Don't require FIRMWAREbDownload() to, first off, unlock a held lock. Thus do all locking in main_usb.c and hold it for a insignificantly shorter period of time. This makes the affected area significantly more readable though. Signed-off-by: Martin Kepplinger <martink@xxxxxxxxx> --- I resend https://lkml.org/lkml/2014/4/30/453 and add devel@xxxxxxxxxxxxxxxxxxxx after waiting a week or so. This fixes a sparse warning too, and uses less spin_lock functions. Maybe someone can think it through. drivers/staging/vt6656/firmware.c | 2 -- drivers/staging/vt6656/main_usb.c | 5 ++++- 2 files changed, 4 insertions(+), 3 deletions(-) diff --git a/drivers/staging/vt6656/firmware.c b/drivers/staging/vt6656/firmware.c index cd2ea76..79129ad 100644 --- a/drivers/staging/vt6656/firmware.c +++ b/drivers/staging/vt6656/firmware.c @@ -54,7 +54,6 @@ int FIRMWAREbDownload(struct vnt_private *pDevice) int ii, rc; DBG_PRT(MSG_LEVEL_DEBUG, KERN_INFO"---->Download firmware\n"); - spin_unlock_irq(&pDevice->lock); rc = request_firmware(&fw, FIRMWARE_NAME, dev); if (rc) { @@ -91,7 +90,6 @@ free_fw: out: kfree(pBuffer); - spin_lock_irq(&pDevice->lock); return result; } MODULE_FIRMWARE(FIRMWARE_NAME); diff --git a/drivers/staging/vt6656/main_usb.c b/drivers/staging/vt6656/main_usb.c index 3c93230..1dc02c4 100644 --- a/drivers/staging/vt6656/main_usb.c +++ b/drivers/staging/vt6656/main_usb.c @@ -319,7 +319,11 @@ static int device_init_registers(struct vnt_private *pDevice) memcpy(pDevice->abySNAP_Bridgetunnel, abySNAP_Bridgetunnel, ETH_ALEN); if (!FIRMWAREbCheckVersion(pDevice)) { + + spin_unlock_irq(&pDevice->lock); if (FIRMWAREbDownload(pDevice) == true) { + + spin_lock_irq(&pDevice->lock); if (FIRMWAREbBrach2Sram(pDevice) == false) { DBG_PRT(MSG_LEVEL_DEBUG, KERN_INFO " FIRMWAREbBrach2Sram fail\n"); @@ -329,7 +333,6 @@ static int device_init_registers(struct vnt_private *pDevice) } else { DBG_PRT(MSG_LEVEL_DEBUG, KERN_INFO " FIRMWAREbDownload fail\n"); - spin_unlock_irq(&pDevice->lock); return false; } } -- 1.7.10.4 _______________________________________________ devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxx http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
