From: Julian Wiedmann <jwi@xxxxxxxxxxxxx> Explain why the plain spin_lock() suffices in current code, even when the stat_lock is also used by zfcp_qdio_int_req() in tasklet context. We could also promote the spin_lock() to a spin_lock_irqsave(), but that would just obfuscate the locking even further. Signed-off-by: Julian Wiedmann <jwi@xxxxxxxxxxxxx> Reviewed-by: Benjamin Block <bblock@xxxxxxxxxxxxx> Reviewed-by: Steffen Maier <maier@xxxxxxxxxxxxx> Signed-off-by: Benjamin Block <bblock@xxxxxxxxxxxxx> --- drivers/s390/scsi/zfcp_qdio.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/drivers/s390/scsi/zfcp_qdio.c b/drivers/s390/scsi/zfcp_qdio.c index 9fc045ddf66d..23ab16d65f2a 100644 --- a/drivers/s390/scsi/zfcp_qdio.c +++ b/drivers/s390/scsi/zfcp_qdio.c @@ -10,6 +10,7 @@ #define KMSG_COMPONENT "zfcp" #define pr_fmt(fmt) KMSG_COMPONENT ": " fmt +#include <linux/lockdep.h> #include <linux/slab.h> #include <linux/module.h> #include "zfcp_ext.h" @@ -283,6 +284,13 @@ int zfcp_qdio_send(struct zfcp_qdio *qdio, struct zfcp_qdio_req *q_req) int retval; u8 sbal_number = q_req->sbal_number; + /* + * This should actually be a spin_lock_bh(stat_lock), to protect against + * zfcp_qdio_int_req() in tasklet context. + * But we can't do so (and are safe), as we always get called with IRQs + * disabled by spin_lock_irq[save](req_q_lock). + */ + lockdep_assert_irqs_disabled(); spin_lock(&qdio->stat_lock); zfcp_qdio_account(qdio); spin_unlock(&qdio->stat_lock); -- 2.26.2
