scsi: qla2xxx: Fix crash in PCIe error handling

BUG: unable to handle kernel NULL pointer dereference at (null)
IP: qla2x00_abort_isp+0x21/0x6b0 [qla2xxx] PGD 0 P4D 0
Oops: 0000 [#1] SMP PTI
CPU: 0 PID: 1715 Comm: kworker/0:2
Tainted: GOE 4.12.14-122.37-default #1 SLE12-SP5
Hardware name: HPE Superdome Flex/Superdome Flex, BIOS
Bundle:3.30.100 SFW:IP147.007.004.017.000.2009211957 09/21/2020
Workqueue: events aer_recover_work_func
task: ffff9e399c14ca80 task.stack: ffffc1c58e4ac000
RIP: 0010:qla2x00_abort_isp+0x21/0x6b0 [qla2xxx]
RSP: 0018:ffffc1c58e4afd50 EFLAGS: 00010282
RAX: 0000000000000000 RBX: ffff9e419cdef480 RCX: 0000000000000000
RDX: ffff9e399c14ca80 RSI: 0000000000000246 RDI: ffff9e419bbc27b8
RBP: ffff9e419bbc27b8 R08: 0000000000000004 R09: 00000000a0440000
R10: 0000000000000000 R11: ffff9e399416d1a0 R12: ffff9e419cdef000
R13: ffff9e3a7cfae800 R14: ffff9e3a7cfae800 R15: 00000000000000c0
FS:  0000000000000000(0000) GS:ffff9e39a0000000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000000000000 CR3: 00000006cd00a005 CR4: 00000000007606f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
PKRU: 55555554
Call Trace:
  qla2xxx_pci_slot_reset+0x141/0x160 [qla2xxx]
  report_slot_reset+0x41/0x80
  ? merge_result.part.4+0x30/0x30
  pci_walk_bus+0x70/0x90
  pcie_do_recovery+0x1db/0x2e0
  aer_recover_work_func+0xc2/0xf0
  process_one_work+0x14c/0x390

Disable board_disable logic where driver resources are freed while OS is in
the process of recovering the adapter.

Link: https://lore.kernel.org/r/20210329085229.4367-9-njavali@marvell.com
Tested-by: Laurence Oberman <loberman@redhat.com>
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

1 files changed, 15 insertions, 1 deletions

diff --git a/drivers/scsi/qla2xxx/qla_dbg.c b/drivers/scsi/qla2xxx/qla_dbg.c
index 144a893e7335..f2d05592c1e2 100644
--- a/drivers/scsi/qla2xxx/qla_dbg.c
+++ b/drivers/scsi/qla2xxx/qla_dbg.c
@@ -113,8 +113,13 @@ qla27xx_dump_mpi_ram(struct qla_hw_data *ha, uint32_t addr, uint32_t *ram,
 	uint32_t stat;
 	ulong i, j, timer = 6000000;
 	int rval = QLA_FUNCTION_FAILED;
+	scsi_qla_host_t *vha = pci_get_drvdata(ha->pdev);
 
 	clear_bit(MBX_INTERRUPT, &ha->mbx_cmd_flags);
+
+	if (qla_pci_disconnected(vha, reg))
+		return rval;
+
 	for (i = 0; i < ram_dwords; i += dwords, addr += dwords) {
 		if (i + dwords > ram_dwords)
 			dwords = ram_dwords - i;
@@ -138,6 +143,9 @@ qla27xx_dump_mpi_ram(struct qla_hw_data *ha, uint32_t addr, uint32_t *ram,
 		while (timer--) {
 			udelay(5);
 
+			if (qla_pci_disconnected(vha, reg))
+				return rval;
+
 			stat = rd_reg_dword(&reg->host_status);
 			/* Check for pending interrupts. */
 			if (!(stat & HSRX_RISC_INT))
@@ -192,9 +200,13 @@ qla24xx_dump_ram(struct qla_hw_data *ha, uint32_t addr, __be32 *ram,
 	uint32_t dwords = qla2x00_gid_list_size(ha) / 4;
 	uint32_t stat;
 	ulong i, j, timer = 6000000;
+	scsi_qla_host_t *vha = pci_get_drvdata(ha->pdev);
 
 	clear_bit(MBX_INTERRUPT, &ha->mbx_cmd_flags);
 
+	if (qla_pci_disconnected(vha, reg))
+		return rval;
+
 	for (i = 0; i < ram_dwords; i += dwords, addr += dwords) {
 		if (i + dwords > ram_dwords)
 			dwords = ram_dwords - i;
@@ -216,8 +228,10 @@ qla24xx_dump_ram(struct qla_hw_data *ha, uint32_t addr, __be32 *ram,
 		ha->flags.mbox_int = 0;
 		while (timer--) {
 			udelay(5);
-			stat = rd_reg_dword(&reg->host_status);
+			if (qla_pci_disconnected(vha, reg))
+				return rval;
 
+			stat = rd_reg_dword(&reg->host_status);
 			/* Check for pending interrupts. */
 			if (!(stat & HSRX_RISC_INT))
 				continue;