Resend it base on v71.1.

-----Original Message-----
From: Dan Williams <[email protected]> 
Sent: Thursday, January 28, 2021 4:47 PM
To: Li, Redhairer <[email protected]>
Cc: linux-nvdimm <[email protected]>
Subject: Re: [PATCH 1/1] ndctl/namespace: Fix disable-namespace accounting 
relative to seed devices

On Fri, Jan 15, 2021 at 9:22 AM Li, Redhairer <[email protected]> wrote:
>
> Hi Dan,
>
> Your comment is make sense.
> ndctl_namespace_disable_safe will return 1 if namespace size is 0.
> I send a new patch out for review.

It looks ok but it does not apply to the current tip of tree now v71.1 can you 
resend?

>
> But I am not sure what do you mean for 2nd patch.
> In cmd_disable_namespace, it already print error if rc<0.
>         rc = do_xaction_namespace(namespace, ACTION_DISABLE, ctx, &disabled);
>         if (rc < 0 && !err_count)
>                 fprintf(stderr, "error disabling namespaces: %s\n",
>                                 strerror(-rc));

Hmm, you're right, once you change to the positive error code the report will 
just work. Did you give it a try does it fix the accounting problem with just 
your first patch?

>
> my patch is based on v70 due to latest one will see "FAIL: create.sh" when 
> make check even not include my change.

I know of at least one create.sh failure that was fixed by:

Kernel commit:

2dd2a1740ee1 libnvdimm/namespace: Fix reaping of invalidated 
block-window-namespace labels

...which is now in v5.11-rc1 and backported to v5.10-rc4. However, that bug 
only started triggering after ndctl changed to reconfigure namespaces in place 
with commit:

d4bc247faeda ndctl/namespace: Reconfigure in-place

..which was only merged into ndctl in v71.

Another kernel change that may be causing your failures is:

d1c5246e08eb x86/mm: Fix leak of pmd ptlock

...which was merged for v5.11-rc3 and backported to v5.10.7.

Can you run latest kernel and ndctl and see if you still see the create.sh 
failure?
_______________________________________________
Linux-nvdimm mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to