[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <86802c440806280330s6c8b1cb6o977290b535dc2343@mail.gmail.com>
Date: Sat, 28 Jun 2008 03:30:39 -0700
From: "Yinghai Lu" <yhlu.kernel@...il.com>
To: "Ingo Molnar" <mingo@...e.hu>
Cc: "Thomas Gleixner" <tglx@...utronix.de>,
"H. Peter Anvin" <hpa@...or.com>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] x86: fix init_memory_mapping over boundary v2
On Sat, Jun 28, 2008 at 12:47 AM, Ingo Molnar <mingo@...e.hu> wrote:
>
> * Yinghai Lu <yhlu.kernel@...il.com> wrote:
>
>> On Sat, Jun 28, 2008 at 12:02 AM, Ingo Molnar <mingo@...e.hu> wrote:
>> >
>> > * Ingo Molnar <mingo@...e.hu> wrote:
>> >
>> >> * Yinghai Lu <yhlu.kernel@...il.com> wrote:
>> >>
>> >> > some end boundary is only page alignment, instead of 2M alignment, so
>> >> > call ker_phycial_mapping_init three times. then don't overmap above
>> >> > the max_low_pfn
>> >>
>> >> applied to tip/x86/unify-setup, thanks Yinghai.
>> >
>> > -tip testing found a 32-bit spontaneous reboot during bootup, with this
>> > config:
>> >
>> > http://redhat.com/~mingo/misc/config-Sat_Jun_28_08_43_52_CEST_2008.bad
>> >
>> > no serial log produced at all.
>> even with console=uart8250,io,0x03f8,115200n8 ?
>
> correct, i use earlyprintk=serial,ttyS0,115200.
>
> i just re-tested it - doing "git merge tip/x86/unify-setup" in the
> tip/master branch triggers the failure with that config on a wide range
> of 32-bit boxes.
please check attached v3..., that is less intrusive.
YH
View attachment "fix_over_mapped.patch" of type "text/x-patch" (2810 bytes)
Powered by blists - more mailing lists