On 4/12/20 5:32 AM, James Smart wrote:
This patch continues the efct driver population.
This patch adds driver definitions for:
Routines to detach xport and hardware objects.
Signed-off-by: Ram Vegesna <ram.vegesna@xxxxxxxxxxxx>
Signed-off-by: James Smart <jsmart2021@xxxxxxxxx>
---
v3:
Removed old patch 28 and merged with 27
---
drivers/scsi/elx/efct/efct_hw.c | 333 +++++++++++++++++++++++++++++++++++++
drivers/scsi/elx/efct/efct_hw.h | 31 ++++
drivers/scsi/elx/efct/efct_xport.c | 291 ++++++++++++++++++++++++++++++++
3 files changed, 655 insertions(+)
diff --git a/drivers/scsi/elx/efct/efct_hw.c b/drivers/scsi/elx/efct/efct_hw.c
index ca2fd237c7d6..a007ca98895d 100644
--- a/drivers/scsi/elx/efct/efct_hw.c
+++ b/drivers/scsi/elx/efct/efct_hw.c
@@ -3503,3 +3503,336 @@ efct_hw_get_host_stats(struct efct_hw *hw, u8 cc,
return rc;
}
+
+static int
+efct_hw_cb_port_control(struct efct_hw *hw, int status, u8 *mqe,
+ void *arg)
+{
+ kfree(mqe);
+ return EFC_SUCCESS;
+}
+
+/* Control a port (initialize, shutdown, or set link configuration) */
+enum efct_hw_rtn
+efct_hw_port_control(struct efct_hw *hw, enum efct_hw_port ctrl,
+ uintptr_t value,
+ void (*cb)(int status, uintptr_t value, void *arg),
+ void *arg)
+{
+ enum efct_hw_rtn rc = EFCT_HW_RTN_ERROR;
+
+ switch (ctrl) {
+ case EFCT_HW_PORT_INIT:
+ {
+ u8 *init_link;
+ u32 speed = 0;
+ u8 reset_alpa = 0;
+
+ u8 *cfg_link;
+
+ cfg_link = kmalloc(SLI4_BMBX_SIZE, GFP_ATOMIC);
+ if (!cfg_link)
+ return EFCT_HW_RTN_NO_MEMORY;
+
Use the predefined mailbox buffer?
Or a memory pool?
+ if (!sli_cmd_config_link(&hw->sli, cfg_link,
+ SLI4_BMBX_SIZE))
+ rc = efct_hw_command(hw, cfg_link,
+ EFCT_CMD_NOWAIT,
+ efct_hw_cb_port_control,
+ NULL);
+
+ if (rc != EFCT_HW_RTN_SUCCESS) {
+ kfree(cfg_link);
+ efc_log_err(hw->os, "CONFIG_LINK failed\n");
+ break;
+ }
+ speed = hw->config.speed;
+ reset_alpa = (u8)(value & 0xff);
+
+ /* Allocate a new buffer for the init_link command */
+ init_link = kmalloc(SLI4_BMBX_SIZE, GFP_ATOMIC);
+ if (!init_link)
+ return EFCT_HW_RTN_NO_MEMORY;
+
+ rc = EFCT_HW_RTN_ERROR;
+ if (!sli_cmd_init_link(&hw->sli, init_link, SLI4_BMBX_SIZE,
+ speed, reset_alpa))
+ rc = efct_hw_command(hw, init_link, EFCT_CMD_NOWAIT,
+ efct_hw_cb_port_control, NULL);
+ /* Free buffer on error, since no callback is coming */
+ if (rc != EFCT_HW_RTN_SUCCESS) {
+ kfree(init_link);
+ efc_log_err(hw->os, "INIT_LINK failed\n");
+ }
+ break;
+ }
+ case EFCT_HW_PORT_SHUTDOWN:
+ {
+ u8 *down_link;
+
+ down_link = kmalloc(SLI4_BMBX_SIZE, GFP_ATOMIC);
+ if (!down_link)
+ return EFCT_HW_RTN_NO_MEMORY;
+
Same here.
+ if (!sli_cmd_down_link(&hw->sli, down_link, SLI4_BMBX_SIZE))
+ rc = efct_hw_command(hw, down_link, EFCT_CMD_NOWAIT,
+ efct_hw_cb_port_control, NULL);
+ /* Free buffer on error, since no callback is coming */
+ if (rc != EFCT_HW_RTN_SUCCESS) {
+ kfree(down_link);
+ efc_log_err(hw->os, "DOWN_LINK failed\n");
+ }
+ break;
+ }
+ default:
+ efc_log_test(hw->os, "unhandled control %#x\n", ctrl);
+ break;
+ }
+
+ return rc;
+}
+
+enum efct_hw_rtn
+efct_hw_teardown(struct efct_hw *hw)
+{
+ u32 i = 0;
+ u32 iters = 10;
+ u32 max_rpi;
+ u32 destroy_queues;
+ u32 free_memory;
+ struct efc_dma *dma;
+ struct efct *efct = hw->os;
+
+ destroy_queues = (hw->state == EFCT_HW_STATE_ACTIVE);
+ free_memory = (hw->state != EFCT_HW_STATE_UNINITIALIZED);
+
+ /* Cancel Sliport Healthcheck */
+ if (hw->sliport_healthcheck) {
+ hw->sliport_healthcheck = 0;
+ efct_hw_config_sli_port_health_check(hw, 0, 0);
+ }
+
+ if (hw->state != EFCT_HW_STATE_QUEUES_ALLOCATED) {
+ hw->state = EFCT_HW_STATE_TEARDOWN_IN_PROGRESS;
+
+ efct_hw_flush(hw);
+
+ /*
+ * If there are outstanding commands, wait for them to complete
+ */
+ while (!list_empty(&hw->cmd_head) && iters) {
+ mdelay(10);
+ efct_hw_flush(hw);
+ iters--;
+ }
+
+ if (list_empty(&hw->cmd_head))
+ efc_log_debug(hw->os,
+ "All commands completed on MQ queue\n");
+ else
+ efc_log_debug(hw->os,
+ "Some cmds still pending on MQ queue\n");
+
+ /* Cancel any remaining commands */
+ efct_hw_command_cancel(hw);
+ } else {
+ hw->state = EFCT_HW_STATE_TEARDOWN_IN_PROGRESS;
+ }
+
+ max_rpi = hw->sli.qinfo.max_qcount[SLI_RSRC_RPI];
+ if (hw->rpi_ref) {
+ for (i = 0; i < max_rpi; i++) {
+ u32 count;
+
+ count = atomic_read(&hw->rpi_ref[i].rpi_count);
+ if (count)
+ efc_log_debug(hw->os,
+ "non-zero ref [%d]=%d\n",
+ i, count);
Ho-hum. So you have a non-zero refcount and _still_ free the structure?
That smells fishy ...
+ }
+ kfree(hw->rpi_ref);
+ hw->rpi_ref = NULL;
+ }
+
Please use proper refcounting here.
Cheers,
Hannes
--
Dr. Hannes Reinecke Teamlead Storage & Networking
hare@xxxxxxx +49 911 74053 688
SUSE Software Solutions GmbH, Maxfeldstr. 5, 90409 Nürnberg
HRB 36809 (AG Nürnberg), Geschäftsführer: Felix Imendörffer
