[PATCH] net: ethernet: miss media cleanup behavior

[PATCH] net: ethernet: miss media cleanup behavior

[Wentao Liang]

In the de21041_media_timer(), line 1081, when media type is locked,
the code jumps to line 1136 to perform cleanup operations. However,
in the de21040_media_timer(), line 991, the same condition leads to
an immediate return without any cleanup.

To address this inconsistency, we have added a jump statement to the
de21040_media_timer() to ensure that cleanup operations are executed
before the function returns.

Signed-off-by: Wentao Liang <liangwentao@iscas.ac.cn>
---
 drivers/net/ethernet/dec/tulip/de2104x.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/dec/tulip/de2104x.c b/drivers/net/ethernet/dec/tulip/de2104x.c
index 0a161a4db242..724c0b3b3966 100644
--- a/drivers/net/ethernet/dec/tulip/de2104x.c
+++ b/drivers/net/ethernet/dec/tulip/de2104x.c
@@ -988,7 +988,7 @@ static void de21040_media_timer (struct timer_list *t)
 	de_link_down(de);
 
 	if (de->media_lock)
-		return;
+		goto set_media;
 
 	if (de->media_type == DE_MEDIA_AUI) {
 		static const u32 next_state = DE_MEDIA_TP;
@@ -998,6 +998,7 @@ static void de21040_media_timer (struct timer_list *t)
 		de_next_media(de, &next_state, 1);
 	}
 
+set_media:
 	spin_lock_irqsave(&de->lock, flags);
 	de_stop_rxtx(de);
 	spin_unlock_irqrestore(&de->lock, flags);
-- 
2.42.0.windows.2

Re: [PATCH] net: ethernet: miss media cleanup behavior

[Jiri Pirko]

Wed, Nov 06, 2024 at 03:11:52PM CET, liangwentao@iscas.ac.cn wrote:
>In the de21041_media_timer(), line 1081, when media type is locked,
>the code jumps to line 1136 to perform cleanup operations. However,
>in the de21040_media_timer(), line 991, the same condition leads to
>an immediate return without any cleanup.

Don't use line numbers please.


>
>To address this inconsistency, we have added a jump statement to the

Who's "we"? Just tell the codebase what to do, what to add, what to
change, etc.


>de21040_media_timer() to ensure that cleanup operations are executed
>before the function returns.
>
>Signed-off-by: Wentao Liang <liangwentao@iscas.ac.cn>

You are missing "Fixes" tag blaming the commit that introduced the
issue.


Actually, just to make things in the same way de21041_media_timer() has
it is not a good reason to do so. What are you trying to fix? What issue
you see. Why the code change is ok?

The code itself looks okay to me.


>---
> drivers/net/ethernet/dec/tulip/de2104x.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
>diff --git a/drivers/net/ethernet/dec/tulip/de2104x.c b/drivers/net/ethernet/dec/tulip/de2104x.c
>index 0a161a4db242..724c0b3b3966 100644
>--- a/drivers/net/ethernet/dec/tulip/de2104x.c
>+++ b/drivers/net/ethernet/dec/tulip/de2104x.c
>@@ -988,7 +988,7 @@ static void de21040_media_timer (struct timer_list *t)
> 	de_link_down(de);
> 
> 	if (de->media_lock)
>-		return;
>+		goto set_media;
> 
> 	if (de->media_type == DE_MEDIA_AUI) {
> 		static const u32 next_state = DE_MEDIA_TP;
>@@ -998,6 +998,7 @@ static void de21040_media_timer (struct timer_list *t)
> 		de_next_media(de, &next_state, 1);
> 	}
> 
>+set_media:
> 	spin_lock_irqsave(&de->lock, flags);
> 	de_stop_rxtx(de);
> 	spin_unlock_irqrestore(&de->lock, flags);
>-- 
>2.42.0.windows.2
>
>