When executing ioctl to allocate tags, if the peer address is 0,
mctp_alloc_local_tag now replaces it with 0xff. However, during tag
dropping, this replacement is not performed, potentially causing the key
not to be dropped as expected.

Signed-off-by: John Wang <wangzhiqiang02@ieisystem.com>

---
v2:
  - Change the subject from 'net' to 'net-next'
  - remove the Change-Id tag
---
 net/mctp/af_mctp.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/net/mctp/af_mctp.c b/net/mctp/af_mctp.c
index de52a9191da0..43288b408fde 100644
--- a/net/mctp/af_mctp.c
+++ b/net/mctp/af_mctp.c
@@ -486,6 +486,9 @@ static int mctp_ioctl_droptag(struct mctp_sock *msk, bool tagv2,
 	tag = ctl.tag & MCTP_TAG_MASK;
 	rc = -EINVAL;
 
+	if (ctl.peer_addr == MCTP_ADDR_NULL)
+		ctl.peer_addr = MCTP_ADDR_ANY;
+
 	spin_lock_irqsave(&net->mctp.keys_lock, flags);
 	hlist_for_each_entry_safe(key, tmp, &msk->keys, sklist) {
 		/* we do an irqsave here, even though we know the irq state,
-- 
2.34.1

On Tue, 30 Jul 2024 16:46:35 +0800 John Wang wrote:
> When executing ioctl to allocate tags, if the peer address is 0,
> mctp_alloc_local_tag now replaces it with 0xff. However, during tag
> dropping, this replacement is not performed, potentially causing the key
> not to be dropped as expected.
> 
> Signed-off-by: John Wang <wangzhiqiang02@ieisystem.com>

Looks sane. Jeremy? Matt?
In netdev we try to review patches within 24-48 hours.
You have willingly boarded this crazy train.. :)

Hi Jakub and John,

> On Tue, 30 Jul 2024 16:46:35 +0800 John Wang wrote:
> > When executing ioctl to allocate tags, if the peer address is 0,
> > mctp_alloc_local_tag now replaces it with 0xff. However, during tag
> > dropping, this replacement is not performed, potentially causing
> > the key
> > not to be dropped as expected.
> > 
> > Signed-off-by: John Wang <wangzhiqiang02@ieisystem.com>
> 
> Looks sane. Jeremy? Matt?

All looks good to me!

Reviewed-by: Jeremy Kerr <jk@codeconstruct.com.au>

(John had already discussed the change with us, so no surprises on my
side)

> In netdev we try to review patches within 24-48 hours.
> You have willingly boarded this crazy train.. :)

Yeah we bought express tickets to netdev town! I just saw that there
were nipa warnings on patchwork, so was waiting on a v3. If it's okay
as-is, I'm happy for a merge.

Cheers,


Jeremy

On Fri, 02 Aug 2024 08:21:46 +0800 Jeremy Kerr wrote:
> > Looks sane. Jeremy? Matt?  
> 
> All looks good to me!
> 
> Reviewed-by: Jeremy Kerr <jk@codeconstruct.com.au>

Thanks!

> (John had already discussed the change with us, so no surprises on my
> side)
> 
> > In netdev we try to review patches within 24-48 hours.
> > You have willingly boarded this crazy train.. :)  
> 
> Yeah we bought express tickets to netdev town! I just saw that there
> were nipa warnings on patchwork, so was waiting on a v3. If it's okay
> as-is, I'm happy for a merge.

It has quite a few false-positives (especially from the in-tree tools
like checkpatch and get_maintainer).