lkml.org 
[lkml]   [2017]   [Nov]   [22]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: mm/percpu.c: use smarter memory allocation for struct pcpu_alloc_info (crisv32 hang)
On Wed, 22 Nov 2017, Jesper Nilsson wrote:

> On Mon, Nov 20, 2017 at 10:50:46PM -0500, Nicolas Pitre wrote:
> > On Mon, 20 Nov 2017, Guenter Roeck wrote:
> > > On Mon, Nov 20, 2017 at 07:28:21PM -0500, Nicolas Pitre wrote:
> > > > On Mon, 20 Nov 2017, Guenter Roeck wrote:
> > > >
> > > > > bdata->node_min_pfn=60000 PFN_PHYS(bdata->node_min_pfn)=c0000000 start_off=536000 region=c0536000
> > > >
> > > > If PFN_PHYS(bdata->node_min_pfn)=c0000000 and
> > > > region=c0536000 that means phys_to_virt() is a no-op.
> > > >
> > > No, it is |= 0x80000000
> >
> > Then the bootmem registration looks very fishy. If you have:
> >
> > > I think the problem is the 0x60000 in bdata->node_min_pfn. It is shifted
> > > left by PFN_PHYS, making it 0xc0000000, which in my understanding is
> > > a virtual address.
> >
> > Exact.
> >
> > #define __pa(x) ((unsigned long)(x) & 0x7fffffff)
> > #define __va(x) ((void *)((unsigned long)(x) | 0x80000000))
> >
> > With that, the only possible physical address range you may have is
> > 0x40000000 - 0x7fffffff, and it better start at 0x40000000. If that's
> > not where your RAM is then something is wrong.
> >
> > This is in fact a very bad idea to define __va() and __pa() using
> > bitwise operations as this hides mistakes like defining physical RAM
> > address at 0xc0000000. Instead, it should look like:
> >
> > #define __pa(x) ((unsigned long)(x) - 0x80000000)
> > #define __va(x) ((void *)((unsigned long)(x) + 0x80000000))
> >
> > This way, bad physical RAM address definitions will be caught
> > immediately.
> >
> > > That doesn't seem to be easy to fix. It seems there is a mixup of physical
> > > and virtual addresses in the architecture.
> >
> > Well... I don't think there is much else to say other than this needs
> > fixing.
>
> The memory map for the ETRAX FS has the SDRAM mapped at both 0x40000000-0x7fffffff
> and 0xc0000000-0xffffffff, and the difference is cached and non-cached.
> That is actively (ab)used in the port, unfortunately, allthough I'm
> uncertain if this is the problem in this case.

It certainly is a problem. If your cached RAM is physically mapped at
0xc0000000 and you want it to be virtually mapped at 0xc0000000 then you
should have:

#define __pa(x) ((unsigned long)(x))
#define __va(x) ((void *)(x))

i.e. no translation. For non-cached RAM access, there are specific
interfaces for that. For example, you could have dma_alloc_coherent()
take advantage of the fact that memory with the top bit cleared becomes
uncached. But __pa() is the wrong interface for obtaining uncached
memory.


Nicolas

\
 
 \ /
  Last update: 2017-11-22 21:18    [W:0.057 / U:7.232 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site