[PATCH v2] net: ethernet: rnpgbe: mark nonfunctional incomplete driver as BROKEN

[PATCH v2] net: ethernet: rnpgbe: mark nonfunctional incomplete driver as BROKEN

[Ethan Nelson-Moore]

The rnpgbe driver as currently shipped in the kernel is incomplete and
has no useful functionality. It will bind to a PCI device and create a
network device, but that device does not function (its .ndo_start_xmit
callback, rnpgbe_xmit_frame, just drops all packets). This situation
means that users could enable this driver and have it load and attach
to their device but not transfer any data. To remove the potential for
user confusion, mark the driver as broken until it is completed and
explain why this was done.

Fixes: ee61c10cd482 ("net: rnpgbe: Add build support for rnpgbe")
Cc: stable@vger.kernel.org # 7.0+
Signed-off-by: Ethan Nelson-Moore <enelsonmoore@gmail.com>
---
Changes from v1: Do not mark vendor section as BROKEN
 drivers/net/ethernet/mucse/Kconfig | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/mucse/Kconfig b/drivers/net/ethernet/mucse/Kconfig
index 0b3e853d625f..810fc1230d97 100644
--- a/drivers/net/ethernet/mucse/Kconfig
+++ b/drivers/net/ethernet/mucse/Kconfig
@@ -16,12 +16,14 @@ config NET_VENDOR_MUCSE
 
 if NET_VENDOR_MUCSE
 
+# This driver is marked as broken because it is incomplete; this avoids users
+# enabling it and expecting it to work.
 config MGBE
 	tristate "Mucse(R) 1GbE PCI Express adapters support"
-	depends on PCI
+	depends on PCI && BROKEN
 	help
 	  This driver supports Mucse(R) 1GbE PCI Express family of
-	  adapters.
+	  adapters. It is incomplete and currently has no useful functionality.
 
 	  More specific information on configuring the driver is in
 	  <file:Documentation/networking/device_drivers/ethernet/mucse/rnpgbe.rst>.
-- 
2.43.0

Re: [PATCH v2] net: ethernet: rnpgbe: mark nonfunctional incomplete driver as BROKEN

[Jakub Kicinski]

On Mon, 27 Apr 2026 20:08:25 -0700 Ethan Nelson-Moore wrote:
> The rnpgbe driver as currently shipped in the kernel is incomplete and
> has no useful functionality. It will bind to a PCI device and create a
> network device, but that device does not function (its .ndo_start_xmit
> callback, rnpgbe_xmit_frame, just drops all packets). This situation
> means that users could enable this driver and have it load and attach
> to their device but not transfer any data. To remove the potential for
> user confusion, mark the driver as broken until it is completed and
> explain why this was done.

I'm having second thoughts about this. I'm worried users will come to
expect that drivers are marked as BROKEN until such time that they
can be considered a sufficiently complete replacement for an OOT /
vendor driver. This will be highly subjective.

Re: [PATCH v2] net: ethernet: rnpgbe: mark nonfunctional incomplete driver as BROKEN

[Ethan Nelson-Moore]

On Wed, Apr 29, 2026 at 5:54 PM Jakub Kicinski <kuba@kernel.org> wrote:
> I'm having second thoughts about this. I'm worried users will come to
> expect that drivers are marked as BROKEN until such time that they
> can be considered a sufficiently complete replacement for an OOT /
> vendor driver. This will be highly subjective.

Hi, Jakub,

I understand your concern, but there is precedent for doing this when
the driver doesn't work at all.

The ntsync driver was marked as broken in commit f5b335dc025c ("misc:
ntsync: mark driver as "broken" to prevent from building") until it
was fully merged. The BROKEN dependency was then removed in commit
c301e1fefc2d ("ntsync: No longer depend on BROKEN.")

If it were my decision, I would remove BROKEN from this driver once it
supports a stable network connection, and perhaps also once it
survives suspends and resumes, since that is expected in modern
desktop use cases. I think that is a fairly objective reading of the
word BROKEN.

It might also be a good idea to agree on expected uses of
CONFIG_BROKEN and document them in init/Kconfig.
Currently it says:
  This option allows you to choose whether you want to try to
  compile (and fix) old drivers that haven't been updated to
  new infrastructure.
which does not fully encompass what it is used for.

Ethan

Re: [PATCH v2] net: ethernet: rnpgbe: mark nonfunctional incomplete driver as BROKEN

[Jakub Kicinski]

On Wed, 29 Apr 2026 18:13:46 -0700 Ethan Nelson-Moore wrote:
> On Wed, Apr 29, 2026 at 5:54 PM Jakub Kicinski <kuba@kernel.org> wrote:
> > I'm having second thoughts about this. I'm worried users will come to
> > expect that drivers are marked as BROKEN until such time that they
> > can be considered a sufficiently complete replacement for an OOT /
> > vendor driver. This will be highly subjective.  
> 
> I understand your concern, but there is precedent for doing this when
> the driver doesn't work at all.
> 
> The ntsync driver was marked as broken in commit f5b335dc025c ("misc:
> ntsync: mark driver as "broken" to prevent from building") until it
> was fully merged. The BROKEN dependency was then removed in commit
> c301e1fefc2d ("ntsync: No longer depend on BROKEN.")

ntsync is mere 1.2kLoC, that's 10x smaller than typical networking
driver these days, and even (slightly) smaller than rnpgbe(!)

Most in-tree networking drivers play constant catch-up with the OOT
version so they are never "fully merged".

> If it were my decision, I would remove BROKEN from this driver once it
> supports a stable network connection, and perhaps also once it
> survives suspends and resumes, since that is expected in modern
> desktop use cases. I think that is a fairly objective reading of the
> word BROKEN.

That'd require us to know if the device is going to be used on
desktop because for datacenter NICs suspend/resume does not matter.

> It might also be a good idea to agree on expected uses of
> CONFIG_BROKEN and document them in init/Kconfig.
> Currently it says:
>   This option allows you to choose whether you want to try to
>   compile (and fix) old drivers that haven't been updated to
>   new infrastructure.
> which does not fully encompass what it is used for.

Maybe it was written in simpler times, maybe the help message was
aspirational to begin with..

Re: [PATCH v2] net: ethernet: rnpgbe: mark nonfunctional incomplete driver as BROKEN

[Ethan Nelson-Moore]

Hi, Jakub,

On Wed, Apr 29, 2026 at 6:27 PM Jakub Kicinski <kuba@kernel.org> wrote:
> That'd require us to know if the device is going to be used on
> desktop because for datacenter NICs suspend/resume does not matter.

Good point - I was thinking about that myself. I would be fine with
only marking it as broken until it can actually transmit and receive
packets. What do you think about that? Of course, as a maintainer, the
decision is up to you in the end.

> Maybe it was written in simpler times, maybe the help message was
> aspirational to begin with..
Probably both. On a related note, I think something needs to be done
about some drivers staying in staging indefinitely without being moved
out of it or removed. The most obvious example is rtl8723bs. Eager new
contributors have spent an inordinate amount of time cleaning it up,
yet it is nowhere near ready to be taken out of staging. One major
blocker is that it doesn't use mac80211, but changing it to do so
would be nearly impossible in its current state. I think it makes the
most sense now to either remove the driver entirely if no one is using
it or add SDIO support to rtl8xxxu (some work has already done on this
out-of-tree).

Ethan