Re: Strange EFAULT on mips64el returned by syscall when another thread is forking

From: "Jiaxun Yang" <jiaxun.yang@flygoat.com>
To: "Linus Torvalds" <torvalds@linux-foundation.org>,
	"Xi Ruoyao" <xry111@xry111.site>,
	tsbogend@alpha.franken.de
Cc: "Andreas Schwab" <schwab@suse.de>,
	"Ben Hutchings" <ben@decadent.org.uk>,
	"linux-mips@vger.kernel.org" <linux-mips@vger.kernel.org>,
	linux-kernel@vger.kernel.org, libc-alpha@sourceware.org
Subject: Re: Strange EFAULT on mips64el returned by syscall when another thread is forking
Date: Thu, 25 Jan 2024 09:28:41 +0000	[thread overview]
Message-ID: <00d1b813-c55f-4365-8d81-d70258e10b16@app.fastmail.com> (raw)
In-Reply-To: <CAHk-=whkEXGOCEZFO2vAZ9rDd8uW8MJwFNYg9KXaC_vZVso6iA@mail.gmail.com>

在2024年1月24日一月 下午10:10，Linus Torvalds写道：
[...]
>
> Anyway, I'm pretty sure this is the bug, now some MIPS person just
> needs to fix the MIPS version of "instruction_pointer()" to do what
> "exception_epc()" already does.

Hi folks,

Kinda MIPS person here, I looked into the problem, and it's not that
easy to fix.

I inspected some existing usage of "instruction_pointer()", and some
of them do want exception return address (which is always CP0_EPC).
Others like this case they want the precise exception instruction
pointer ("exception_epc()" for MIPS).

I'm planning to make "instruction_pointer()" always point to exception
instructions, and implemented a new function called "return_pc()"or
whatever for "exception_epc()". Do you have a better name in mind?

Besides isa16 stuff do require kernel to read user space fault
instruction to determine delay slot size... I don't think it's always
possible when "instruction_pointer()" is called. MIPS16/microMIPS
is rarely used today though.

Thanks

>
>               Linus

-- 
- Jiaxun