linux-kernel - Re: [vfs] 8bb3c61baf: vm-scalability.median -23.7% regression

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: <20190908234722.GE1131@ZenIV.linux.org.uk>
Date:   Mon, 9 Sep 2019 00:47:22 +0100
From:   Al Viro <viro@...iv.linux.org.uk>
To:     kernel test robot <rong.a.chen@...el.com>
Cc:     David Howells <dhowells@...hat.com>,
        Hugh Dickins <hughd@...gle.com>,
        LKML <linux-kernel@...r.kernel.org>,
        linux-fsdevel@...r.kernel.org, lkp@...org
Subject: Re: [vfs]  8bb3c61baf:  vm-scalability.median -23.7% regression

On Sun, Sep 08, 2019 at 10:46:01PM +0100, Al Viro wrote:
> On Tue, Sep 03, 2019 at 04:41:22PM +0800, kernel test robot wrote:
> > Greeting,
> > 
> > FYI, we noticed a -23.7% regression of vm-scalability.median due to commit:
> > 
> > 
> > commit: 8bb3c61bafa8c1cd222ada602bb94ff23119e738 ("vfs: Convert ramfs, shmem, tmpfs, devtmpfs, rootfs to use the new mount API")
> > https://kernel.googlesource.com/pub/scm/linux/kernel/git/viro/vfs.git work.mount
> > 
> > in testcase: vm-scalability
> > on test machine: 88 threads Intel(R) Xeon(R) CPU E5-2699 v4 @ 2.20GHz with 128G memory
> > with following parameters:
> > 
> > 	runtime: 300s
> > 	size: 16G
> > 	test: shm-pread-rand
> > 	cpufreq_governor: performance
> > 	ucode: 0xb000036
> 
> That thing loses size=... option.  Both size= and nr_blocks= affect the
> same thing (->max_blocks), but the parser keeps track of the options
> it has seen and applying the parsed data to superblock checks only
> whether nr_blocks= had been there.  IOW, size= gets parsed, but the
> result goes nowhere.
> 
> I'm not sure whether it's better to fix the patch up or redo it from
> scratch - it needs to be carved up anyway and it's highly non-transparent,
> so I'm probably going to replace the damn thing entirely with something
> that would be easier to follow.

... and this
+       { Opt_huge,     "deny",         SHMEM_HUGE_DENY },
+       { Opt_huge,     "force",        SHMEM_HUGE_FORCE },
had been wrong - huge=deny and huge=force should not be accepted _and_
fs_parameter_enum is not suitable for negative constants right now
anyway.