linux-kernel - Re: [PATCH] MIPS: mm: tlb-r4k: Uniquify TLB entries on init

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] [day] [month] [year] [list]

Message-Id: <40ff6065-2f2b-4a07-b247-02e11fd46eb8@app.fastmail.com>
Date: Sat, 07 Jun 2025 11:36:52 +0100
From: "Jiaxun Yang" <jiaxun.yang@...goat.com>
To: "Maciej W. Rozycki" <macro@...am.me.uk>
Cc: "Thomas Bogendoerfer" <tsbogend@...ha.franken.de>,
 "linux-mips@...r.kernel.org" <linux-mips@...r.kernel.org>,
 linux-kernel@...r.kernel.org
Subject: Re: [PATCH] MIPS: mm: tlb-r4k: Uniquify TLB entries on init

在2025年6月7日周六 上午3:48，Maciej W. Rozycki写道：
> On Thu, 5 Jun 2025, Jiaxun Yang wrote:
>
>> +static unsigned long r4k_safe_entryhi(void)
>> +{
>> +	int entry = current_cpu_data.tlbsize;
>> +	int old_index;
>> +
>> +	old_index = read_c0_index();
>> +	while (entry >= 0) {
> [...]
>> +		entry++;
>> +	}
>
>  Hmm, how is it supposed to work: you start from say 48 and then iterate 
> until 0x80000000 before giving up?  Also a signed overflow condition is UB 
> pre-C23, so the compiler may well optimise the loop control away.

Ah sorry I didn't realise it's a UB. However, this loop is likely to be
terminated very early as it should be fairly easily to find a free EntryHi
value with TLBP.

>
>  How did you verify this code?

Boot tested on Boston as well as M5150 simulation. I tried to manually setup a
TLB entry collision situation with Boot-MIPS.

>
>   Maciej

Thanks

-- 
- Jiaxun