ibmveth: Use WARN_ON with error handling rather than BUG_ON
PATCH 1/1 ibmveth: Use WARN_ON with error handling rather than BUG_ON
- Replaced BUG_ON calls with WARN_ON calls with error handling, with calls to a new ibmveth_reset routine, which resets the device.
- Added KUnit tests for ibmveth_remove_buffer_from_pool and ibmveth_rxq_get_buffer under new IBMVETH_KUNIT_TEST config option.
- Removed unneeded forward declaration of ibmveth_rxq_harvest_buffer.
--- drivers/net/ethernet/ibm/Kconfig | 13 ++ drivers/net/ethernet/ibm/ibmveth.c | 242 ++++++++++++++++++++++++++--- drivers/net/ethernet/ibm/ibmveth.h | 65 ++++---- 3 files changed, 269 insertions(+), 51 deletions(-)
drivers/net/ethernet/ibm/Kconfig +13 -0
@@ -27,6 +27,19 @@ config IBMVETH To compile this driver as a module, choose M here. The module will be called ibmveth. +config IBMVETH_KUNIT_TEST + bool "KUnit test for IBM LAN Virtual Ethernet support" if !KUNIT_ALL_TESTS + depends on KUNIT + depends on KUNIT=y && IBMVETH=y + default KUNIT_ALL_TESTS + help + This builds unit tests for the IBM LAN Virtual Ethernet driver. + + For more information on KUnit and unit tests in general, please refer + to the KUnit documentation in Documentation/dev-tools/kunit/. + + If unsure, say N. + source "drivers/net/ethernet/ibm/emac/Kconfig" config EHEA
drivers/net/ethernet/ibm/ibmveth.c +223 -19
352 lines not on this page · raw message
drivers/net/ethernet/ibm/ibmveth.h +33 -32
71 lines not on this page · raw message
Signature · 1 line
2.49.0
Quoted 5 lines · Dave Marquardt
- Replaced
BUG_ONcalls withWARN_ONcalls with error handling, with calls to a newibmveth_resetroutine, which resets the device.
- Added KUnit tests foribmveth_remove_buffer_from_poolandibmveth_rxq_get_bufferunder newIBMVETH_KUNIT_TESTconfig option.
- Removed unneeded forward declaration ofibmveth_rxq_harvest_buffer.
It will be great if you split this patch into 3 patches according to your description.
Quoted 54 lines
Signed-off-by Dave Marquardt <[email protected]>--- drivers/net/ethernet/ibm/Kconfig | 13 ++ drivers/net/ethernet/ibm/ibmveth.c | 242 ++++++++++++++++++++++++++--- drivers/net/ethernet/ibm/ibmveth.h | 65 ++++---- 3 files changed, 269 insertions(+), 51 deletions(-)drivers/net/ethernet/ibm/Kconfig +13
-0@@ -27,6 +27,19 @@ config IBMVETH To compile this driver as a module, choose M here. The module will be called ibmveth. +config IBMVETH_KUNIT_TEST + bool "KUnit test for IBM LAN Virtual Ethernet support" if !KUNIT_ALL_TESTS + depends on KUNIT + depends on KUNIT=y && IBMVETH=y + default KUNIT_ALL_TESTS + help + This builds unit tests for the IBM LAN Virtual Ethernet driver. + + For more information on KUnit and unit tests in general, please refer + to the KUnit documentation in Documentation/dev-tools/kunit/. + + If unsure, say N. + source "drivers/net/ethernet/ibm/emac/Kconfig" config EHEAdrivers/net/ethernet/ibm/ibmveth.c +1
-2@@ -28,6 +28,7 @@#include <linux/ip.h> #include <linux/ipv6.h> #include <linux/slab.h> +#include <linux/workqueue.h> #include <asm/hvcall.h> #include <linux/atomic.h> #include <asm/vio.h> @@ -39,8 +40,6 @@#include "ibmveth.h" static irqreturn_t ibmveth_interrupt (int irq,void *dev_instance); -static void ibmveth_rxq_harvest_buffer (struct ibmveth_adapter *adapter, -bool reuse);static unsigned long ibmveth_get_desired_dma (struct vio_dev *vdev);static struct kobj_type ktype_veth_pool;
drivers/net/ethernet/ibm/ibmveth.c static void ibmveth_replenish_buffer_pool(struct ibmveth_adapter *adapter, · line 234
Quoted 6 lines
drivers/net/ethernet/ibm/ibmveth.c +2
-1@@ -231,7 +230,10 @@ static void ibmveth_replenish_buffer_pool(struct ibmveth_adapter *adapter, index = pool->free_map[free_index]; skb =NULL ; -BUG_ON (index ==IBM_VETH_INVALID_MAP ); +if (WARN_ON (index ==IBM_VETH_INVALID_MAP )) { + (void )schedule_work (&adapter->work);
What is the purpose of void casting here (and in other places in this patch)?
drivers/net/ethernet/ibm/ibmveth.c static void ibmveth_replenish_buffer_pool(struct ibmveth_adapter *adapter, · line 235
Quoted 1 line
drivers/net/ethernet/ibm/ibmveth.c +1
-0+ goto failure2;
Maybe increment_buffer_failure, or sth that is telling what happen after goto.
Quoted 16 lines
drivers/net/ethernet/ibm/ibmveth.c +2
-0+ }/* are we allocating a new buffer or recycling an old one */ if (pool->skbuff[index])@@ -300,6 +302,7 @@ static void ibmveth_replenish_buffer_pool(struct ibmveth_adapter *adapter, DMA_FROM_DEVICE);dev_kfree_skb_any (pool->skbuff[index]); pool->skbuff[index] =NULL ; +failure2: adapter->replenish_add_buff_failure++;mb ();@@ -370,20 +373,36 @@ static void ibmveth_free_buffer_pool(struct ibmveth_adapter *adapter, } }
[...]
Quoted 9 lines · Michal Swiatkowski
Quoted 5 lines · Dave Marquardt
- Replaced
BUG_ONcalls withWARN_ONcalls with error handling, with calls to a newibmveth_resetroutine, which resets the device.
- Added KUnit tests foribmveth_remove_buffer_from_poolandibmveth_rxq_get_bufferunder newIBMVETH_KUNIT_TESTconfig option.
- Removed unneeded forward declaration ofibmveth_rxq_harvest_buffer.It will be great if you split this patch into 3 patches according to your description.
Thanks. I debated the right approach here. Thanks for the guidance.
Quoted 11 lines
drivers/net/ethernet/ibm/ibmveth.c static void ibmveth_replenish_buffer_pool(struct ibmveth_adapter *adapter, · line 234
Quoted 8 lines
static struct
kobj_typektype_veth_pool;drivers/net/ethernet/ibm/ibmveth.c +2
-1@@ -231,7 +230,10 @@ static void ibmveth_replenish_buffer_pool(struct ibmveth_adapter *adapter, index = pool->free_map[free_index]; skb =NULL ; -BUG_ON (index ==IBM_VETH_INVALID_MAP ); +if (WARN_ON (index ==IBM_VETH_INVALID_MAP )) { + (void )schedule_work (&adapter->work);What is the purpose of void casting here (and in other places in this patch)?
I'm indicating that I'm ignoring the bool returned by schedule_work(). Since this seemed odd to you, I take it the convention in Linux code is not doing this.
Quoted 4 lines
drivers/net/ethernet/ibm/ibmveth.c static void ibmveth_replenish_buffer_pool(struct ibmveth_adapter *adapter, · line 235
Quoted 1 line
drivers/net/ethernet/ibm/ibmveth.c +1
-0+ goto failure2;Maybe
increment_buffer_failure, or sth that is telling what happen after goto.
Okay, I can change that.
Quoted 11 lines
Quoted 11 lines
drivers/net/ethernet/ibm/ibmveth.c +2
-0+ }/* are we allocating a new buffer or recycling an old one */ if (pool->skbuff[index])@@ -300,6 +302,7 @@ static void ibmveth_replenish_buffer_pool(struct ibmveth_adapter *adapter, DMA_FROM_DEVICE);dev_kfree_skb_any (pool->skbuff[index]); pool->skbuff[index] =NULL ; +failure2: adapter->replenish_add_buff_failure++;
Quoted 7 lines
Quoted 5 lines
(quoted diff) +0
-0mb();@@ -370,20 +373,36 @@ static void ibmveth_free_buffer_pool(struct ibmveth_adapter *adapter, } }[...]
Thanks for your review!
-Dave