| Message ID | 20230327152112.15635-1-wsa+renesas@sang-engineering.com (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | cdeccd13a03fc6e97c090c12a27566d02dace6d9 |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [net-next,v2] Revert "sh_eth: remove open coded netif_running()" | expand |
On 3/27/23 6:21 PM, Wolfram Sang wrote: > This reverts commit ce1fdb065695f49ef6f126d35c1abbfe645d62d5. It turned > out this actually introduces a race condition. netif_running() is not a > suitable check for get_stats. > > Reported-by: Sergey Shtylyov <s.shtylyov@omp.ru> > Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com> [...] Reviewed-by: Sergey Shtylyov <s.shtylyov@omp.ru> MBR, Sergey
On Mon, Mar 27, 2023 at 05:21:12PM +0200, Wolfram Sang wrote: > This reverts commit ce1fdb065695f49ef6f126d35c1abbfe645d62d5. It turned > out this actually introduces a race condition. netif_running() is not a > suitable check for get_stats. > > Reported-by: Sergey Shtylyov <s.shtylyov@omp.ru> > Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com> Reviewed-by: Simon Horman <simon.horman@corigine.com>
Hello: This patch was applied to netdev/net-next.git (main) by Jakub Kicinski <kuba@kernel.org>: On Mon, 27 Mar 2023 17:21:12 +0200 you wrote: > This reverts commit ce1fdb065695f49ef6f126d35c1abbfe645d62d5. It turned > out this actually introduces a race condition. netif_running() is not a > suitable check for get_stats. > > Reported-by: Sergey Shtylyov <s.shtylyov@omp.ru> > Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com> > > [...] Here is the summary with links: - [net-next,v2] Revert "sh_eth: remove open coded netif_running()" https://git.kernel.org/netdev/net-next/c/cdeccd13a03f You are awesome, thank you!
diff --git a/drivers/net/ethernet/renesas/sh_eth.c b/drivers/net/ethernet/renesas/sh_eth.c index 2d9787231099..d8ec729825be 100644 --- a/drivers/net/ethernet/renesas/sh_eth.c +++ b/drivers/net/ethernet/renesas/sh_eth.c @@ -2441,6 +2441,8 @@ static int sh_eth_open(struct net_device *ndev) netif_start_queue(ndev); + mdp->is_opened = 1; + return ret; out_free_irq: @@ -2563,7 +2565,7 @@ static struct net_device_stats *sh_eth_get_stats(struct net_device *ndev) if (mdp->cd->no_tx_cntrs) return &ndev->stats; - if (!netif_running(ndev)) + if (!mdp->is_opened) return &ndev->stats; sh_eth_update_stat(ndev, &ndev->stats.tx_dropped, TROCR); @@ -2612,6 +2614,8 @@ static int sh_eth_close(struct net_device *ndev) /* Free all the skbuffs in the Rx queue and the DMA buffer. */ sh_eth_ring_free(ndev); + mdp->is_opened = 0; + pm_runtime_put(&mdp->pdev->dev); return 0; diff --git a/drivers/net/ethernet/renesas/sh_eth.h b/drivers/net/ethernet/renesas/sh_eth.h index f56dbc8a064a..a5c07c6ff44a 100644 --- a/drivers/net/ethernet/renesas/sh_eth.h +++ b/drivers/net/ethernet/renesas/sh_eth.h @@ -560,6 +560,7 @@ struct sh_eth_private { unsigned no_ether_link:1; unsigned ether_link_active_low:1; + unsigned is_opened:1; unsigned wol_enabled:1; };
This reverts commit ce1fdb065695f49ef6f126d35c1abbfe645d62d5. It turned out this actually introduces a race condition. netif_running() is not a suitable check for get_stats. Reported-by: Sergey Shtylyov <s.shtylyov@omp.ru> Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com> --- Change since v1: * added 'net-next' to $subject drivers/net/ethernet/renesas/sh_eth.c | 6 +++++- drivers/net/ethernet/renesas/sh_eth.h | 1 + 2 files changed, 6 insertions(+), 1 deletion(-)