| Message ID | 20220223142337.41757-1-maukka@ext.kapsi.fi (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | 42404d8f1c01861b22ccfa1d70f950242720ae57 |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [v2] net: mv643xx_eth: process retval from of_get_mac_address | expand |
On Wed, 23 Feb 2022 16:23:37 +0200 Mauri Sandberg wrote: > Obtaining a MAC address may be deferred in cases when the MAC is stored > in an NVMEM block, for example, and it may not be ready upon the first > retrieval attempt and return EPROBE_DEFER. > > It is also possible that a port that does not rely on NVMEM has been > already created when getting the defer request. Thus, also the resources > allocated previously must be freed when doing a roll-back. > > Signed-off-by: Mauri Sandberg <maukka@ext.kapsi.fi> > Cc: Andrew Lunn <andrew@lunn.ch> While we wait for Andrew's ack, is this the correct fixes tag? Fixes: 76723bca2802 ("net: mv643xx_eth: add DT parsing support")
On Thu, Feb 24, 2022 at 08:57:54AM -0800, Jakub Kicinski wrote: > On Wed, 23 Feb 2022 16:23:37 +0200 Mauri Sandberg wrote: > > Obtaining a MAC address may be deferred in cases when the MAC is stored > > in an NVMEM block, for example, and it may not be ready upon the first > > retrieval attempt and return EPROBE_DEFER. > > > > It is also possible that a port that does not rely on NVMEM has been > > already created when getting the defer request. Thus, also the resources > > allocated previously must be freed when doing a roll-back. > > > > Signed-off-by: Mauri Sandberg <maukka@ext.kapsi.fi> > > Cc: Andrew Lunn <andrew@lunn.ch> > > While we wait for Andrew's ack, is this the correct fixes tag? > > Fixes: 76723bca2802 ("net: mv643xx_eth: add DT parsing support") Yes, that looks correct. The history around here is convoluted, and anybody trying to backport that far is going to need a few different versions. Reviewed-by: Andrew Lunn <andrew@lunn.ch> Andrew
Hello: This patch was applied to netdev/net.git (master) by Jakub Kicinski <kuba@kernel.org>: On Wed, 23 Feb 2022 16:23:37 +0200 you wrote: > Obtaining a MAC address may be deferred in cases when the MAC is stored > in an NVMEM block, for example, and it may not be ready upon the first > retrieval attempt and return EPROBE_DEFER. > > It is also possible that a port that does not rely on NVMEM has been > already created when getting the defer request. Thus, also the resources > allocated previously must be freed when doing a roll-back. > > [...] Here is the summary with links: - [v2] net: mv643xx_eth: process retval from of_get_mac_address https://git.kernel.org/netdev/net/c/42404d8f1c01 You are awesome, thank you!
diff --git a/drivers/net/ethernet/marvell/mv643xx_eth.c b/drivers/net/ethernet/marvell/mv643xx_eth.c index 105247582684..143ca8be5eb5 100644 --- a/drivers/net/ethernet/marvell/mv643xx_eth.c +++ b/drivers/net/ethernet/marvell/mv643xx_eth.c @@ -2704,6 +2704,16 @@ MODULE_DEVICE_TABLE(of, mv643xx_eth_shared_ids); static struct platform_device *port_platdev[3]; +static void mv643xx_eth_shared_of_remove(void) +{ + int n; + + for (n = 0; n < 3; n++) { + platform_device_del(port_platdev[n]); + port_platdev[n] = NULL; + } +} + static int mv643xx_eth_shared_of_add_port(struct platform_device *pdev, struct device_node *pnp) { @@ -2740,7 +2750,9 @@ static int mv643xx_eth_shared_of_add_port(struct platform_device *pdev, return -EINVAL; } - of_get_mac_address(pnp, ppd.mac_addr); + ret = of_get_mac_address(pnp, ppd.mac_addr); + if (ret) + return ret; mv643xx_eth_property(pnp, "tx-queue-size", ppd.tx_queue_size); mv643xx_eth_property(pnp, "tx-sram-addr", ppd.tx_sram_addr); @@ -2804,21 +2816,13 @@ static int mv643xx_eth_shared_of_probe(struct platform_device *pdev) ret = mv643xx_eth_shared_of_add_port(pdev, pnp); if (ret) { of_node_put(pnp); + mv643xx_eth_shared_of_remove(); return ret; } } return 0; } -static void mv643xx_eth_shared_of_remove(void) -{ - int n; - - for (n = 0; n < 3; n++) { - platform_device_del(port_platdev[n]); - port_platdev[n] = NULL; - } -} #else static inline int mv643xx_eth_shared_of_probe(struct platform_device *pdev) {
Obtaining a MAC address may be deferred in cases when the MAC is stored in an NVMEM block, for example, and it may not be ready upon the first retrieval attempt and return EPROBE_DEFER. It is also possible that a port that does not rely on NVMEM has been already created when getting the defer request. Thus, also the resources allocated previously must be freed when doing a roll-back. Signed-off-by: Mauri Sandberg <maukka@ext.kapsi.fi> Cc: Andrew Lunn <andrew@lunn.ch> --- v1 -> v2 - escalate all error values from of_get_mac_address() - move mv643xx_eth_shared_of_remove() before mv643xx_eth_shared_of_probe() - release all resources potentially allocated for previous port nodes - update commit title and message --- drivers/net/ethernet/marvell/mv643xx_eth.c | 24 +++++++++++++--------- 1 file changed, 14 insertions(+), 10 deletions(-) base-commit: cfb92440ee71adcc2105b0890bb01ac3cddb8507