netdev - Re: [PATCH net-next] net: bgmac: support MDIO described in DT

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <3b468d96-8c36-ec04-8993-97f1de12c34f@gmail.com>
Date:   Sat, 2 Oct 2021 13:46:08 +0200
From:   Rafał Miłecki <zajec5@...il.com>
To:     Florian Fainelli <f.fainelli@...il.com>,
        "David S . Miller" <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>
Cc:     netdev@...r.kernel.org, devicetree@...r.kernel.org,
        bcm-kernel-feedback-list@...adcom.com,
        Rafał Miłecki <rafal@...ecki.pl>
Subject: Re: [PATCH net-next] net: bgmac: support MDIO described in DT

On 01.10.2021 01:04, Florian Fainelli wrote:
> On 9/30/21 7:29 AM, Rafał Miłecki wrote:
>> On 20.09.2021 19:57, Rafał Miłecki wrote:
>>> On 20.09.2021 18:11, Florian Fainelli wrote:
>>>> I believe this leaks np and the use case is not exactly clear to me
>>>> here. AFAICT the Northstar SoCs have two MDIO controllers: one for
>>>> internal PHYs and one for external PHYs which how you would attach a
>>>> switch to the chip (in chipcommonA). Is 53573 somewhat different here?
>>>> What is the MDIO bus driver that is being used?
>>>
>>> of_get_child_by_name() doesn't seem to increase refcount or anything and
>>> I think it's how most drivers handle it. I don't think it should leak.
>>>
>>> BCM53573 is a built with some older blocks. Please check:
>>>
>>> 4ebd50472899 ("ARM: BCM53573: Initial support for Broadcom BCM53573
>>> SoCs")
>>>       BCM53573 series is a new family with embedded wireless. By marketing
>>>       people it's sometimes called Northstar but it uses different CPU
>>> and has
>>>       different architecture so we need a new symbol for it.
>>>       Fortunately it shares some peripherals with other iProc based
>>> SoCs so we
>>>       will be able to reuse some drivers/bindings.
>>>
>>> e90d2d51c412 ("ARM: BCM5301X: Add basic dts for BCM53573 based Tenda
>>> AC9")
>>>       BCM53573 seems to be low priced alternative for Northstar
>>> chipsts. It
>>>       uses single core Cortex-A7 and doesn't have SDU or local (TWD)
>>> timer. It
>>>       was also stripped out of independent SPI controller and 2 GMACs.
>>>
>>> Northstar uses SRAB which is some memory based (0x18007000) access to
>>> switch register space.
>>> BCM53573 uses different blocks & mappings and it doesn't include SRAB at
>>> 0x18007000. Accessing switch registers is handled over MDIO.
>>
>> Florian: did my explanations help reviewing this patch? Would you ack it
>> now?
> 
> Thanks for providing the background.
> 
> You still appear to be needing an of_node_put() after
> of_mdiobus_register() because that function does increase the reference
> count.

I really can't find code increasing refcount.

I even attempted to runtime test it and I still can't see a leaking ref. See:

[    1.168863] bgmac_bcma bcma0:5: [bcma_mdio_mii_register] BEFORE count:2
[    1.176235] libphy: bcma_mdio mii bus: probed
[    1.181513] bcm53xx bcma_mdio-0-0:1e: found switch: BCM53125, rev 4
[    1.187936] bcm53xx bcma_mdio-0-0:1e: failed to register switch: -517
[    1.194610] bgmac_bcma bcma0:5: [bcma_mdio_mii_register]  AFTER count:2


diff --git a/drivers/net/ethernet/broadcom/bgmac-bcma-mdio.c b/drivers/net/ethernet/broadcom/bgmac-bcma-mdio.c
index 086739e4f..e52a3d8b7 100644
--- a/drivers/net/ethernet/broadcom/bgmac-bcma-mdio.c
+++ b/drivers/net/ethernet/broadcom/bgmac-bcma-mdio.c
@@ -233,11 +233,14 @@ struct mii_bus *bcma_mdio_mii_register(struct bgmac *bgmac)

         np = of_get_child_by_name(core->dev.of_node, "mdio");

+
+       dev_info(&core->dev, "[%s] BEFORE count:%d\n", __func__, refcount_read(&np->kobj.kref.refcount));
         err = of_mdiobus_register(mii_bus, np);
         if (err) {
                 dev_err(&core->dev, "Registration of mii bus failed\n");
                 goto err_free_bus;
         }
+       dev_info(&core->dev, "[%s]  AFTER count:%d\n", __func__, refcount_read(&np->kobj.kref.refcount));

         return mii_bus;