Linux CXL
 help / color / mirror / Atom feed
* [PATCH] daxctl: Fix memblock enumeration off-by-one
@ 2023-02-02 20:56 Dan Williams
  2023-02-03  0:03 ` Eliot Moss
  2023-02-03 23:07 ` Verma, Vishal L
  0 siblings, 2 replies; 5+ messages in thread
From: Dan Williams @ 2023-02-02 20:56 UTC (permalink / raw)
  To: vishal.l.verma; +Cc: nvdimm, linux-cxl

A memblock is an inclusive memory range. Bound the search by the last
address in the memory block.

Found by wondering why an offline 32-block (at 128MB == 4GB) range was
reported as 33 blocks with one online.

Signed-off-by: Dan Williams <dan.j.williams@intel.com>
---
 daxctl/lib/libdaxctl.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/daxctl/lib/libdaxctl.c b/daxctl/lib/libdaxctl.c
index 5703992f5b88..d990479d8585 100644
--- a/daxctl/lib/libdaxctl.c
+++ b/daxctl/lib/libdaxctl.c
@@ -1477,7 +1477,7 @@ static int memblock_in_dev(struct daxctl_memory *mem, const char *memblock)
 		err(ctx, "%s: Unable to determine resource\n", devname);
 		return -EACCES;
 	}
-	dev_end = dev_start + daxctl_dev_get_size(dev);
+	dev_end = dev_start + daxctl_dev_get_size(dev) - 1;
 
 	memblock_size = daxctl_memory_get_block_size(mem);
 	if (!memblock_size) {


^ permalink raw reply related	[flat|nested] 5+ messages in thread

* Re: [PATCH] daxctl: Fix memblock enumeration off-by-one
  2023-02-02 20:56 [PATCH] daxctl: Fix memblock enumeration off-by-one Dan Williams
@ 2023-02-03  0:03 ` Eliot Moss
  2023-02-03  0:11   ` Dan Williams
  2023-02-03 23:07 ` Verma, Vishal L
  1 sibling, 1 reply; 5+ messages in thread
From: Eliot Moss @ 2023-02-03  0:03 UTC (permalink / raw)
  To: Dan Williams, vishal.l.verma; +Cc: nvdimm, linux-cxl

On 2/3/2023 7:56 AM, Dan Williams wrote:
> A memblock is an inclusive memory range. Bound the search by the last
> address in the memory block.
> 
> Found by wondering why an offline 32-block (at 128MB == 4GB) range was
> reported as 33 blocks with one online.
> 
> Signed-off-by: Dan Williams <dan.j.williams@intel.com>
> ---
>   daxctl/lib/libdaxctl.c |    2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/daxctl/lib/libdaxctl.c b/daxctl/lib/libdaxctl.c
> index 5703992f5b88..d990479d8585 100644
> --- a/daxctl/lib/libdaxctl.c
> +++ b/daxctl/lib/libdaxctl.c
> @@ -1477,7 +1477,7 @@ static int memblock_in_dev(struct daxctl_memory *mem, const char *memblock)
>   		err(ctx, "%s: Unable to determine resource\n", devname);
>   		return -EACCES;
>   	}
> -	dev_end = dev_start + daxctl_dev_get_size(dev);
> +	dev_end = dev_start + daxctl_dev_get_size(dev) - 1;
>   
>   	memblock_size = daxctl_memory_get_block_size(mem);
>   	if (!memblock_size) {

Might this address the bug I reported?

Regards - Eliot Moss

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] daxctl: Fix memblock enumeration off-by-one
  2023-02-03  0:03 ` Eliot Moss
@ 2023-02-03  0:11   ` Dan Williams
  2023-02-03  0:59     ` Eliot Moss
  0 siblings, 1 reply; 5+ messages in thread
From: Dan Williams @ 2023-02-03  0:11 UTC (permalink / raw)
  To: Eliot Moss, Dan Williams, vishal.l.verma; +Cc: nvdimm, linux-cxl

Eliot Moss wrote:
> On 2/3/2023 7:56 AM, Dan Williams wrote:
> > A memblock is an inclusive memory range. Bound the search by the last
> > address in the memory block.
> > 
> > Found by wondering why an offline 32-block (at 128MB == 4GB) range was
> > reported as 33 blocks with one online.
> > 
> > Signed-off-by: Dan Williams <dan.j.williams@intel.com>
> > ---
> >   daxctl/lib/libdaxctl.c |    2 +-
> >   1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/daxctl/lib/libdaxctl.c b/daxctl/lib/libdaxctl.c
> > index 5703992f5b88..d990479d8585 100644
> > --- a/daxctl/lib/libdaxctl.c
> > +++ b/daxctl/lib/libdaxctl.c
> > @@ -1477,7 +1477,7 @@ static int memblock_in_dev(struct daxctl_memory *mem, const char *memblock)
> >   		err(ctx, "%s: Unable to determine resource\n", devname);
> >   		return -EACCES;
> >   	}
> > -	dev_end = dev_start + daxctl_dev_get_size(dev);
> > +	dev_end = dev_start + daxctl_dev_get_size(dev) - 1;
> >   
> >   	memblock_size = daxctl_memory_get_block_size(mem);
> >   	if (!memblock_size) {
> 
> Might this address the bug I reported?

This one?

http://lore.kernel.org/r/558d0ff1-4658-a11b-5a6d-0be0a3a6799c@cs.umass.edu

I don't think so, that one seems to have something to do with the file
extent layout that causes fs/dax.c to fallback to 4K mappings.

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] daxctl: Fix memblock enumeration off-by-one
  2023-02-03  0:11   ` Dan Williams
@ 2023-02-03  0:59     ` Eliot Moss
  0 siblings, 0 replies; 5+ messages in thread
From: Eliot Moss @ 2023-02-03  0:59 UTC (permalink / raw)
  To: Dan Williams, vishal.l.verma; +Cc: nvdimm, linux-cxl

On 2/3/2023 11:11 AM, Dan Williams wrote:
> Eliot Moss wrote:
>> On 2/3/2023 7:56 AM, Dan Williams wrote:
>>> A memblock is an inclusive memory range. Bound the search by the last
>>> address in the memory block.
>>>
>>> Found by wondering why an offline 32-block (at 128MB == 4GB) range was
>>> reported as 33 blocks with one online.
>>>
>>> Signed-off-by: Dan Williams <dan.j.williams@intel.com>
>>> ---
>>>    daxctl/lib/libdaxctl.c |    2 +-
>>>    1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/daxctl/lib/libdaxctl.c b/daxctl/lib/libdaxctl.c
>>> index 5703992f5b88..d990479d8585 100644
>>> --- a/daxctl/lib/libdaxctl.c
>>> +++ b/daxctl/lib/libdaxctl.c
>>> @@ -1477,7 +1477,7 @@ static int memblock_in_dev(struct daxctl_memory *mem, const char *memblock)
>>>    		err(ctx, "%s: Unable to determine resource\n", devname);
>>>    		return -EACCES;
>>>    	}
>>> -	dev_end = dev_start + daxctl_dev_get_size(dev);
>>> +	dev_end = dev_start + daxctl_dev_get_size(dev) - 1;
>>>    
>>>    	memblock_size = daxctl_memory_get_block_size(mem);
>>>    	if (!memblock_size) {
>>
>> Might this address the bug I reported?
> 
> This one?
> 
> http://lore.kernel.org/r/558d0ff1-4658-a11b-5a6d-0be0a3a6799c@cs.umass.edu
> 
> I don't think so, that one seems to have something to do with the file
> extent layout that causes fs/dax.c to fallback to 4K mappings.

That was the one - I think you're right; it's not immediately related.

EM

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH] daxctl: Fix memblock enumeration off-by-one
  2023-02-02 20:56 [PATCH] daxctl: Fix memblock enumeration off-by-one Dan Williams
  2023-02-03  0:03 ` Eliot Moss
@ 2023-02-03 23:07 ` Verma, Vishal L
  1 sibling, 0 replies; 5+ messages in thread
From: Verma, Vishal L @ 2023-02-03 23:07 UTC (permalink / raw)
  To: Williams, Dan J; +Cc: linux-cxl@vger.kernel.org, nvdimm@lists.linux.dev

On Thu, 2023-02-02 at 12:56 -0800, Dan Williams wrote:
> A memblock is an inclusive memory range. Bound the search by the last
> address in the memory block.
> 
> Found by wondering why an offline 32-block (at 128MB == 4GB) range was
> reported as 33 blocks with one online.
> 
> Signed-off-by: Dan Williams <dan.j.williams@intel.com>
> ---
>  daxctl/lib/libdaxctl.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)

Good find! Applied, thanks.

> 
> diff --git a/daxctl/lib/libdaxctl.c b/daxctl/lib/libdaxctl.c
> index 5703992f5b88..d990479d8585 100644
> --- a/daxctl/lib/libdaxctl.c
> +++ b/daxctl/lib/libdaxctl.c
> @@ -1477,7 +1477,7 @@ static int memblock_in_dev(struct daxctl_memory *mem, const char *memblock)
>                 err(ctx, "%s: Unable to determine resource\n", devname);
>                 return -EACCES;
>         }
> -       dev_end = dev_start + daxctl_dev_get_size(dev);
> +       dev_end = dev_start + daxctl_dev_get_size(dev) - 1;
>  
>         memblock_size = daxctl_memory_get_block_size(mem);
>         if (!memblock_size) {
> 


^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2023-02-03 23:08 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2023-02-02 20:56 [PATCH] daxctl: Fix memblock enumeration off-by-one Dan Williams
2023-02-03  0:03 ` Eliot Moss
2023-02-03  0:11   ` Dan Williams
2023-02-03  0:59     ` Eliot Moss
2023-02-03 23:07 ` Verma, Vishal L

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox