Message ID | 20180926125356.2026635-1-arnd@arndb.de |
---|---|
State | New |
Headers | show |
Series | mt76: fix building without CONFIG_MT76x0U | expand |
On Wed, Sep 26, 2018 at 02:51:59PM +0200, Arnd Bergmann wrote: > The recent rework of the mt76 driver caused build failures in > configurations that leave the mt76x0u support disabled: > > ERROR: "mt76u_mcu_deinit" [drivers/net/wireless/mediatek/mt76/mt76x0/mt76x0-common.ko] undefined! > ERROR: "mt76x02u_set_txinfo" [drivers/net/wireless/mediatek/mt76/mt76x0/mt76x0-common.ko] undefined! > ERROR: "mt76u_queues_deinit" [drivers/net/wireless/mediatek/mt76/mt76x0/mt76x0-common.ko] undefined! > ERROR: "mt76u_stop_stat_wk" [drivers/net/wireless/mediatek/mt76/mt76x0/mt76x0-common.ko] undefined! > > The mt76x0_push_txwi()/mt76x0_tx_prepare_skb()/mt76x0_cleanup() functions > that cause some of these are only called from the usb portion, and can > be hidden in an #ifdef in this case. > > mt76u_stop_stat_wk() is called mt76x0_mac_stop(), which is in turn > shared between multiple callers. Calling it only when the USB driver > is enabled avoids the link error but it is not clear to me whether this > can be called from a context where it would not do the right thing. I think there not yet posted Lorenzo patches that fix this build problem. Regards Stanislaw
On Wed, Sep 26, 2018 at 4:06 PM Stanislaw Gruszka <sgruszka@redhat.com> wrote: > > On Wed, Sep 26, 2018 at 02:51:59PM +0200, Arnd Bergmann wrote: > > The recent rework of the mt76 driver caused build failures in > > configurations that leave the mt76x0u support disabled: > > > > ERROR: "mt76u_mcu_deinit" [drivers/net/wireless/mediatek/mt76/mt76x0/mt76x0-common.ko] undefined! > > ERROR: "mt76x02u_set_txinfo" [drivers/net/wireless/mediatek/mt76/mt76x0/mt76x0-common.ko] undefined! > > ERROR: "mt76u_queues_deinit" [drivers/net/wireless/mediatek/mt76/mt76x0/mt76x0-common.ko] undefined! > > ERROR: "mt76u_stop_stat_wk" [drivers/net/wireless/mediatek/mt76/mt76x0/mt76x0-common.ko] undefined! > > > > The mt76x0_push_txwi()/mt76x0_tx_prepare_skb()/mt76x0_cleanup() functions > > that cause some of these are only called from the usb portion, and can > > be hidden in an #ifdef in this case. > > > > mt76u_stop_stat_wk() is called mt76x0_mac_stop(), which is in turn > > shared between multiple callers. Calling it only when the USB driver > > is enabled avoids the link error but it is not clear to me whether this > > can be called from a context where it would not do the right thing. > > I think there not yet posted Lorenzo patches that fix this build > problem. Ok, good. I assume that is a better fix than mine then. I also noticed another problem that my patch does not address, but won't send a follow up patch for that. Please just ignore this version. Arnd
Hi Arnd, I love your patch! Yet something to improve: [auto build test ERROR on wireless-drivers-next/master] [also build test ERROR on next-20180926] [cannot apply to v4.19-rc5] [if your patch is applied to the wrong git tree, please drop us a note to help improve the system] url: https://github.com/0day-ci/linux/commits/Arnd-Bergmann/mt76-fix-building-without-CONFIG_MT76x0U/20180927-101346 base: https://git.kernel.org/pub/scm/linux/kernel/git/kvalo/wireless-drivers-next.git master config: i386-allmodconfig (attached as .config) compiler: gcc-7 (Debian 7.3.0-1) 7.3.0 reproduce: # save the attached .config to linux build tree make ARCH=i386 All errors (new ones prefixed by >>): >> ERROR: "mt76x0_tx_prepare_skb" [drivers/net/wireless/mediatek/mt76/mt76x0/mt76x0u.ko] undefined! >> ERROR: "mt76x0_cleanup" [drivers/net/wireless/mediatek/mt76/mt76x0/mt76x0u.ko] undefined! --- 0-DAY kernel test infrastructure Open Source Technology Center https://lists.01.org/pipermail/kbuild-all Intel Corporation
diff --git a/drivers/net/wireless/mediatek/mt76/mt76x0/init.c b/drivers/net/wireless/mediatek/mt76/mt76x0/init.c index 3a88be267daf..9d0f0e1bf07a 100644 --- a/drivers/net/wireless/mediatek/mt76/mt76x0/init.c +++ b/drivers/net/wireless/mediatek/mt76/mt76x0/init.c @@ -377,7 +377,8 @@ void mt76x0_mac_stop(struct mt76x0_dev *dev) { cancel_delayed_work_sync(&dev->cal_work); cancel_delayed_work_sync(&dev->mac_work); - mt76u_stop_stat_wk(&dev->mt76); + if (IS_ENABLED(CONFIG_MT76x0U)) + mt76u_stop_stat_wk(&dev->mt76); mt76x0_mac_stop_hw(dev); } EXPORT_SYMBOL_GPL(mt76x0_mac_stop); @@ -457,6 +458,7 @@ int mt76x0_init_hardware(struct mt76x0_dev *dev) } EXPORT_SYMBOL_GPL(mt76x0_init_hardware); +#ifdef CONFIG_MT76x0U void mt76x0_cleanup(struct mt76x0_dev *dev) { clear_bit(MT76_STATE_INITIALIZED, &dev->mt76.state); @@ -465,6 +467,7 @@ void mt76x0_cleanup(struct mt76x0_dev *dev) mt76u_mcu_deinit(&dev->mt76); } EXPORT_SYMBOL_GPL(mt76x0_cleanup); +#endif struct mt76x0_dev * mt76x0_alloc_device(struct device *pdev, const struct mt76_driver_ops *drv_ops) diff --git a/drivers/net/wireless/mediatek/mt76/mt76x0/tx.c b/drivers/net/wireless/mediatek/mt76/mt76x0/tx.c index c6d8ba01feb1..2d3c20e82ce1 100644 --- a/drivers/net/wireless/mediatek/mt76/mt76x0/tx.c +++ b/drivers/net/wireless/mediatek/mt76/mt76x0/tx.c @@ -17,6 +17,7 @@ #include "../mt76x02_util.h" #include "../mt76x02_usb.h" +#ifdef CONFIG_MT76x0U static struct mt76x02_txwi * mt76x0_push_txwi(struct mt76x0_dev *dev, struct sk_buff *skb, struct ieee80211_sta *sta, struct mt76_wcid *wcid, @@ -53,6 +54,7 @@ mt76x0_push_txwi(struct mt76x0_dev *dev, struct sk_buff *skb, return txwi; } +#endif void mt76x0_tx(struct ieee80211_hw *hw, struct ieee80211_tx_control *control, struct sk_buff *skb) @@ -82,6 +84,7 @@ void mt76x0_tx(struct ieee80211_hw *hw, struct ieee80211_tx_control *control, mt76_tx(&dev->mt76, control->sta, wcid, skb); } +#ifdef CONFIG_MT76x0U int mt76x0_tx_prepare_skb(struct mt76_dev *mdev, void *data, struct sk_buff *skb, struct mt76_queue *q, struct mt76_wcid *wcid, struct ieee80211_sta *sta, @@ -97,6 +100,7 @@ int mt76x0_tx_prepare_skb(struct mt76_dev *mdev, void *data, return mt76x02u_set_txinfo(skb, wcid, q2ep(q->hw_idx)); } EXPORT_SYMBOL_GPL(mt76x0_tx_prepare_skb); +#endif void mt76x0_queue_rx_skb(struct mt76_dev *mdev, enum mt76_rxq_id q, struct sk_buff *skb)
The recent rework of the mt76 driver caused build failures in configurations that leave the mt76x0u support disabled: ERROR: "mt76u_mcu_deinit" [drivers/net/wireless/mediatek/mt76/mt76x0/mt76x0-common.ko] undefined! ERROR: "mt76x02u_set_txinfo" [drivers/net/wireless/mediatek/mt76/mt76x0/mt76x0-common.ko] undefined! ERROR: "mt76u_queues_deinit" [drivers/net/wireless/mediatek/mt76/mt76x0/mt76x0-common.ko] undefined! ERROR: "mt76u_stop_stat_wk" [drivers/net/wireless/mediatek/mt76/mt76x0/mt76x0-common.ko] undefined! The mt76x0_push_txwi()/mt76x0_tx_prepare_skb()/mt76x0_cleanup() functions that cause some of these are only called from the usb portion, and can be hidden in an #ifdef in this case. mt76u_stop_stat_wk() is called mt76x0_mac_stop(), which is in turn shared between multiple callers. Calling it only when the USB driver is enabled avoids the link error but it is not clear to me whether this can be called from a context where it would not do the right thing. Signed-off-by: Arnd Bergmann <arnd@arndb.de> --- drivers/net/wireless/mediatek/mt76/mt76x0/init.c | 5 ++++- drivers/net/wireless/mediatek/mt76/mt76x0/tx.c | 4 ++++ 2 files changed, 8 insertions(+), 1 deletion(-) -- 2.18.0