[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20181105180004.GA4923@bombadil.infradead.org>
Date: Mon, 5 Nov 2018 10:00:04 -0800
From: Matthew Wilcox <willy@...radead.org>
To: Linus Torvalds <torvalds@...ux-foundation.org>
Cc: rong.a.chen@...el.com,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
lkp@...org
Subject: Re: [LKP] [xarray] 0e9446c35a:
WARNING:at_include/linux/xarray.h:#xa_mk_value
On Mon, Nov 05, 2018 at 09:57:47AM -0800, Linus Torvalds wrote:
> Willy,
>
> On Sun, Nov 4, 2018 at 11:03 PM kernel test robot <rong.a.chen@...el.com> wrote:
> >
> > commit: 0e9446c35a80931044b6d8d2d74a9cabd248539f ("xarray: Add range store functionality")
> ...
> > [ 11.880031] WARNING: CPU: 0 PID: 1 at include/linux/xarray.h:54 xa_mk_value+0x7/0x10
> > [ 11.881944] EIP: xa_mk_value+0x7/0x10
> > [ 11.883565] Call Trace:
> > [ 11.883565] __check_store_range+0xe/0x171
> > [ 11.883565] check_store_range+0x5d/0x87
>
> If I read this right, it's simply this case:
>
> __check_store_range(xa, UINT_MAX + i, UINT_MAX + j);
>
> which turns into negative values on a 32-bit machine, and then
> xa_mk_value() warns.
>
> Should that part of the test be surrounded by #ifdef CONFIG_64BIT? Or
> did you _want_ the warnings?
I should have just made that (1 << 24) instead of UINT_MAX. I had forgotten
about 32-bit machines while writing that test.
That prompted me to actually build the test suite on 32-bit and there's
a few other things in there that want fixing. I'll have a pull request
for you later today.
Powered by blists - more mailing lists