4.9-stable review patch. If anyone has any objections, please let me know.
------------------
From: Bryant G. Ly <bryan...@linux.vnet.ibm.com>
commit a5b0e4062fb225155189e593699bbfcd0597f8b5 upstream.
Currently, dma_alloc_coherent is being called with a GFP_KERNEL
flag which allows it to sleep in an interrupt context, need to
change to GFP_ATOMIC.
Tested-by: Steven Royer <sero...@linux.vnet.ibm.com>
Reviewed-by: Michael Cyr <mike...@linux.vnet.ibm.com>
Signed-off-by: Bryant G. Ly <bryan...@linux.vnet.ibm.com>
Signed-off-by: Bart Van Assche <bart.vanass...@sandisk.com>
Signed-off-by: Greg Kroah-Hartman <gre...@linuxfoundation.org>
---
drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
--- a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
+++ b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c
@@ -1239,7 +1239,7 @@ static long ibmvscsis_adapter_info(struc
}
info = dma_alloc_coherent(&vscsi->dma_dev->dev, sizeof(*info), &token,
- GFP_KERNEL);
+ GFP_ATOMIC);
if (!info) {
dev_err(&vscsi->dev, "bad dma_alloc_coherent %p\n",
iue->target);
@@ -1357,7 +1357,7 @@ static int ibmvscsis_cap_mad(struct scsi
}
cap = dma_alloc_coherent(&vscsi->dma_dev->dev, olen, &token,
- GFP_KERNEL);
+ GFP_ATOMIC);
if (!cap) {
dev_err(&vscsi->dev, "bad dma_alloc_coherent %p\n",
iue->target);
