[PATCH] 8139too in linux-3.18.0: some potential bugs

[PATCH] 8139too in linux-3.18.0: some potential bugs

[Jia-Ju Bai]

I have actually tested 8139too driver on the real hardware(Realtek RTL8139D
PCI Ethernet Controller), and find some potential bugs:
The target file is drivers/net/ethernet/realtek/8139too.c, which is used to
build 8139too.ko.

(1) In the normal process of 8139too, netif_napi_add is called in
rtl8139_init_one, but netif_napi_del is not called in rtl8139_remove_one.
However, many other ethernet card drivers call them in pairs, even in the
error handling paths, such as r8169 and igb.

(2) In the normal process of 8139too, pci_enable_device and
pci_disable_device are called in pairs in rtl8139_init_board(in
rtl8139_init_one) and rtl8139_remove_one. However, when pci_enable_device
has been called and pci_request_regions is failed in rtl8139_init_board,
"err_out" segment in rtl8139_init_board is executed immediately to exit, but
pci_disable_device is not called because "disable_dev_on_err = 0".


Meanwhile, I also write the patch to fix the bugs. I have run the patch on
the hardware, it can work normally and fix the above bugs.

diff --git a/drivers/net/ethernet/realtek/8139too.c
b/drivers/net/ethernet/realtek/8139too.c
index 007b38c..f277c5f 100644
--- a/drivers/net/ethernet/realtek/8139too.c
+++ b/drivers/net/ethernet/realtek/8139too.c
@@ -782,11 +782,11 @@ static struct net_device *rtl8139_init_board(struct
pci_dev *pdev)
 	rc = pci_enable_device (pdev);
 	if (rc)
 		goto err_out;
+	disable_dev_on_err = 1;
 
 	rc = pci_request_regions (pdev, DRV_NAME);
 	if (rc)
 		goto err_out;
-	disable_dev_on_err = 1;
 
 	pci_set_master (pdev);
 
@@ -1098,6 +1098,7 @@ static int rtl8139_init_one(struct pci_dev *pdev,
 	return 0;
 
 err_out:
+	netif_napi_del (&tp->napi);
 	__rtl8139_cleanup_dev (dev);
 	pci_disable_device (pdev);
 	return i;
@@ -1112,6 +1113,7 @@ static void rtl8139_remove_one(struct pci_dev *pdev)
 	assert (dev != NULL);
 
 	cancel_delayed_work_sync(&tp->thread);
+	netif_napi_del (&tp->napi);
 
 	unregister_netdev (dev);

Thanks!

Re: [PATCH] 8139too in linux-3.18.0: some potential bugs

[Sergei Shtylyov]

On 12/20/2014 4:20 PM, Jia-Ju Bai wrote:

> I have actually tested 8139too driver on the real hardware(Realtek RTL8139D
> PCI Ethernet Controller), and find some potential bugs:
> The target file is drivers/net/ethernet/realtek/8139too.c, which is used to
> build 8139too.ko.

> (1) In the normal process of 8139too, netif_napi_add is called in
> rtl8139_init_one, but netif_napi_del is not called in rtl8139_remove_one.
> However, many other ethernet card drivers call them in pairs, even in the
> error handling paths, such as r8169 and igb.

> (2) In the normal process of 8139too, pci_enable_device and
> pci_disable_device are called in pairs in rtl8139_init_board(in
> rtl8139_init_one) and rtl8139_remove_one. However, when pci_enable_device
> has been called and pci_request_regions is failed in rtl8139_init_board,
> "err_out" segment in rtl8139_init_board is executed immediately to exit, but
> pci_disable_device is not called because "disable_dev_on_err = 0".

    Again, please fix one issue per patch.

> Meanwhile, I also write the patch to fix the bugs. I have run the patch on
> the hardware, it can work normally and fix the above bugs.

    Too bad you didn't provide a sign-off...

WBR, Sergei

Re: [PATCH] 8139too in linux-3.18.0: some potential bugs

[Jia-Ju Bai]

I have written it in the e-mail. Does not it display??

Okay, I send it again:

diff --git a/drivers/net/ethernet/realtek/8139too.c
b/drivers/net/ethernet/realtek/8139too.c
index 007b38c..f277c5f 100644
--- a/drivers/net/ethernet/realtek/8139too.c
+++ b/drivers/net/ethernet/realtek/8139too.c
@@ -782,11 +782,11 @@ static struct net_device *rtl8139_init_board(struct
pci_dev *pdev)
 	rc = pci_enable_device (pdev);
 	if (rc)
 		goto err_out;
+	disable_dev_on_err = 1;
 
 	rc = pci_request_regions (pdev, DRV_NAME);
 	if (rc)
 		goto err_out;
-	disable_dev_on_err = 1;
 
 	pci_set_master (pdev);
 
@@ -1098,6 +1098,7 @@ static int rtl8139_init_one(struct pci_dev *pdev,
 	return 0;
 
 err_out:
+	netif_napi_del (&tp->napi);
 	__rtl8139_cleanup_dev (dev);
 	pci_disable_device (pdev);
 	return i;
@@ -1112,6 +1113,7 @@ static void rtl8139_remove_one(struct pci_dev *pdev)
 	assert (dev != NULL);
 
 	cancel_delayed_work_sync(&tp->thread);
+	netif_napi_del (&tp->napi);
 
 	unregister_netdev (dev);

Re: [PATCH] 8139too in linux-3.18.0: some potential bugs

[Jia-Ju Bai]

Sorry, I know your meaning.


Signed-off-by: Jia-Ju Bai <baijiaju1990@163.com>
diff --git a/drivers/net/ethernet/realtek/8139too.c
b/drivers/net/ethernet/realtek/8139too.c
index 007b38c..f277c5f 100644
--- a/drivers/net/ethernet/realtek/8139too.c
+++ b/drivers/net/ethernet/realtek/8139too.c
@@ -782,11 +782,11 @@ static struct net_device *rtl8139_init_board(struct
pci_dev *pdev)
 	rc = pci_enable_device (pdev);
 	if (rc)
 		goto err_out;
+	disable_dev_on_err = 1;
 
 	rc = pci_request_regions (pdev, DRV_NAME);
 	if (rc)
 		goto err_out;
-	disable_dev_on_err = 1;
 
 	pci_set_master (pdev);
 
@@ -1098,6 +1098,7 @@ static int rtl8139_init_one(struct pci_dev *pdev,
 	return 0;
 
 err_out:
+	netif_napi_del(&tp->napi);
 	__rtl8139_cleanup_dev (dev);
 	pci_disable_device (pdev);
 	return i;
@@ -1112,6 +1113,7 @@ static void rtl8139_remove_one(struct pci_dev *pdev)
 	assert (dev != NULL);
 
 	cancel_delayed_work_sync(&tp->thread);
+	netif_napi_del(&tp->napi);
 
 	unregister_netdev (dev);


-----邮件原件-----
发件人: Sergei Shtylyov [mailto:sergei.shtylyov@cogentembedded.com] 
发送时间: 2014年12月20日 21:48
收件人: Jia-Ju Bai; netdev@vger.kernel.org; jgarzik@pobox.com;
shangh@realtek.com.tw
主题: Re: [PATCH] 8139too in linux-3.18.0: some potential bugs

On 12/20/2014 4:20 PM, Jia-Ju Bai wrote:

> I have actually tested 8139too driver on the real hardware(Realtek 
> RTL8139D PCI Ethernet Controller), and find some potential bugs:
> The target file is drivers/net/ethernet/realtek/8139too.c, which is 
> used to build 8139too.ko.

> (1) In the normal process of 8139too, netif_napi_add is called in 
> rtl8139_init_one, but netif_napi_del is not called in rtl8139_remove_one.
> However, many other ethernet card drivers call them in pairs, even in 
> the error handling paths, such as r8169 and igb.

> (2) In the normal process of 8139too, pci_enable_device and 
> pci_disable_device are called in pairs in rtl8139_init_board(in
> rtl8139_init_one) and rtl8139_remove_one. However, when 
> pci_enable_device has been called and pci_request_regions is failed in 
> rtl8139_init_board, "err_out" segment in rtl8139_init_board is 
> executed immediately to exit, but pci_disable_device is not called because
"disable_dev_on_err = 0".

    Again, please fix one issue per patch.

> Meanwhile, I also write the patch to fix the bugs. I have run the 
> patch on the hardware, it can work normally and fix the above bugs.

    Too bad you didn't provide a sign-off...

WBR, Sergei

Re: [PATCH] 8139too in linux-3.18.0: some potential bugs

[Sergei Shtylyov]

On 12/20/2014 5:15 PM, Jia-Ju Bai wrote:

> Sorry, I know your meaning.

    No, apparently you don't. Please:

1) don't send patches with the description stripped;
2) split your patch into 2 patches to solve one issue per patch.

> Signed-off-by: Jia-Ju Bai <baijiaju1990@163.com>

WBR, Sergei

[PATCH v2 1/2] 8139too in linux-3.18.0: Fix the lack of pci_disable_device

[Jia-Ju Bai]

When pci_request_regions is failed in rtl8139_init_board, 
pci_disable_device is not called to disable the device which are enabled 
by pci_enable_device, because disable_dev_on_err is not assigned 1.
This patch fix this problem.

Signed-off-by: Jia-Ju Bai <baijiaju1990@163.com>
---
drivers/net/ethernet/realtek/8139too.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/realtek/8139too.c 
b/drivers/net/ethernet/realtek/8139too.c
index 007b38c..49bbcf3 100644
--- a/drivers/net/ethernet/realtek/8139too.c
+++ b/drivers/net/ethernet/realtek/8139too.c
@@ -783,10 +783,10 @@ static struct net_device 
*rtl8139_init_board(struct pci_dev *pdev)
if (rc)
goto err_out;

+ disable_dev_on_err = 1;
rc = pci_request_regions (pdev, DRV_NAME);
if (rc)
goto err_out;
- disable_dev_on_err = 1;

pci_set_master (pdev);

Re: [PATCH v2 1/2] 8139too in linux-3.18.0: Fix the lack of pci_disable_device

[David Miller]

From: Jia-Ju Bai <baijiaju1990@163.com>
Date: Sun, 21 Dec 2014 15:36:43 +0800

> When pci_request_regions is failed in rtl8139_init_board,
> pci_disable_device is not called to disable the device which are
> enabled by pci_enable_device, because disable_dev_on_err is not
> assigned 1.
> This patch fix this problem.
> 
> Signed-off-by: Jia-Ju Bai <baijiaju1990@163.com>

Your patches are severely corrupted by your email client, and thus
are not usable.