Re: [PATCH] riscv: Enable interrupt during exception handling

From: "Clément Léger" <cleger@rivosinc.com>
To: "Maciej W. Rozycki" <macro@orcam.me.uk>,
	Palmer Dabbelt <palmer@dabbelt.com>
Cc: namcao@linutronix.de, Paul Walmsley <paul.walmsley@sifive.com>,
	aou@eecs.berkeley.edu, Alexandre Ghiti <alex@ghiti.fr>,
	bigeasy@linutronix.de, clrkwllms@kernel.org, rostedt@goodmis.org,
	linux-rt-devel@lists.linux.dev, linux-riscv@lists.infradead.org,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH] riscv: Enable interrupt during exception handling
Date: Tue, 24 Jun 2025 13:37:13 +0200	[thread overview]
Message-ID: <19cf2371-7293-4e71-a17d-669db4c8270a@rivosinc.com> (raw)
In-Reply-To: <alpine.DEB.2.21.2506240303170.61655@angie.orcam.me.uk>

On 24/06/2025 04:09, Maciej W. Rozycki wrote:
> On Mon, 23 Jun 2025, Palmer Dabbelt wrote:
> 
>> I'm kind of split on a Fixes tag here.  One could argue it's a regression, as
>> having interrupts disabled during exceptions is going to cause all sorts of
>> performance issues for users.  Seems a bit risk to just backport, though...
>>
>> That said, if nobody noticed then it's probably a good sign nobody is really
>> paying attention and we should just backport it before anyone notices...
> 
>  Oh, someone did notice and it's not only performance, cf. 
> <https://lore.kernel.org/r/alpine.DEB.2.21.2501070143250.18889@angie.orcam.me.uk/>.

I also had a series which was doing so for misaligned accesses handling,
but after discussion, it was not ok to do so.:

https://lore.kernel.org/linux-riscv/20250422094419.GC14170@noisy.programming.kicks-ass.net/

So this series should probably be modified to only reenable interrupts
when explicitly needed, ie page faults and for force_sig_fault().

Thanks,

Clément

> 
>   Maciej
> 
> _______________________________________________
> linux-riscv mailing list
> linux-riscv@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-riscv