riscv: Ensure isa-ext static keys are writable

Open Source and information security mailing list archives

Message-ID: <20220817153150.fvht2isdiqjpk3my@kamzik>
Date:   Wed, 17 Aug 2022 17:31:50 +0200
From:   Andrew Jones <ajones@...tanamicro.com>
To:     Jisheng Zhang <jszhang@...nel.org>
Cc:     Palmer Dabbelt <palmer@...belt.com>, atishp@...shpatra.org,
        linux-riscv@...ts.infradead.org, kvm-riscv@...ts.infradead.org,
        linux-kernel@...r.kernel.org, daolu@...osinc.com,
        Conor.Dooley@...rochip.com, re@...z.net, apatel@...tanamicro.com
Subject: Re: [PATCH] riscv: Ensure isa-ext static keys are writable

On Wed, Aug 17, 2022 at 10:30:07PM +0800, Jisheng Zhang wrote:
> On Tue, Aug 16, 2022 at 09:29:36PM -0700, Palmer Dabbelt wrote:
> > On Tue, 16 Aug 2022 16:41:46 PDT (-0700), atishp@...shpatra.org wrote:
> > > On Tue, Aug 16, 2022 at 9:31 AM Andrew Jones <ajones@...tanamicro.com> wrote:
> > > > 
> > > > riscv_isa_ext_keys[] is an array of static keys used in the unified
> > > > ISA extension framework. The keys added to this array may be used
> > > > anywhere, including in modules. Ensure the keys remain writable by
> > > > placing them in the data section.
> > > > 
> > > > The need to change riscv_isa_ext_keys[]'s section was found when the
> > > > kvm module started failing to load. Commit 8eb060e10185 ("arch/riscv:
> > > > add Zihintpause support") adds a static branch check for a newly
> > > > added isa-ext key to cpu_relax(), which kvm uses.
> > >> > > 
> > > > Fixes: c360cbec3511 ("riscv: introduce unified static key mechanism for ISA extensions")
> > > > Signed-off-by: Andrew Jones <ajones@...tanamicro.com>
> > > > > > ---
> > > >  arch/riscv/kernel/cpufeature.c | 2 +-
> > > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > > 
> > > > diff --git a/arch/riscv/kernel/cpufeature.c b/arch/riscv/kernel/cpufeature.c
> > > > index 553d755483ed..3b5583db9d80 100644
> > > > --- a/arch/riscv/kernel/cpufeature.c
> > > > +++ b/arch/riscv/kernel/cpufeature.c
> > > > @@ -28,7 +28,7 @@ unsigned long elf_hwcap __read_mostly;
> > > >  /* Host ISA bitmap */
> > > >  static DECLARE_BITMAP(riscv_isa, RISCV_ISA_EXT_MAX) __read_mostly;
> > > > 
> > > > -__ro_after_init DEFINE_STATIC_KEY_ARRAY_FALSE(riscv_isa_ext_keys, RISCV_ISA_EXT_KEY_MAX);
> > > > +DEFINE_STATIC_KEY_ARRAY_FALSE(riscv_isa_ext_keys, RISCV_ISA_EXT_KEY_MAX);
> > > >  EXPORT_SYMBOL(riscv_isa_ext_keys);
> > > > 
> > > >  /**
> > > > --
> > > > 2.37.1
> > > > 
> > > > 
> > > > --
> > > > kvm-riscv mailing list
> > > > kvm-riscv@...ts.infradead.org
> > > > http://lists.infradead.org/mailman/listinfo/kvm-riscv
> > > 
> > > Thanks for the quick fix. Tested with kvm guests booting in Qemu.
> > > 
> > > Tested-by: Atish Patra <atishp@...osinc.com>
> 
> Reviewed-by: Jisheng Zhang <jszhang@...nel.org>
> 
> Thanks for catching the bug in my code.
> 
> > 
> > Thanks, I hadn't realized how static keys work but looks like having them as
> > __ro_after_init was always bogus.  Sorry to break stuff, looks like I should be
> 
> Sorry for breaking the static isa extension key usage in modules.
> It's a good idea to make them readonly, but need to cope with the usage
> in module properly. This may be a good improvement item.

While some static keys do use __ro_after_init, they're almost all locally
defined (static __ro_after_init). I see a few instances of __read_mostly
global static key definitions, though. That may make sense for
riscv_isa_ext_keys[] as well.

Thanks,
drew

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