On Fri, 2009-10-09 at 10:08 +0800, Wu Zhangjin wrote:
> On Thu, 2009-10-08 at 22:44 +0200, Ralf Baechle wrote:
> > On Thu, Oct 08, 2009 at 10:21:12PM +0200, Rafael J. Wysocki wrote:
> >
> > > > Here, we fix it via checking pfn is in the "System RAM" or not. and
> > > > Seems pfn_valid() is not called in assembly code, we move it to
> > > > "!__ASSEMBLY__" to ensure we can simply declare it via "extern int
> > > > pfn_valid(unsigned long)" without Compiling Error.
> > > >
> > > > (This -v1 version incorporates feedback from Pavel Machek <pavel@xxxxxx>
> > > > and Sergei Shtylyov <sshtylyov@xxxxxxxxxxxxx>)
> > >
> > > Hmm. What exactly would be wrong with using register_nosave_region() or
> > > register_nosave_region_late() like x86 does?
>
> Have tried to use register_nosave_region(), it works, but seems there is
> something else wrong there.
>
> >
> > That seems to be more the fix than pfn_valid / PageReserved fiddlery we were
> > discussing in the other thread. Thanks!
>
> Just checked the arch/mips/loongson/common/mem.c, Seems it did not
> register any reserved pages, in reality, two sections of memory are
> reserved.
>
> here should be the patch, tested it, works.
>
> diff --git a/arch/mips/loongson/common/mem.c
> b/arch/mips/loongson/common/mem.c
> index 7c92f79..069d20b 100644
> --- a/arch/mips/loongson/common/mem.c
> +++ b/arch/mips/loongson/common/mem.c
> @@ -15,11 +15,17 @@
>
> void __init prom_init_memory(void)
> {
> - add_memory_region(0x0, (memsize << 20), BOOT_MEM_RAM);
> + add_memory_region(0x0, memsize << 20, BOOT_MEM_RAM);
> +
> + add_memory_region(memsize << 20, LOONGSON_PCI_MEM_START - (memsize
> <<
> + 20), BOOT_MEM_RESERVED);
> #ifdef CONFIG_64BIT
> if (highmemsize > 0)
> add_memory_region(LOONGSON_HIGHMEM_START,
> highmemsize << 20, BOOT_MEM_RAM);
> +
> + add_memory_region(LOONGSON_PCI_MEM_END + 1, LOONGSON_HIGHMEM_START
> -
> + LOONGSON_PCI_MEM_END - 1, BOOT_MEM_RESERVED);
> #endif /* CONFIG_64BIT */
> }
The above patch can not fix the problem when enabled FLATMEM in
linux-2.6.32-rc3, the real problem should be that we need register the
"pci memory space" as nosave pages, and also, the above "reserved"(not
memory) pages should be registered as nosave pages. but the simpler
solution should be the pfn_valid() I sent out in this E-mail thread, we
just need to check whether they are "valid", if they are "System
RAM"(BOOT_MEM_RAM or BOOT_MEM_ROM_DATA), they should be valid.
and what's more? should be register "pci memory space" as nosave pages
for all architecture?
Regards,
Wu Zhangjin
