* [PATCH v3 01/10] mpi3mr: Skip device shutdown during unload per controller configuration
2026-07-24 10:24 [PATCH v3 00/10] mpi3mr: Few Enhancements and minor fixes Ranjan Kumar
@ 2026-07-24 10:24 ` Ranjan Kumar
2026-07-24 10:24 ` [PATCH v3 02/10] mpi3mr: Update MPI Headers to revision 41 Ranjan Kumar
` (8 subsequent siblings)
9 siblings, 0 replies; 14+ messages in thread
From: Ranjan Kumar @ 2026-07-24 10:24 UTC (permalink / raw)
To: linux-scsi, martin.petersen
Cc: sathya.prakash, chandrakanth.patil, vishakhavc, ipylypiv,
Ranjan Kumar, Sashiko
The controller may be configured through Driver Page 1 to suppress
device shutdown requests during driver unload. Cache this setting and
skip the device shutdown request during IOC shutdown when unloading
the driver.
Additionally, ensure the driver_pg1 fields are properly converted
from little-endian to CPU endianness using le32_to_cpu() and le16_to_cpu()
before evaluating the shutdown disable flag and allocating diag buffers.
This prevents failures and massive memory allocation errors on big-endian
architectures.
Reported-by: Sashiko <sashiko-bot@kernel.org>
Closes: https://sashiko.dev/#/patchset/20260626114109.43685-1-ranjan.kumar@broadcom.com?part=1
Closes: https://sashiko.dev/#/patchset/20260708183305.244485-1-ranjan.kumar@broadcom.com?part=1
Signed-off-by: Chandrakanth Patil <chandrakanth.patil@broadcom.com>
Signed-off-by: Ranjan Kumar <ranjan.kumar@broadcom.com>
---
drivers/scsi/mpi3mr/mpi3mr.h | 3 +++
drivers/scsi/mpi3mr/mpi3mr_app.c | 24 ++++++++++++----------
drivers/scsi/mpi3mr/mpi3mr_fw.c | 35 ++++++++++++++++++++++++--------
drivers/scsi/mpi3mr/mpi3mr_os.c | 2 ++
4 files changed, 44 insertions(+), 20 deletions(-)
diff --git a/drivers/scsi/mpi3mr/mpi3mr.h b/drivers/scsi/mpi3mr/mpi3mr.h
index c25525fe0671..39096004c60a 100644
--- a/drivers/scsi/mpi3mr/mpi3mr.h
+++ b/drivers/scsi/mpi3mr/mpi3mr.h
@@ -1410,6 +1410,9 @@ struct mpi3mr_ioc {
struct dma_pool *trace_buf_pool;
struct segments *trace_buf;
u8 invalid_io_comp;
+ bool is_unload;
+ bool skip_dev_shutdown_on_unload;
+
};
diff --git a/drivers/scsi/mpi3mr/mpi3mr_app.c b/drivers/scsi/mpi3mr/mpi3mr_app.c
index 1353a8ff9c85..49243ef86dca 100644
--- a/drivers/scsi/mpi3mr/mpi3mr_app.c
+++ b/drivers/scsi/mpi3mr/mpi3mr_app.c
@@ -141,25 +141,27 @@ void mpi3mr_alloc_diag_bufs(struct mpi3mr_ioc *mrioc)
trace_min_size = fw_min_size = MPI3MR_DEFAULT_HDB_MIN_SZ;
} else {
- trace_size = driver_pg1.host_diag_trace_max_size * 1024;
- trace_dec_size = driver_pg1.host_diag_trace_decrement_size
+ trace_size = le16_to_cpu(driver_pg1.host_diag_trace_max_size) * 1024;
+ trace_dec_size = le16_to_cpu(driver_pg1.host_diag_trace_decrement_size)
* 1024;
trace_min_size = driver_pg1.host_diag_trace_min_size * 1024;
- fw_size = driver_pg1.host_diag_fw_max_size * 1024;
- fw_dec_size = driver_pg1.host_diag_fw_decrement_size * 1024;
- fw_min_size = driver_pg1.host_diag_fw_min_size * 1024;
+ fw_size = le16_to_cpu(driver_pg1.host_diag_fw_max_size) * 1024;
+ fw_dec_size = le16_to_cpu(driver_pg1.host_diag_fw_decrement_size) * 1024;
+ fw_min_size = le16_to_cpu(driver_pg1.host_diag_fw_min_size) * 1024;
dprint_init(mrioc,
"%s:trace diag buffer sizes read from driver\n"
"page1: maximum size = %dKB, decrement size = %dKB\n"
- ", minimum size = %dKB\n", __func__, driver_pg1.host_diag_trace_max_size,
- driver_pg1.host_diag_trace_decrement_size,
- driver_pg1.host_diag_trace_min_size);
+ ", minimum size = %dKB\n", __func__,
+ le16_to_cpu(driver_pg1.host_diag_trace_max_size),
+ le16_to_cpu(driver_pg1.host_diag_trace_decrement_size),
+ le16_to_cpu(driver_pg1.host_diag_trace_min_size));
dprint_init(mrioc,
"%s:firmware diag buffer sizes read from driver\n"
"page1: maximum size = %dKB, decrement size = %dKB\n"
- ", minimum size = %dKB\n", __func__, driver_pg1.host_diag_fw_max_size,
- driver_pg1.host_diag_fw_decrement_size,
- driver_pg1.host_diag_fw_min_size);
+ ", minimum size = %dKB\n", __func__,
+ le16_to_cpu(driver_pg1.host_diag_fw_max_size),
+ le16_to_cpu(driver_pg1.host_diag_fw_decrement_size),
+ le16_to_cpu(driver_pg1.host_diag_fw_min_size));
if ((trace_size == 0) && (fw_size == 0))
return;
}
diff --git a/drivers/scsi/mpi3mr/mpi3mr_fw.c b/drivers/scsi/mpi3mr/mpi3mr_fw.c
index 31b19ed1528e..59241038f689 100644
--- a/drivers/scsi/mpi3mr/mpi3mr_fw.c
+++ b/drivers/scsi/mpi3mr/mpi3mr_fw.c
@@ -4107,26 +4107,35 @@ static int mpi3mr_repost_diag_bufs(struct mpi3mr_ioc *mrioc)
}
/**
- * mpi3mr_read_tsu_interval - Update time stamp interval
+ * mpi3mr_read_driver_page1 - Read Driver Page 1 parameters
* @mrioc: Adapter instance reference
*
- * Update time stamp interval if its defined in driver page 1,
- * otherwise use default value.
+ * Reads and caches Driver Page 1 parameters such as
+ * timestamp update interval and driver behavior flags.
*
* Return: Nothing
*/
static void
-mpi3mr_read_tsu_interval(struct mpi3mr_ioc *mrioc)
+mpi3mr_read_driver_page1(struct mpi3mr_ioc *mrioc)
{
struct mpi3_driver_page1 driver_pg1;
u16 pg_sz = sizeof(driver_pg1);
int retval = 0;
mrioc->ts_update_interval = MPI3MR_TSUPDATE_INTERVAL;
+ mrioc->skip_dev_shutdown_on_unload = 0;
retval = mpi3mr_cfg_get_driver_pg1(mrioc, &driver_pg1, pg_sz);
- if (!retval && driver_pg1.time_stamp_update)
+
+ if (retval)
+ return;
+
+ if (driver_pg1.time_stamp_update)
mrioc->ts_update_interval = (driver_pg1.time_stamp_update * 60);
+
+ mrioc->skip_dev_shutdown_on_unload =
+ (le32_to_cpu(driver_pg1.flags) &
+ MPI3_DRIVER1_FLAGS_DEVICE_SHUTDOWN_ON_UNLOAD_DISABLE) ? 1 : 0;
}
/**
@@ -4432,7 +4441,7 @@ int mpi3mr_init_ioc(struct mpi3mr_ioc *mrioc)
goto out_failed_noretry;
}
- mpi3mr_read_tsu_interval(mrioc);
+ mpi3mr_read_driver_page1(mrioc);
mpi3mr_print_ioc_info(mrioc);
dprint_init(mrioc, "allocating host diag buffers\n");
@@ -4604,7 +4613,7 @@ int mpi3mr_reinit_ioc(struct mpi3mr_ioc *mrioc, u8 is_resume)
goto out_failed_noretry;
}
- mpi3mr_read_tsu_interval(mrioc);
+ mpi3mr_read_driver_page1(mrioc);
mpi3mr_print_ioc_info(mrioc);
if (is_resume) {
@@ -5089,8 +5098,16 @@ static void mpi3mr_issue_ioc_shutdown(struct mpi3mr_ioc *mrioc)
return;
}
- shutdown_action = MPI3_SYSIF_IOC_CONFIG_SHUTDOWN_NORMAL |
- MPI3_SYSIF_IOC_CONFIG_DEVICE_SHUTDOWN_SEND_REQ;
+ shutdown_action = MPI3_SYSIF_IOC_CONFIG_SHUTDOWN_NORMAL;
+
+ if (!(mrioc->is_unload && mrioc->skip_dev_shutdown_on_unload))
+ shutdown_action |=
+ MPI3_SYSIF_IOC_CONFIG_DEVICE_SHUTDOWN_SEND_REQ;
+ else
+ ioc_info(mrioc,
+ "The shutdown request is issued without the device shutdown bit set\n"
+ "as indicated by the controller configuration\n");
+
ioc_config = readl(&mrioc->sysif_regs->ioc_configuration);
ioc_config |= shutdown_action;
diff --git a/drivers/scsi/mpi3mr/mpi3mr_os.c b/drivers/scsi/mpi3mr/mpi3mr_os.c
index 402d1f35d214..d2a20f2721db 100644
--- a/drivers/scsi/mpi3mr/mpi3mr_os.c
+++ b/drivers/scsi/mpi3mr/mpi3mr_os.c
@@ -5665,6 +5665,8 @@ static void mpi3mr_remove(struct pci_dev *pdev)
return;
mrioc = shost_priv(shost);
+ mrioc->is_unload = true;
+
while (mrioc->reset_in_progress || mrioc->is_driver_loading)
ssleep(1);
--
2.47.3
^ permalink raw reply related [flat|nested] 14+ messages in thread* [PATCH v3 02/10] mpi3mr: Update MPI Headers to revision 41
2026-07-24 10:24 [PATCH v3 00/10] mpi3mr: Few Enhancements and minor fixes Ranjan Kumar
2026-07-24 10:24 ` [PATCH v3 01/10] mpi3mr: Skip device shutdown during unload per controller configuration Ranjan Kumar
@ 2026-07-24 10:24 ` Ranjan Kumar
2026-07-24 10:24 ` [PATCH v3 03/10] mpi3mr: Add early timestamp synchronization after driver load Ranjan Kumar
` (7 subsequent siblings)
9 siblings, 0 replies; 14+ messages in thread
From: Ranjan Kumar @ 2026-07-24 10:24 UTC (permalink / raw)
To: linux-scsi, martin.petersen
Cc: sathya.prakash, chandrakanth.patil, vishakhavc, ipylypiv,
Ranjan Kumar
Update MPI Headers to revision 41
Signed-off-by: Ranjan Kumar <ranjan.kumar@broadcom.com>
---
drivers/scsi/mpi3mr/mpi/mpi30_cnfg.h | 77 +++++++++++++++++++++--
drivers/scsi/mpi3mr/mpi/mpi30_image.h | 7 ++-
drivers/scsi/mpi3mr/mpi/mpi30_ioc.h | 15 +++--
drivers/scsi/mpi3mr/mpi/mpi30_transport.h | 2 +-
4 files changed, 87 insertions(+), 14 deletions(-)
diff --git a/drivers/scsi/mpi3mr/mpi/mpi30_cnfg.h b/drivers/scsi/mpi3mr/mpi/mpi30_cnfg.h
index 33dd303c97bb..7cf16a5c15b7 100644
--- a/drivers/scsi/mpi3mr/mpi/mpi30_cnfg.h
+++ b/drivers/scsi/mpi3mr/mpi/mpi30_cnfg.h
@@ -72,6 +72,12 @@
#define MPI3_SECURITY_PGAD_SLOT_GROUP_SHIFT (8)
#define MPI3_SECURITY_PGAD_SLOT_MASK (0x000000ff)
#define MPI3_INSTANCE_PGAD_INSTANCE_MASK (0x0000ffff)
+#define MPI3_INSTANCE_PGAD_INSTANCE_SHIFT (0)
+#define MPI3_INTERFACE_PGAD_INTERFACE_MASK (0x0000000f)
+#define MPI3_INTERFACE_PGAD_INTERFACE_SHIFT (0)
+#define MPI3_INTERFACE_PGAD_INTERFACE_MPI (0)
+#define MPI3_INTERFACE_PGAD_INTERFACE_NVME_VD (1)
+#define MPI3_INTERFACE_PGAD_INTERFACE_NVME_PD (2)
struct mpi3_config_request {
__le16 host_tag;
u8 ioc_use_only02;
@@ -492,10 +498,31 @@ struct mpi3_man10_istwi_ctrlr_entry {
};
#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_BUS_SPEED_MASK (0x000c)
-#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_BUS_SPEED_100K (0x0000)
-#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_BUS_SPEED_400K (0x0004)
+#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_BUS_SPEED_100_KHZ (0x0000)
+#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_BUS_SPEED_400_KHZ (0x0004)
#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_TARGET_ENABLED (0x0002)
#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_INITIATOR_ENABLED (0x0001)
+#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_I2C_GLITCH_FLTR_MASK (0xc000)
+#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_I2C_GLITCH_FLTR_SHIFT (14)
+#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_I2C_GLITCH_FLTR_50_NS (0x0000)
+#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_I2C_GLITCH_FLTR_10_NS (0x4000)
+#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_I2C_GLITCH_FLTR_5_NS (0x8000)
+#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_I2C_GLITCH_FLTR_0_NS (0xc000)
+#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_BUS_TYPE_MASK (0x3000)
+#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_BUS_TYPE_SHIFT (12)
+#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_BUS_TYPE_I2C (0x0000)
+#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_BUS_TYPE_I3C (0x1000)
+#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_BUS_TYPE_AUTO (0x2000)
+#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_I3C_MAX_DATA_RATE_MASK (0x0e00)
+#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_I3C_MAX_DATA_RATE_SHIFT (9)
+#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_I3C_MAX_DATA_RATE_12_5_MHZ (0x0000)
+#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_I3C_MAX_DATA_RATE_8_MHZ (0x0200)
+#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_I3C_MAX_DATA_RATE_6_MHZ (0x0400)
+#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_I3C_MAX_DATA_RATE_4_MHZ (0x0600)
+#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_I3C_MAX_DATA_RATE_2_MHZ (0x0800)
+#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_BUS_SPEED_MASK (0x000c)
+#define MPI3_MAN10_ISTWI_CTRLR_FLAGS_BUS_SPEED_SHIFT (0)
+
#ifndef MPI3_MAN10_ISTWI_CTRLR_MAX
#define MPI3_MAN10_ISTWI_CTRLR_MAX (1)
#endif
@@ -1027,6 +1054,16 @@ struct mpi3_io_unit_page5 {
#define MPI3_IOUNIT5_DEVICE_SHUTDOWN_SWITCH_ATTACHED (0x02)
#define MPI3_IOUNIT5_DEVICE_SHUTDOWN_DIRECT_AND_EXPANDER (0x03)
#define MPI3_IOUNIT5_DEVICE_SHUTDOWN_DIRECT_AND_SWITCH (0x03)
+#define MPI3_IOUNIT5_DEVICE_SHUTDOWN_HDD_SPINDOWN_MASK (0xc000)
+#define MPI3_IOUNIT5_DEVICE_SHUTDOWN_HDD_SPINDOWN_NONE (0x0000)
+#define MPI3_IOUNIT5_DEVICE_SHUTDOWN_HDD_SPINDOWN_ALL (0x4000)
+#define MPI3_IOUNIT5_DEVICE_SHUTDOWN_HDD_SPINDOWN_FILTERED (0x8000)
+#define MPI3_IOUNIT5_DEVICE_SHUTDOWN_HDD_SPINDOWN_RESERVED (0xc000)
+#define MPI3_IOUNIT5_DEVICE_SHUTDOWN_SYNC_CACHE_MASK (0x3000)
+#define MPI3_IOUNIT5_DEVICE_SHUTDOWN_SYNC_CACHE_ALL (0x0000)
+#define MPI3_IOUNIT5_DEVICE_SHUTDOWN_SYNC_CACHE_FILTERED (0x1000)
+#define MPI3_IOUNIT5_DEVICE_SHUTDOWN_SYNC_CACHE_NONE (0x2000)
+#define MPI3_IOUNIT5_DEVICE_SHUTDOWN_SYNC_CACHE_RESERVED (0x3000)
#define MPI3_IOUNIT5_DEVICE_SHUTDOWN_SATA_HDD_MASK (0x0300)
#define MPI3_IOUNIT5_DEVICE_SHUTDOWN_SATA_HDD_SHIFT (8)
#define MPI3_IOUNIT5_DEVICE_SHUTDOWN_SAS_HDD_MASK (0x00c0)
@@ -1069,7 +1106,8 @@ struct mpi3_io_unit_page8 {
struct mpi3_config_page_header header;
u8 sb_mode;
u8 sb_state;
- __le16 reserved0a;
+ u8 flags;
+ u8 reserved0b;
u8 num_slots;
u8 slots_available;
u8 current_key_encryption_algo;
@@ -1088,6 +1126,8 @@ struct mpi3_io_unit_page8 {
#define MPI3_IOUNIT8_SBSTATE_SVN_UPDATE_PENDING (0x04)
#define MPI3_IOUNIT8_SBSTATE_KEY_UPDATE_PENDING (0x02)
#define MPI3_IOUNIT8_SBSTATE_SECURE_BOOT_ENABLED (0x01)
+#define MPI3_IOUNIT8_FLAGS_FWQR_CAPABLE (0x80)
+#define MPI3_IOUNIT8_FLAGS_FWQR_SECURED (0x40)
#define MPI3_IOUNIT8_SBMODE_CURRENT_KEY_IOUNIT17 (0x10)
#define MPI3_IOUNIT8_SBMODE_HARD_SECURE_RECERTIFIED (0x08)
struct mpi3_io_unit_page9 {
@@ -1174,10 +1214,16 @@ struct mpi3_io_unit_page12 {
#define MPI3_IOUNIT12_FLAGS_NUMPASSES_32 (0x00000200)
#define MPI3_IOUNIT12_FLAGS_NUMPASSES_64 (0x00000300)
#define MPI3_IOUNIT12_FLAGS_PASSPERIOD_MASK (0x00000003)
+#define MPI3_IOUNIT12_FLAGS_PASSPERIOD_SHIFT (0)
#define MPI3_IOUNIT12_FLAGS_PASSPERIOD_DISABLED (0x00000000)
#define MPI3_IOUNIT12_FLAGS_PASSPERIOD_500US (0x00000001)
#define MPI3_IOUNIT12_FLAGS_PASSPERIOD_1MS (0x00000002)
#define MPI3_IOUNIT12_FLAGS_PASSPERIOD_2MS (0x00000003)
+#define MPI3_IOUNIT12_FLAGS_INTERFACE_MASK (0x0000000c)
+#define MPI3_IOUNIT12_FLAGS_INTERFACE_SHIFT (2)
+#define MPI3_IOUNIT12_FLAGS_INTERFACE_MPI (0x00000000)
+#define MPI3_IOUNIT12_FLAGS_INTERFACE_NVME_VD (0x00000004)
+#define MPI3_IOUNIT12_FLAGS_INTERFACE_NVME_PD (0x00000008)
#ifndef MPI3_IOUNIT13_FUNC_MAX
#define MPI3_IOUNIT13_FUNC_MAX (1)
#endif
@@ -1238,6 +1284,7 @@ struct mpi3_io_unit_page15 {
#define MPI3_IOUNIT15_PAGEVERSION (0x00)
#define MPI3_IOUNIT15_FLAGS_EPRINIT_INITREQUIRED (0x04)
#define MPI3_IOUNIT15_FLAGS_EPRSUPPORT_MASK (0x03)
+#define MPI3_IOUNIT15_FLAGS_EPRSUPPORT_SHIFT (0)
#define MPI3_IOUNIT15_FLAGS_EPRSUPPORT_NOT_SUPPORTED (0x00)
#define MPI3_IOUNIT15_FLAGS_EPRSUPPORT_WITHOUT_POWER_BRAKE_GPIO (0x01)
#define MPI3_IOUNIT15_FLAGS_EPRSUPPORT_WITH_POWER_BRAKE_GPIO (0x02)
@@ -1255,6 +1302,9 @@ struct mpi3_io_unit_page17 {
__le32 current_key[];
};
#define MPI3_IOUNIT17_PAGEVERSION (0x00)
+#define MPI3_IOUNIT17_FLAGS_KEYROOT_MASK (0x01)
+#define MPI3_IOUNIT17_FLAGS_KEYROOT_HW (0x00)
+#define MPI3_IOUNIT17_FLAGS_KEYROOT_FW (0x01)
struct mpi3_io_unit_page18 {
struct mpi3_config_page_header header;
u8 flags;
@@ -1640,11 +1690,28 @@ struct mpi3_security_page3 {
};
#define MPI3_SECURITY3_PAGEVERSION (0x00)
-#define MPI3_SECURITY3_FLAGS_TYPE_MASK (0x0f)
+#define MPI3_SECURITY3_FLAGS_TYPE_MASK (0x1f)
#define MPI3_SECURITY3_FLAGS_TYPE_SHIFT (0)
#define MPI3_SECURITY3_FLAGS_TYPE_NOT_VALID (0)
#define MPI3_SECURITY3_FLAGS_TYPE_MLDSA_PRIVATE (1)
#define MPI3_SECURITY3_FLAGS_TYPE_MLDSA_PUBLIC (2)
+union mpi3_security_digest {
+ __le32 dword[16];
+ __le16 word[32];
+ u8 byte[64];
+};
+struct mpi3_security_page4 {
+ struct mpi3_config_page_header header;
+ __le32 reserved08[2];
+ union mpi3_security_mac mac;
+ union mpi3_security_nonce nonce;
+ u8 num_digests;
+ u8 hash_algorithm;
+ __le16 reserved92;
+ __le32 reserved94[3];
+ union mpi3_security_digest digest[];
+};
+#define MPI3_SECURITY4_PAGEVERSION (0x00)
struct mpi3_security_page10 {
struct mpi3_config_page_header header;
__le32 reserved08[2];
@@ -2074,7 +2141,7 @@ struct mpi3_sas_phy3_phy_event_config {
#define MPI3_SASPHY3_EVENT_CODE_LCCONN_TIME (0xd5)
#define MPI3_SASPHY3_EVENT_CODE_SSP_TX_START_TRANSMIT (0xd6)
#define MPI3_SASPHY3_EVENT_CODE_SATA_TX_START (0xd7)
-#define MPI3_SASPHY3_EVENT_CODE_SMP_TX_START_TRANSMT (0xd8)
+#define MPI3_SASPHY3_EVENT_CODE_SMP_TX_START_TRANSMIT (0xd8)
#define MPI3_SASPHY3_EVENT_CODE_TX_SMP_BREAK_CONN (0xd9)
#define MPI3_SASPHY3_EVENT_CODE_SSP_RX_START_RECEIVE (0xda)
#define MPI3_SASPHY3_EVENT_CODE_SATA_RX_START_RECEIVE (0xdb)
diff --git a/drivers/scsi/mpi3mr/mpi/mpi30_image.h b/drivers/scsi/mpi3mr/mpi/mpi30_image.h
index 62ddf094d46c..5fa09fa79358 100644
--- a/drivers/scsi/mpi3mr/mpi/mpi30_image.h
+++ b/drivers/scsi/mpi3mr/mpi/mpi30_image.h
@@ -18,7 +18,7 @@ struct mpi3_hash_exclusion_format {
__le32 size;
};
-#define MPI3_IMAGE_HASH_EXCUSION_NUM (4)
+#define MPI3_IMAGE_HASH_EXCLUSION_NUM (4)
struct mpi3_component_image_header {
__le32 signature0;
__le32 load_address;
@@ -42,7 +42,7 @@ struct mpi3_component_image_header {
union mpi3_version_union rmc_interface_version;
union mpi3_version_union etp_interface_version;
struct mpi3_comp_image_version component_image_version;
- struct mpi3_hash_exclusion_format hash_exclusion[MPI3_IMAGE_HASH_EXCUSION_NUM];
+ struct mpi3_hash_exclusion_format hash_exclusion[MPI3_IMAGE_HASH_EXCLUSION_NUM];
__le32 next_image_header_offset;
union mpi3_version_union security_version;
__le32 reserved84[31];
@@ -347,7 +347,8 @@ struct mpi3_encrypted_hash_entry {
struct mpi3_encrypted_hash_data {
u8 image_version;
u8 num_hash;
- __le16 reserved02;
+ u8 fw_num_hash;
+ u8 reserved03;
__le32 reserved04;
struct mpi3_encrypted_hash_entry encrypted_hash_entry[MPI3_ENCRYPTED_HASH_ENTRY_MAX];
};
diff --git a/drivers/scsi/mpi3mr/mpi/mpi30_ioc.h b/drivers/scsi/mpi3mr/mpi/mpi30_ioc.h
index 68efa0d51345..aa42fba7f930 100644
--- a/drivers/scsi/mpi3mr/mpi/mpi30_ioc.h
+++ b/drivers/scsi/mpi3mr/mpi/mpi30_ioc.h
@@ -428,10 +428,10 @@ struct mpi3_event_data_sas_discovery {
#define MPI3_EVENT_SAS_DISC_FLAGS_IN_PROGRESS (0x01)
#define MPI3_EVENT_SAS_DISC_RC_STARTED (0x01)
#define MPI3_EVENT_SAS_DISC_RC_COMPLETED (0x02)
-#define MPI3_SAS_DISC_STATUS_MAX_ENCLOSURES_EXCEED (0x80000000)
-#define MPI3_SAS_DISC_STATUS_MAX_EXPANDERS_EXCEED (0x40000000)
-#define MPI3_SAS_DISC_STATUS_MAX_DEVICES_EXCEED (0x20000000)
-#define MPI3_SAS_DISC_STATUS_MAX_TOPO_PHYS_EXCEED (0x10000000)
+#define MPI3_SAS_DISC_STATUS_MAX_ENCLOSURES_EXCEEDED (0x80000000)
+#define MPI3_SAS_DISC_STATUS_MAX_EXPANDERS_EXCEEDED (0x40000000)
+#define MPI3_SAS_DISC_STATUS_MAX_DEVICES_EXCEEDED (0x20000000)
+#define MPI3_SAS_DISC_STATUS_MAX_TOPO_PHYS_EXCEEDED (0x10000000)
#define MPI3_SAS_DISC_STATUS_INVALID_CEI (0x00010000)
#define MPI3_SAS_DISC_STATUS_FECEI_MISMATCH (0x00008000)
#define MPI3_SAS_DISC_STATUS_MULTIPLE_DEVICES_IN_SLOT (0x00004000)
@@ -965,7 +965,7 @@ struct mpi3_ci_download_reply {
u8 flags;
u8 cache_dirty;
u8 pending_count;
- u8 reserved13;
+ u8 additional_flags;
};
#define MPI3_CI_DOWNLOAD_FLAGS_DOWNLOAD_IN_PROGRESS (0x80)
@@ -979,6 +979,11 @@ struct mpi3_ci_download_reply {
#define MPI3_CI_DOWNLOAD_FLAGS_ACTIVATION_STATUS_ONLINE_PENDING (0x04)
#define MPI3_CI_DOWNLOAD_FLAGS_ACTIVATION_STATUS_OFFLINE_PENDING (0x06)
#define MPI3_CI_DOWNLOAD_FLAGS_COMPATIBLE (0x01)
+#define MPI3_CI_DOWNLOAD_ADDITIONALFLAGS_REDUNDANCYRESTORATION_MASK (0x03)
+#define MPI3_CI_DOWNLOAD_ADDITIONALFLAGS_REDUNDANCYRESTORATION_SHIFT (0)
+#define MPI3_CI_DOWNLOAD_ADDITIONALFLAGS_REDUNDANCYRESTORATION_NONE (0x00)
+#define MPI3_CI_DOWNLOAD_ADDITIONALFLAGS_REDUNDANCYRESTORATION_PRIMARY (0x01)
+#define MPI3_CI_DOWNLOAD_ADDITIONALFLAGS_REDUNDANCYRESTORATION_SECONDARY (0x02)
struct mpi3_ci_upload_request {
__le16 host_tag;
u8 ioc_use_only02;
diff --git a/drivers/scsi/mpi3mr/mpi/mpi30_transport.h b/drivers/scsi/mpi3mr/mpi/mpi30_transport.h
index 290a1f5c2924..794ecc778945 100644
--- a/drivers/scsi/mpi3mr/mpi/mpi30_transport.h
+++ b/drivers/scsi/mpi3mr/mpi/mpi30_transport.h
@@ -18,7 +18,7 @@ union mpi3_version_union {
#define MPI3_VERSION_MAJOR (3)
#define MPI3_VERSION_MINOR (0)
-#define MPI3_VERSION_UNIT (39)
+#define MPI3_VERSION_UNIT (41)
#define MPI3_VERSION_DEV (0)
#define MPI3_DEVHANDLE_INVALID (0xffff)
struct mpi3_sysif_oper_queue_indexes {
--
2.47.3
^ permalink raw reply related [flat|nested] 14+ messages in thread* [PATCH v3 03/10] mpi3mr: Add early timestamp synchronization after driver load
2026-07-24 10:24 [PATCH v3 00/10] mpi3mr: Few Enhancements and minor fixes Ranjan Kumar
2026-07-24 10:24 ` [PATCH v3 01/10] mpi3mr: Skip device shutdown during unload per controller configuration Ranjan Kumar
2026-07-24 10:24 ` [PATCH v3 02/10] mpi3mr: Update MPI Headers to revision 41 Ranjan Kumar
@ 2026-07-24 10:24 ` Ranjan Kumar
2026-07-24 10:24 ` [PATCH v3 04/10] mpi3mr: Fix NVMe page size caching for non-operational devices Ranjan Kumar
` (6 subsequent siblings)
9 siblings, 0 replies; 14+ messages in thread
From: Ranjan Kumar @ 2026-07-24 10:24 UTC (permalink / raw)
To: linux-scsi, martin.petersen
Cc: sathya.prakash, chandrakanth.patil, vishakhavc, ipylypiv,
Ranjan Kumar
When the driver is loaded from initramfs, the controller timestamp may
be initialized before the system clock has been synchronized. As a
result, the controller can operate with a stale timestamp until the
first periodic synchronization occurs.
Currently, the first controller timestamp synchronization occurs only
after the configured ts_update_interval expires (15 minutes by default).
Add an early timestamp synchronization 60 seconds after driver load,
followed by the existing periodic synchronization interval.
Signed-off-by: Ranjan Kumar <ranjan.kumar@broadcom.com>
---
drivers/scsi/mpi3mr/mpi3mr.h | 3 +++
drivers/scsi/mpi3mr/mpi3mr_fw.c | 25 +++++++++++++++++++------
2 files changed, 22 insertions(+), 6 deletions(-)
diff --git a/drivers/scsi/mpi3mr/mpi3mr.h b/drivers/scsi/mpi3mr/mpi3mr.h
index 39096004c60a..1f2f0951b560 100644
--- a/drivers/scsi/mpi3mr/mpi3mr.h
+++ b/drivers/scsi/mpi3mr/mpi3mr.h
@@ -125,6 +125,7 @@ extern atomic64_t event_counter;
#define MPI3MR_RESETTM_TIMEOUT 60
#define MPI3MR_RESET_HOST_IOWAIT_TIMEOUT 5
#define MPI3MR_TSUPDATE_INTERVAL 900
+#define MPI3MR_EARLY_TSUPDATE_SECONDS 60
#define MPI3MR_DEFAULT_SHUTDOWN_TIME 120
#define MPI3MR_RAID_ERRREC_RESET_TIMEOUT 180
#define MPI3MR_PREPARE_FOR_RESET_TIMEOUT 180
@@ -1118,6 +1119,7 @@ struct scmd_priv {
* @evtack_cmds_bitmap: Event Ack bitmap
* @delayed_evtack_cmds_list: Delayed event acknowledgment list
* @ts_update_counter: Timestamp update counter
+ * @early_ts_sync_done: Early (1 min) timestamp sync completed after load
* @ts_update_interval: Timestamp update interval
* @reset_in_progress: Reset in progress flag
* @unrecoverable: Controller unrecoverable flag
@@ -1318,6 +1320,7 @@ struct mpi3mr_ioc {
struct list_head delayed_evtack_cmds_list;
u16 ts_update_counter;
+ u8 early_ts_sync_done;
u16 ts_update_interval;
u8 reset_in_progress;
u8 unrecoverable;
diff --git a/drivers/scsi/mpi3mr/mpi3mr_fw.c b/drivers/scsi/mpi3mr/mpi3mr_fw.c
index 59241038f689..434b66f7b502 100644
--- a/drivers/scsi/mpi3mr/mpi3mr_fw.c
+++ b/drivers/scsi/mpi3mr/mpi3mr_fw.c
@@ -2870,8 +2870,9 @@ static int mpi3mr_print_pkg_ver(struct mpi3mr_ioc *mrioc)
* @work: work struct
*
* Watch dog work periodically executed (1 second interval) to
- * monitor firmware fault and to issue periodic timer sync to
- * the firmware.
+ * monitor firmware fault and perform timestamp synchronization
+ * to firmware, with an early sync 1 minute after load followed
+ * by periodic updates at ts_update_interval seconds (default 15 minutes).
*
* Return: Nothing.
*/
@@ -2917,11 +2918,23 @@ static void mpi3mr_watchdog_work(struct work_struct *work)
}
if (!(mrioc->facts.ioc_capabilities &
- MPI3_IOCFACTS_CAPABILITY_NON_SUPERVISOR_IOC) &&
- (mrioc->ts_update_counter++ >= mrioc->ts_update_interval)) {
+ MPI3_IOCFACTS_CAPABILITY_NON_SUPERVISOR_IOC)) {
+ if (!mrioc->early_ts_sync_done) {
+ /*
+ * Send time sync 1 min after load
+ */
+ if (mrioc->ts_update_counter++ >=
+ MPI3MR_EARLY_TSUPDATE_SECONDS) {
+ mrioc->early_ts_sync_done = 1;
+ mrioc->ts_update_counter = 0;
+ mpi3mr_sync_timestamp(mrioc);
+ }
+ } else if (mrioc->ts_update_counter++ >=
+ mrioc->ts_update_interval) {
+ mrioc->ts_update_counter = 0;
+ mpi3mr_sync_timestamp(mrioc);
+ }
- mrioc->ts_update_counter = 0;
- mpi3mr_sync_timestamp(mrioc);
}
if ((mrioc->prepare_for_reset) &&
--
2.47.3
^ permalink raw reply related [flat|nested] 14+ messages in thread* [PATCH v3 04/10] mpi3mr: Fix NVMe page size caching for non-operational devices
2026-07-24 10:24 [PATCH v3 00/10] mpi3mr: Few Enhancements and minor fixes Ranjan Kumar
` (2 preceding siblings ...)
2026-07-24 10:24 ` [PATCH v3 03/10] mpi3mr: Add early timestamp synchronization after driver load Ranjan Kumar
@ 2026-07-24 10:24 ` Ranjan Kumar
2026-07-24 10:44 ` sashiko-bot
2026-07-24 10:25 ` [PATCH v3 05/10] mpi3mr: Fix performance regression caused by extended IRQ poll sleep Ranjan Kumar
` (5 subsequent siblings)
9 siblings, 1 reply; 14+ messages in thread
From: Ranjan Kumar @ 2026-07-24 10:24 UTC (permalink / raw)
To: linux-scsi, martin.petersen
Cc: sathya.prakash, chandrakanth.patil, vishakhavc, ipylypiv,
Ranjan Kumar, Sashiko
For NVMe devices in an error state, the cached PCIe page size remains
unset. This causes management IOCTL validation to fail, preventing
requests from reaching firmware and returning incorrect errors to
userspace. Populate the page size attribute irrespective of
device access status so firmware can process IOCTLs and report
appropriate errors.
Additionally, harden the device initialization path against invalid
firmware data for non-operational devices:
1. Add bounds checking for page_size, falling back to 4096 bytes (shift
exponent 12) to prevent undefined shift behavior and kernel panics.
2. Initialize reset_to and abort_to timeouts with default values to
prevent IOCTLs from failing instantly and triggering unintended full
controller resets.
Reported-by: Sashiko <sashiko-bot@kernel.org>
Closes: https://sashiko.dev/#/patchset/20260626114109.43685-1-ranjan.kumar@broadcom.com?part=4
Closes: https://sashiko.dev/#/patchset/20260708183305.244485-1-ranjan.kumar@broadcom.com?part=4
Signed-off-by: Chandrakanth Patil <chandrakanth.patil@broadcom.com>
Signed-off-by: Ranjan Kumar <ranjan.kumar@broadcom.com>
---
drivers/scsi/mpi3mr/mpi3mr_os.c | 10 +++++++---
1 file changed, 7 insertions(+), 3 deletions(-)
diff --git a/drivers/scsi/mpi3mr/mpi3mr_os.c b/drivers/scsi/mpi3mr/mpi3mr_os.c
index d2a20f2721db..7b86152922ba 100644
--- a/drivers/scsi/mpi3mr/mpi3mr_os.c
+++ b/drivers/scsi/mpi3mr/mpi3mr_os.c
@@ -1354,12 +1354,16 @@ static void mpi3mr_update_tgtdev(struct mpi3mr_ioc *mrioc,
tgtdev->dev_spec.pcie_inf.capb =
le32_to_cpu(pcieinf->capabilities);
tgtdev->dev_spec.pcie_inf.mdts = MPI3MR_DEFAULT_MDTS;
- /* 2^12 = 4096 */
- tgtdev->dev_spec.pcie_inf.pgsz = 12;
+ tgtdev->dev_spec.pcie_inf.reset_to = MPI3MR_INTADMCMD_TIMEOUT;
+ tgtdev->dev_spec.pcie_inf.abort_to = MPI3MR_INTADMCMD_TIMEOUT;
+ /* Validate firmware page size to prevent undefined shift behavior */
+ if (pcieinf->page_size > 0 && pcieinf->page_size < 31)
+ tgtdev->dev_spec.pcie_inf.pgsz = pcieinf->page_size;
+ else
+ tgtdev->dev_spec.pcie_inf.pgsz = 12; /* Default to 4096 (2^12) */
if (dev_pg0->access_status == MPI3_DEVICE0_ASTATUS_NO_ERRORS) {
tgtdev->dev_spec.pcie_inf.mdts =
le32_to_cpu(pcieinf->maximum_data_transfer_size);
- tgtdev->dev_spec.pcie_inf.pgsz = pcieinf->page_size;
tgtdev->dev_spec.pcie_inf.reset_to =
max_t(u8, pcieinf->controller_reset_to,
MPI3MR_INTADMCMD_TIMEOUT);
--
2.47.3
^ permalink raw reply related [flat|nested] 14+ messages in thread* Re: [PATCH v3 04/10] mpi3mr: Fix NVMe page size caching for non-operational devices
2026-07-24 10:24 ` [PATCH v3 04/10] mpi3mr: Fix NVMe page size caching for non-operational devices Ranjan Kumar
@ 2026-07-24 10:44 ` sashiko-bot
0 siblings, 0 replies; 14+ messages in thread
From: sashiko-bot @ 2026-07-24 10:44 UTC (permalink / raw)
To: Ranjan Kumar; +Cc: linux-scsi
Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [Medium] Double-write of `reset_to` and `abort_to` timeouts introduces a race condition where concurrent readers may observe a reduced default timeout.
- [Medium] The bounds check for `page_size` accepts invalid NVMe page size shifts (1 to 11), which causes management IOCTL validation to fail before reaching the firmware.
--
--- Patch [4]: [PATCH v3 04/10] mpi3mr: Fix NVMe page size caching for non-operational devices ---
commit b27f8d86d47076feb2c7ac1adf8a15a3291cf854
Author: Ranjan Kumar <ranjan.kumar@broadcom.com>
mpi3mr: Fix NVMe page size caching for non-operational devices
This commit populates the NVMe page size attribute for devices in an error
state so management IOCTLs can reach the firmware. It adds bounds checking
for page_size and initializes reset_to and abort_to with default values.
Closes: https://sashiko.dev/#/patchset/20260626114109.43685-1-ranjan.kumar@broadcom.com?part=4
Closes: https://sashiko.dev/#/patchset/20260708183305.244485-1-ranjan.kumar@broadcom.com?part=4
> diff --git a/drivers/scsi/mpi3mr/mpi3mr_os.c b/drivers/scsi/mpi3mr/mpi3mr_os.c
> index d2a20f2721dbf..7b86152922bad 100644
> --- a/drivers/scsi/mpi3mr/mpi3mr_os.c
> +++ b/drivers/scsi/mpi3mr/mpi3mr_os.c
> @@ -1354,12 +1354,16 @@ static void mpi3mr_update_tgtdev(struct mpi3mr_ioc *mrioc,
> tgtdev->dev_spec.pcie_inf.capb =
> le32_to_cpu(pcieinf->capabilities);
> tgtdev->dev_spec.pcie_inf.mdts = MPI3MR_DEFAULT_MDTS;
> - /* 2^12 = 4096 */
> - tgtdev->dev_spec.pcie_inf.pgsz = 12;
> + tgtdev->dev_spec.pcie_inf.reset_to = MPI3MR_INTADMCMD_TIMEOUT;
> + tgtdev->dev_spec.pcie_inf.abort_to = MPI3MR_INTADMCMD_TIMEOUT;
[Severity: Medium]
Can a concurrent reader observe the intermediate default timeout values here?
If a Task Management Request (via mpi3mr_issue_tm()) executes concurrently
with a DevStatusChange event processed by mpi3mr_devinfochg_evt_bh(), it might
read these prematurely shortened 60-second timeouts.
Since mpi3mr_update_tgtdev() modifies tgtdev without holding tgtdev_lock, the
reader can observe this intermediate value before it is overwritten with the
final firmware-provided timeout a few lines further down. This could cause
the command to time out early and trigger an unintended full controller reset.
> + /* Validate firmware page size to prevent undefined shift behavior */
> + if (pcieinf->page_size > 0 && pcieinf->page_size < 31)
> + tgtdev->dev_spec.pcie_inf.pgsz = pcieinf->page_size;
[Severity: Medium]
Does this bounds check accept values that are logically invalid for NVMe
memory page sizes?
If the firmware provides corrupted device page0 data with a page_size between
1 and 11, this condition incorrectly accepts it.
Later, when an application issues an IOCTL, the PRP builder in mpi3mr_app.c
evaluates dev_pgsz = 1 << pgsz to a small value (e.g., 2 to 2048 bytes). This
fails size constraints and rejects the IOCTL before it reaches the firmware,
returning incorrect errors to userspace.
> + else
> + tgtdev->dev_spec.pcie_inf.pgsz = 12; /* Default to 4096 (2^12) */
> if (dev_pg0->access_status == MPI3_DEVICE0_ASTATUS_NO_ERRORS) {
> tgtdev->dev_spec.pcie_inf.mdts =
> le32_to_cpu(pcieinf->maximum_data_transfer_size);
> - tgtdev->dev_spec.pcie_inf.pgsz = pcieinf->page_size;
> tgtdev->dev_spec.pcie_inf.reset_to =
> max_t(u8, pcieinf->controller_reset_to,
> MPI3MR_INTADMCMD_TIMEOUT);
--
Sashiko AI review · https://sashiko.dev/#/patchset/20260724102505.115136-1-ranjan.kumar@broadcom.com?part=4
^ permalink raw reply [flat|nested] 14+ messages in thread
* [PATCH v3 05/10] mpi3mr: Fix performance regression caused by extended IRQ poll sleep
2026-07-24 10:24 [PATCH v3 00/10] mpi3mr: Few Enhancements and minor fixes Ranjan Kumar
` (3 preceding siblings ...)
2026-07-24 10:24 ` [PATCH v3 04/10] mpi3mr: Fix NVMe page size caching for non-operational devices Ranjan Kumar
@ 2026-07-24 10:25 ` Ranjan Kumar
2026-07-24 10:50 ` sashiko-bot
2026-07-24 10:25 ` [PATCH v3 06/10] mpi3mr: Fix memory leak on operational queue creation failure Ranjan Kumar
` (4 subsequent siblings)
9 siblings, 1 reply; 14+ messages in thread
From: Ranjan Kumar @ 2026-07-24 10:25 UTC (permalink / raw)
To: linux-scsi, martin.petersen
Cc: sathya.prakash, chandrakanth.patil, vishakhavc, ipylypiv,
Ranjan Kumar, Sashiko
Commit 24d7071d9645 ("scsi: mpi3mr: A performance fix") increased the
threaded IRQ poll sleep range from 2-20 us to 20-21 us to work around a
timer slack issue.
On kernels unaffected by the timer slack issue, the longer sleep interval
reduces reply queue processing efficiency and causes an approximately 7%
throughput regression on NVMe direct-attached RAID10 configurations.
Restore the IRQ poll sleep range to 2-20 us to recover the lost
throughput.
Additionally, resolve the following issues in the reply queue processing
and polling logic:
1. Add missing dma_rmb() memory barriers in the admin and operational
reply queue processing loops. This ensures that the descriptor
payload is only read after the phase bit check is complete, preventing
weakly ordered architectures from speculatively processing stale data.
2. Add bounds checking for `request_queue_id` in
mpi3mr_process_op_reply_q() to prevent out-of-bounds memory
corruption if the hardware provides an invalid queue ID.
3. Fix a TOCTOU race condition in mpi3mr_process_op_reply_q() by
implementing a double-check pattern with dma_rmb() and
atomic_add_unless() when exiting the loop. This prevents lost
interrupts if a descriptor arrives exactly as the thread drops the
in_use lock.
4. Replace a direct panic() call with a safe ioc_err() log and abort in
mpi3mr_process_op_reply_desc() when mpi3mr_get_reply_virt_addr()
returns NULL. This prevents a single malformed DMA reply address from
crashing the entire host OS.
Note: The unbounded busy-wait loop (usleep_range) in mpi3mr_isr_poll()
flagged by automated review is intentionally retained. This short sleep
polling mechanism is critical for batching completions and achieving the
target throughput on high-performance NVMe configurations.
Reported-by: Sashiko <sashiko-bot@kernel.org>
Closes: https://sashiko.dev/#/patchset/20260626114109.43685-1-ranjan.kumar@broadcom.com?part=5
Closes: https://sashiko.dev/#/patchset/20260708183305.244485-1-ranjan.kumar@broadcom.com?part=5
Signed-off-by: Chandrakanth Patil <chandrakanth.patil@broadcom.com>
Signed-off-by: Ranjan Kumar <ranjan.kumar@broadcom.com>
---
drivers/scsi/mpi3mr/mpi3mr.h | 2 +-
drivers/scsi/mpi3mr/mpi3mr_fw.c | 56 ++++++++++++++++++++++++++++++---
drivers/scsi/mpi3mr/mpi3mr_os.c | 3 +-
3 files changed, 54 insertions(+), 7 deletions(-)
diff --git a/drivers/scsi/mpi3mr/mpi3mr.h b/drivers/scsi/mpi3mr/mpi3mr.h
index 1f2f0951b560..1d11d7c69536 100644
--- a/drivers/scsi/mpi3mr/mpi3mr.h
+++ b/drivers/scsi/mpi3mr/mpi3mr.h
@@ -178,7 +178,7 @@ extern atomic64_t event_counter;
#define MPI3MR_DEFAULT_SDEV_QD 32
/* Definitions for Threaded IRQ poll*/
-#define MPI3MR_IRQ_POLL_SLEEP 20
+#define MPI3MR_IRQ_POLL_SLEEP 2
#define MPI3MR_IRQ_POLL_TRIGGER_IOCOUNT 8
/* Definitions for the controller security status*/
diff --git a/drivers/scsi/mpi3mr/mpi3mr_fw.c b/drivers/scsi/mpi3mr/mpi3mr_fw.c
index 434b66f7b502..9f7cee26ebcd 100644
--- a/drivers/scsi/mpi3mr/mpi3mr_fw.c
+++ b/drivers/scsi/mpi3mr/mpi3mr_fw.c
@@ -473,6 +473,12 @@ int mpi3mr_process_admin_reply_q(struct mpi3mr_ioc *mrioc)
return 0;
}
+ /*
+ * Ensure that the descriptor payload is read only after
+ * the phase bit check is complete.
+ */
+ dma_rmb();
+
do {
if (mrioc->unrecoverable || mrioc->io_admin_reset_sync)
break;
@@ -493,6 +499,13 @@ int mpi3mr_process_admin_reply_q(struct mpi3mr_ioc *mrioc)
if ((le16_to_cpu(reply_desc->reply_flags) &
MPI3_REPLY_DESCRIPT_FLAGS_PHASE_MASK) != exp_phase)
break;
+
+ /*
+ * Ensure that the descriptor payload is read only after
+ * the phase bit check is complete.
+ */
+ dma_rmb();
+
if (threshold_comps == MPI3MR_THRESHOLD_REPLY_COUNT) {
writel(admin_reply_ci,
&mrioc->sysif_regs->admin_reply_queue_ci);
@@ -565,14 +578,33 @@ int mpi3mr_process_op_reply_q(struct mpi3mr_ioc *mrioc,
if ((le16_to_cpu(reply_desc->reply_flags) &
MPI3_REPLY_DESCRIPT_FLAGS_PHASE_MASK) != exp_phase) {
atomic_dec(&op_reply_q->in_use);
+ /* Check for a TOCTOU race condition */
+ dma_rmb();
+ if ((le16_to_cpu(reply_desc->reply_flags) &
+ MPI3_REPLY_DESCRIPT_FLAGS_PHASE_MASK) == exp_phase) {
+ if (atomic_add_unless(&op_reply_q->in_use, 1, 1))
+ goto process_desc;
+ }
return 0;
}
+process_desc:
+ /*
+ * Ensure that the descriptor payload is read only after
+ * the phase bit check is complete.
+ */
+ dma_rmb();
do {
if (mrioc->unrecoverable || mrioc->io_admin_reset_sync)
break;
req_q_idx = le16_to_cpu(reply_desc->request_queue_id) - 1;
+
+ if (unlikely(req_q_idx >= mrioc->num_op_req_q)) {
+ ioc_err(mrioc, "Invalid request queue id %d\n", req_q_idx + 1);
+ break;
+ }
+
op_req_q = &mrioc->req_qinfo[req_q_idx];
WRITE_ONCE(op_req_q->ci, le16_to_cpu(reply_desc->request_queue_ci));
@@ -592,8 +624,23 @@ int mpi3mr_process_op_reply_q(struct mpi3mr_ioc *mrioc,
reply_desc = mpi3mr_get_reply_desc(op_reply_q, reply_ci);
if ((le16_to_cpu(reply_desc->reply_flags) &
- MPI3_REPLY_DESCRIPT_FLAGS_PHASE_MASK) != exp_phase)
+ MPI3_REPLY_DESCRIPT_FLAGS_PHASE_MASK) != exp_phase) {
+ atomic_dec(&op_reply_q->in_use);
+ /* Check for a TOCTOU race condition */
+ dma_rmb();
+ if ((le16_to_cpu(reply_desc->reply_flags) &
+ MPI3_REPLY_DESCRIPT_FLAGS_PHASE_MASK) == exp_phase) {
+ /* Descriptor arrived, try to reclaim ownership */
+ if (atomic_add_unless(&op_reply_q->in_use, 1, 1))
+ continue;
+ }
break;
+ }
+ /*
+ * Ensure that the descriptor payload is read only after
+ * the phase bit check is complete.
+ */
+ dma_rmb();
#ifndef CONFIG_PREEMPT_RT
/*
* Exit completion loop to avoid CPU lockup
@@ -743,11 +790,12 @@ static irqreturn_t mpi3mr_isr_poll(int irq, void *privdata)
num_op_reply +=
mpi3mr_process_op_reply_q(mrioc,
intr_info->op_reply_q);
+ if (!atomic_read(&intr_info->op_reply_q->pend_ios))
+ break;
- usleep_range(MPI3MR_IRQ_POLL_SLEEP, MPI3MR_IRQ_POLL_SLEEP + 1);
+ usleep_range(MPI3MR_IRQ_POLL_SLEEP, 10 * MPI3MR_IRQ_POLL_SLEEP);
- } while (atomic_read(&intr_info->op_reply_q->pend_ios) &&
- (num_op_reply < mrioc->max_host_ios));
+ } while (num_op_reply < mrioc->max_host_ios);
intr_info->op_reply_q->enable_irq_poll = false;
enable_irq(intr_info->os_irq);
diff --git a/drivers/scsi/mpi3mr/mpi3mr_os.c b/drivers/scsi/mpi3mr/mpi3mr_os.c
index 7b86152922ba..39624fae9131 100644
--- a/drivers/scsi/mpi3mr/mpi3mr_os.c
+++ b/drivers/scsi/mpi3mr/mpi3mr_os.c
@@ -3426,8 +3426,7 @@ void mpi3mr_process_op_reply_desc(struct mpi3mr_ioc *mrioc,
scsi_reply = mpi3mr_get_reply_virt_addr(mrioc,
*reply_dma);
if (!scsi_reply) {
- panic("%s: scsi_reply is NULL, this shouldn't happen\n",
- mrioc->name);
+ ioc_err(mrioc, "scsi_reply is NULL, invalid reply_frame_address\n");
goto out;
}
host_tag = le16_to_cpu(scsi_reply->host_tag);
--
2.47.3
^ permalink raw reply related [flat|nested] 14+ messages in thread* Re: [PATCH v3 05/10] mpi3mr: Fix performance regression caused by extended IRQ poll sleep
2026-07-24 10:25 ` [PATCH v3 05/10] mpi3mr: Fix performance regression caused by extended IRQ poll sleep Ranjan Kumar
@ 2026-07-24 10:50 ` sashiko-bot
0 siblings, 0 replies; 14+ messages in thread
From: sashiko-bot @ 2026-07-24 10:50 UTC (permalink / raw)
To: Ranjan Kumar; +Cc: linux-scsi
Thank you for your contribution! Sashiko AI review found 5 potential issue(s) to consider:
- [Critical] Leaked invalid `reply_dma` address allows arbitrary hardware DMA memory corruption.
- [High] Double-decrement of the `op_reply_q->in_use` lock breaks queue mutual exclusion.
- [High] Improper use of `continue` inside the TOCTOU fallback skips critical DMA memory barriers and batch consumer index updates.
- [High] Breaking out of the processing loop on an invalid `request_queue_id` without advancing the consumer index causes an unrecoverable infinite polling loop and CPU lockup.
- [High] Reclaiming the queue lock after a TOCTOU race without reloading local indices causes data corruption via double-processing of descriptors.
--
--- Patch [5]: [PATCH v3 05/10] mpi3mr: Fix performance regression caused by extended IRQ poll sleep ---
commit 166312d7a3fc2817bfd92d0a670bc3744b27cc2e
Author: Ranjan Kumar <ranjan.kumar@broadcom.com>
mpi3mr: Fix performance regression caused by extended IRQ poll sleep
This patch restores the IRQ poll sleep range to 2-20 us to recover lost
throughput on unaffected kernels. It also introduces missing dma_rmb()
barriers, bounds checking for request_queue_id, TOCTOU race fixes, and
replaces a panic() with a safe log in the reply queue processing logic.
Link: https://sashiko.dev/#/patchset/20260626114109.43685-1-ranjan.kumar@broadcom.com?part=5
Link: https://sashiko.dev/#/patchset/20260708183305.244485-1-ranjan.kumar@broadcom.com?part=5
[ ... ]
> diff --git a/drivers/scsi/mpi3mr/mpi3mr_fw.c b/drivers/scsi/mpi3mr/mpi3mr_fw.c
> index 434b66f7b502d..9f7cee26ebcd3 100644
> --- a/drivers/scsi/mpi3mr/mpi3mr_fw.c
> +++ b/drivers/scsi/mpi3mr/mpi3mr_fw.c
[ ... ]
> @@ -565,14 +578,33 @@ int mpi3mr_process_op_reply_q(struct mpi3mr_ioc *mrioc,
> if ((le16_to_cpu(reply_desc->reply_flags) &
> MPI3_REPLY_DESCRIPT_FLAGS_PHASE_MASK) != exp_phase) {
> atomic_dec(&op_reply_q->in_use);
> + /* Check for a TOCTOU race condition */
> + dma_rmb();
> + if ((le16_to_cpu(reply_desc->reply_flags) &
> + MPI3_REPLY_DESCRIPT_FLAGS_PHASE_MASK) == exp_phase) {
> + if (atomic_add_unless(&op_reply_q->in_use, 1, 1))
> + goto process_desc;
[Severity: High]
If we reclaim the lock here and jump to process_desc, won't we use the local
exp_phase and reply_ci variables?
If another thread already processed this descriptor and updated the global
op_reply_q->ci and op_reply_q->ephase, could we end up double-processing
the same descriptor because our local indices weren't reloaded?
> + }
> return 0;
> }
> +process_desc:
> + /*
> + * Ensure that the descriptor payload is read only after
> + * the phase bit check is complete.
> + */
> + dma_rmb();
>
> do {
> if (mrioc->unrecoverable || mrioc->io_admin_reset_sync)
> break;
>
> req_q_idx = le16_to_cpu(reply_desc->request_queue_id) - 1;
> +
> + if (unlikely(req_q_idx >= mrioc->num_op_req_q)) {
> + ioc_err(mrioc, "Invalid request queue id %d\n", req_q_idx + 1);
> + break;
[Severity: High]
If we break out of the loop here, we don't advance the consumer index
(reply_ci) or decrement op_reply_q->pend_ios.
In mpi3mr_isr_poll(), if pend_ios remains > 0 and num_op_reply is 0,
could this cause an unrecoverable infinite polling loop?
> + }
> +
> op_req_q = &mrioc->req_qinfo[req_q_idx];
>
> WRITE_ONCE(op_req_q->ci, le16_to_cpu(reply_desc->request_queue_ci));
> @@ -592,8 +624,23 @@ int mpi3mr_process_op_reply_q(struct mpi3mr_ioc *mrioc,
> reply_desc = mpi3mr_get_reply_desc(op_reply_q, reply_ci);
>
> if ((le16_to_cpu(reply_desc->reply_flags) &
> - MPI3_REPLY_DESCRIPT_FLAGS_PHASE_MASK) != exp_phase)
> + MPI3_REPLY_DESCRIPT_FLAGS_PHASE_MASK) != exp_phase) {
> + atomic_dec(&op_reply_q->in_use);
> + /* Check for a TOCTOU race condition */
> + dma_rmb();
> + if ((le16_to_cpu(reply_desc->reply_flags) &
> + MPI3_REPLY_DESCRIPT_FLAGS_PHASE_MASK) == exp_phase) {
> + /* Descriptor arrived, try to reclaim ownership */
> + if (atomic_add_unless(&op_reply_q->in_use, 1, 1))
> + continue;
[Severity: High]
If we successfully reclaim ownership and use continue to jump back to the
start of the loop, doesn't this bypass the trailing dma_rmb() and the
threshold_comps exact-match check at the end of the loop?
Could missing the exact-match threshold check prevent the consumer index
from being batched to the hardware, potentially stalling IO?
> + }
> break;
[Severity: High]
If the phase bit check fails here and we break out of the loop, we have
already called atomic_dec(&op_reply_q->in_use) above.
Since the function cleanup block at the end also unconditionally executes
atomic_dec(&op_reply_q->in_use), will this result in a double-decrement
that drops the lock value below zero and breaks queue mutual exclusion?
> + }
[ ... ]
> diff --git a/drivers/scsi/mpi3mr/mpi3mr_os.c b/drivers/scsi/mpi3mr/mpi3mr_os.c
> index 7b86152922bad..39624fae9131f 100644
> --- a/drivers/scsi/mpi3mr/mpi3mr_os.c
> +++ b/drivers/scsi/mpi3mr/mpi3mr_os.c
> @@ -3426,8 +3426,7 @@ void mpi3mr_process_op_reply_desc(struct mpi3mr_ioc *mrioc,
> scsi_reply = mpi3mr_get_reply_virt_addr(mrioc,
> *reply_dma);
> if (!scsi_reply) {
> - panic("%s: scsi_reply is NULL, this shouldn't happen\n",
> - mrioc->name);
> + ioc_err(mrioc, "scsi_reply is NULL, invalid reply_frame_address\n");
> goto out;
[Severity: Critical]
When mpi3mr_get_reply_virt_addr() returns NULL, we log an error and jump
to the 'out' label, but *reply_dma remains populated with the unvalidated
address.
In the caller mpi3mr_process_op_reply_q(), we have:
mpi3mr_process_op_reply_desc(mrioc, reply_desc, &reply_dma, reply_qidx);
if (reply_dma)
mpi3mr_repost_reply_buf(mrioc, reply_dma);
Does this mean we will unknowingly repost this unvalidated DMA address back
to the hardware, allowing arbitrary hardware DMA memory corruption?
> }
> host_tag = le16_to_cpu(scsi_reply->host_tag);
--
Sashiko AI review · https://sashiko.dev/#/patchset/20260724102505.115136-1-ranjan.kumar@broadcom.com?part=5
^ permalink raw reply [flat|nested] 14+ messages in thread
* [PATCH v3 06/10] mpi3mr: Fix memory leak on operational queue creation failure
2026-07-24 10:24 [PATCH v3 00/10] mpi3mr: Few Enhancements and minor fixes Ranjan Kumar
` (4 preceding siblings ...)
2026-07-24 10:25 ` [PATCH v3 05/10] mpi3mr: Fix performance regression caused by extended IRQ poll sleep Ranjan Kumar
@ 2026-07-24 10:25 ` Ranjan Kumar
2026-07-24 10:51 ` sashiko-bot
2026-07-24 10:25 ` [PATCH v3 07/10] mpi3mr: Fix firmware event reference leak during cleanup Ranjan Kumar
` (3 subsequent siblings)
9 siblings, 1 reply; 14+ messages in thread
From: Ranjan Kumar @ 2026-07-24 10:25 UTC (permalink / raw)
To: linux-scsi, martin.petersen
Cc: sathya.prakash, chandrakanth.patil, vishakhavc, ipylypiv,
Ranjan Kumar, Sashiko
When operational queue creation fails after one or more queues have
been created, the error path frees the queue information arrays but
does not release the DMA memory segments associated with the created
queues, resulting in a memory leak. Fix this by ensuring that partially
allocated segments are freed immediately if a queue fails to create.
Furthermore, the error handling path leaves dangling pointers in
mrioc->intr_info[*].op_reply_q. If a spurious interrupt fires on an IRQ
vector associated with a downgraded or freed queue, it leads to a
Use-After-Free or NULL pointer dereference. Resolve this by ensuring
the corresponding intr_info pointer is cleared whenever a reply queue's
memory segments are freed.
Finally, harden the error handling path by checking for NULL pointers
before freeing segments, and reset the operational queue counts to zero
on failure. This prevents a deferred kernel panic during driver cleanup
if queue creation fails during a controller reset.
Reported-by: Sashiko <sashiko-bot@kernel.org>
Closes: https://sashiko.dev/#/patchset/20260626114109.43685-1-ranjan.kumar@broadcom.com?part=6
Closes: https://sashiko.dev/#/patchset/20260708183305.244485-1-ranjan.kumar@broadcom.com?part=6
Signed-off-by: Chandrakanth Patil <chandrakanth.patil@broadcom.com>
Signed-off-by: Ranjan Kumar <ranjan.kumar@broadcom.com>
---
drivers/scsi/mpi3mr/mpi3mr_fw.c | 26 ++++++++++++++++++++++----
1 file changed, 22 insertions(+), 4 deletions(-)
diff --git a/drivers/scsi/mpi3mr/mpi3mr_fw.c b/drivers/scsi/mpi3mr/mpi3mr_fw.c
index 9f7cee26ebcd..370ad8568117 100644
--- a/drivers/scsi/mpi3mr/mpi3mr_fw.c
+++ b/drivers/scsi/mpi3mr/mpi3mr_fw.c
@@ -2020,6 +2020,10 @@ static void mpi3mr_free_op_reply_q_segments(struct mpi3mr_ioc *mrioc, u16 q_idx)
u16 j;
int size;
struct segments *segments;
+ u16 midx = REPLY_QUEUE_IDX_TO_MSIX_IDX(q_idx, mrioc->op_reply_q_offset);
+
+ if (midx < mrioc->intr_info_count)
+ mrioc->intr_info[midx].op_reply_q = NULL;
segments = mrioc->op_reply_qinfo[q_idx].q_segments;
if (!segments)
@@ -2504,7 +2508,7 @@ static int mpi3mr_create_op_req_q(struct mpi3mr_ioc *mrioc, u16 idx,
static int mpi3mr_create_op_queues(struct mpi3mr_ioc *mrioc)
{
int retval = 0;
- u16 num_queues = 0, i = 0, msix_count_op_q = 1;
+ u16 num_queues = 0, i = 0, j = 0, msix_count_op_q = 1;
u32 ioc_status;
enum mpi3mr_iocstate ioc_state;
@@ -2556,6 +2560,13 @@ static int mpi3mr_create_op_queues(struct mpi3mr_ioc *mrioc)
}
}
+ if (i < num_queues) {
+ for (j = i; j < num_queues; j++) {
+ mpi3mr_free_op_req_q_segments(mrioc, j);
+ mpi3mr_free_op_reply_q_segments(mrioc, j);
+ }
+ }
+
if (i == 0) {
/* Not even one queue is created successfully*/
retval = -1;
@@ -2577,11 +2588,18 @@ static int mpi3mr_create_op_queues(struct mpi3mr_ioc *mrioc)
return retval;
out_failed:
- kfree(mrioc->req_qinfo);
- mrioc->req_qinfo = NULL;
-
+ if (mrioc->req_qinfo) {
+ for (j = 0; j < i; j++) {
+ mpi3mr_free_op_req_q_segments(mrioc, j);
+ mpi3mr_free_op_reply_q_segments(mrioc, j);
+ }
+ kfree(mrioc->req_qinfo);
+ mrioc->req_qinfo = NULL;
+ }
+ mrioc->num_op_req_q = 0;
kfree(mrioc->op_reply_qinfo);
mrioc->op_reply_qinfo = NULL;
+ mrioc->num_op_reply_q = 0;
return retval;
}
--
2.47.3
^ permalink raw reply related [flat|nested] 14+ messages in thread* Re: [PATCH v3 06/10] mpi3mr: Fix memory leak on operational queue creation failure
2026-07-24 10:25 ` [PATCH v3 06/10] mpi3mr: Fix memory leak on operational queue creation failure Ranjan Kumar
@ 2026-07-24 10:51 ` sashiko-bot
0 siblings, 0 replies; 14+ messages in thread
From: sashiko-bot @ 2026-07-24 10:51 UTC (permalink / raw)
To: Ranjan Kumar; +Cc: linux-scsi
Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
Pre-existing issues:
- [High] Race condition in interrupt pointer clearing causes Use-After-Free and TOCTOU NULL pointer dereferences.
- [High] Permanent DMA memory leak of `q_segment_list` when `kzalloc_objs` fails during queue allocation.
--
--- Patch [6]: [PATCH v3 06/10] mpi3mr: Fix memory leak on operational queue creation failure ---
commit 30cab2ccb258127e2dd2cf1a6d7a0d940a8b72fa
Author: Ranjan Kumar <ranjan.kumar@broadcom.com>
mpi3mr: Fix memory leak on operational queue creation failure
When operational queue creation fails after one or more queues have
been created, the error path frees the queue information arrays but
does not release the DMA memory segments associated with the created
queues, resulting in a memory leak. Fix this by ensuring that partially
allocated segments are freed immediately if a queue fails to create.
Closes: https://sashiko.dev/#/patchset/20260626114109.43685-1-ranjan.kumar@broadcom.com?part=6
Closes: https://sashiko.dev/#/patchset/20260708183305.244485-1-ranjan.kumar@broadcom.com?part=6
> diff --git a/drivers/scsi/mpi3mr/mpi3mr_fw.c b/drivers/scsi/mpi3mr/mpi3mr_fw.c
> index 9f7cee26ebcd3..370ad85681171 100644
> --- a/drivers/scsi/mpi3mr/mpi3mr_fw.c
> +++ b/drivers/scsi/mpi3mr/mpi3mr_fw.c
> @@ -2020,6 +2020,10 @@ static void mpi3mr_free_op_reply_q_segments(struct mpi3mr_ioc *mrioc, u16 q_idx)
> u16 j;
> int size;
> struct segments *segments;
> + u16 midx = REPLY_QUEUE_IDX_TO_MSIX_IDX(q_idx, mrioc->op_reply_q_offset);
> +
> + if (midx < mrioc->intr_info_count)
> + mrioc->intr_info[midx].op_reply_q = NULL;
[Severity: High]
This isn't a bug introduced by this patch, but can this clear the pointer
while an interrupt is actively accessing it?
Since we clear mrioc->intr_info[midx].op_reply_q without disabling the
IRQ or calling synchronize_irq(), a hardware interrupt firing concurrently
could access the queue locklessly.
For example, in drivers/scsi/mpi3mr/mpi3mr_fw.c:mpi3mr_isr_poll():
if (!atomic_read(&intr_info->op_reply_q->pend_ios))
break;
The pointer is evaluated unconditionally and could be set to NULL
concurrently before being read again.
Similarly, in mpi3mr_isr_primary():
if (intr_info->op_reply_q)
num_op_reply = mpi3mr_process_op_reply_q(mrioc,
intr_info->op_reply_q);
There is a race window where op_reply_q can be set to NULL after the
check, resulting in processing a queue whose segments are actively
being freed.
>
> segments = mrioc->op_reply_qinfo[q_idx].q_segments;
> if (!segments)
> return;
[Severity: High]
This is a pre-existing issue, but could this early return bypass the
cleanup for q_segment_list?
In mpi3mr_alloc_op_reply_q_segments(), if the kzalloc_objs() call for
q_segments fails, it returns without cleaning up the successfully allocated
q_segment_list:
op_reply_q->q_segments = kzalloc_objs(struct segments,
op_reply_q->num_segments);
if (!op_reply_q->q_segments)
return -ENOMEM;
Because q_segments is NULL in this failure path, the cleanup function
mpi3mr_free_op_reply_q_segments() returns early here, bypassing the
dma_free_coherent() cleanup for q_segment_list. This could cause a
permanent DMA memory leak of q_segment_list.
Does this same vulnerability exist in mpi3mr_alloc_op_req_q_segments()
and mpi3mr_free_op_req_q_segments() where similar early returns occur?
--
Sashiko AI review · https://sashiko.dev/#/patchset/20260724102505.115136-1-ranjan.kumar@broadcom.com?part=6
^ permalink raw reply [flat|nested] 14+ messages in thread
* [PATCH v3 07/10] mpi3mr: Fix firmware event reference leak during cleanup
2026-07-24 10:24 [PATCH v3 00/10] mpi3mr: Few Enhancements and minor fixes Ranjan Kumar
` (5 preceding siblings ...)
2026-07-24 10:25 ` [PATCH v3 06/10] mpi3mr: Fix memory leak on operational queue creation failure Ranjan Kumar
@ 2026-07-24 10:25 ` Ranjan Kumar
2026-07-24 10:25 ` [PATCH v3 08/10] mpi3mr: Fix SAS port allocation and registration error handling Ranjan Kumar
` (2 subsequent siblings)
9 siblings, 0 replies; 14+ messages in thread
From: Ranjan Kumar @ 2026-07-24 10:25 UTC (permalink / raw)
To: linux-scsi, martin.petersen
Cc: sathya.prakash, chandrakanth.patil, vishakhavc, ipylypiv,
Ranjan Kumar, Sashiko
During firmware event cleanup, when an event is currently executing or
pending at the SCSI mid-layer, the driver sets a discard flag and exits
the cleanup routine early. This early exit skips the normal cancel path,
resulting in the firmware event reference count not being decremented,
leading to a reference leak.
Further analysis of the firmware event handling revealed and fixed
additional concurrency issues:
1. TOCTOU Race: mpi3mr_cleanup_fwevt_list() read current_event
locklessly, allowing the worker thread to free it concurrently.
Fix this by safely acquiring the reference under fwevt_lock.
2. Use-After-Free: mpi3mr_dequeue_fwevt() dropped the event reference
before returning it to the caller. Remove this drop and add a
balancing put to the end of mpi3mr_cancel_work() so the caller
retains the reference during cancellation.
3. Soft Lockup/Deadlock: mpi3mr_fwevt_bh() temporarily dropped
fwevt_lock while moving an event from the list to current_event.
This race window allowed driver unload (rmmod) to intervene and
deadlock. Fix this by inlining the list deletion so the lock is
held continuously.
Reported-by: Sashiko <sashiko-bot@kernel.org>
Closes: https://sashiko.dev/#/patchset/20260626114109.43685-1-ranjan.kumar@broadcom.com?part=7
Closes: https://sashiko.dev/#/patchset/20260708183305.244485-1-ranjan.kumar@broadcom.com?part=7
Signed-off-by: Chandrakanth Patil <chandrakanth.patil@broadcom.com>
Signed-off-by: Ranjan Kumar <ranjan.kumar@broadcom.com>
---
drivers/scsi/mpi3mr/mpi3mr_os.c | 84 +++++++++++++++++++--------------
1 file changed, 48 insertions(+), 36 deletions(-)
diff --git a/drivers/scsi/mpi3mr/mpi3mr_os.c b/drivers/scsi/mpi3mr/mpi3mr_os.c
index 39624fae9131..545570d490fa 100644
--- a/drivers/scsi/mpi3mr/mpi3mr_os.c
+++ b/drivers/scsi/mpi3mr/mpi3mr_os.c
@@ -282,32 +282,6 @@ void mpi3mr_hdb_trigger_data_event(struct mpi3mr_ioc *mrioc,
mpi3mr_fwevt_add_to_list(mrioc, fwevt);
}
-/**
- * mpi3mr_fwevt_del_from_list - Delete firmware event from list
- * @mrioc: Adapter instance reference
- * @fwevt: Firmware event reference
- *
- * Delete the given firmware event from the firmware event list.
- *
- * Return: Nothing.
- */
-static void mpi3mr_fwevt_del_from_list(struct mpi3mr_ioc *mrioc,
- struct mpi3mr_fwevt *fwevt)
-{
- unsigned long flags;
-
- spin_lock_irqsave(&mrioc->fwevt_lock, flags);
- if (!list_empty(&fwevt->list)) {
- list_del_init(&fwevt->list);
- /*
- * Put fwevt reference count after
- * removing it from fwevt_list
- */
- mpi3mr_fwevt_put(fwevt);
- }
- spin_unlock_irqrestore(&mrioc->fwevt_lock, flags);
-}
-
/**
* mpi3mr_dequeue_fwevt - Dequeue firmware event from the list
* @mrioc: Adapter instance reference
@@ -327,11 +301,7 @@ static struct mpi3mr_fwevt *mpi3mr_dequeue_fwevt(
fwevt = list_first_entry(&mrioc->fwevt_list,
struct mpi3mr_fwevt, list);
list_del_init(&fwevt->list);
- /*
- * Put fwevt reference count after
- * removing it from fwevt_list
- */
- mpi3mr_fwevt_put(fwevt);
+
}
spin_unlock_irqrestore(&mrioc->fwevt_lock, flags);
@@ -365,6 +335,11 @@ static void mpi3mr_cancel_work(struct mpi3mr_fwevt *fwevt)
*/
mpi3mr_fwevt_put(fwevt);
}
+
+ /*
+ * Drop the reference count that was acquired by the caller.
+ */
+ mpi3mr_fwevt_put(fwevt);
}
/**
@@ -379,16 +354,39 @@ static void mpi3mr_cancel_work(struct mpi3mr_fwevt *fwevt)
void mpi3mr_cleanup_fwevt_list(struct mpi3mr_ioc *mrioc)
{
struct mpi3mr_fwevt *fwevt = NULL;
+ unsigned long flags;
+ /*
+ * Safely read current_event under lock to prevent TOCTOU race
+ * with the firmware event worker thread.
+ */
+ spin_lock_irqsave(&mrioc->fwevt_lock, flags);
if ((list_empty(&mrioc->fwevt_list) && !mrioc->current_event) ||
- !mrioc->fwevt_worker_thread)
+ !mrioc->fwevt_worker_thread) {
+ spin_unlock_irqrestore(&mrioc->fwevt_lock, flags);
return;
+ }
+ spin_unlock_irqrestore(&mrioc->fwevt_lock, flags);
while ((fwevt = mpi3mr_dequeue_fwevt(mrioc)))
mpi3mr_cancel_work(fwevt);
- if (mrioc->current_event) {
- fwevt = mrioc->current_event;
+ /*
+ * Safely read current_event under lock to prevent TOCTOU race
+ * with the firmware event worker thread.
+ */
+ spin_lock_irqsave(&mrioc->fwevt_lock, flags);
+ fwevt = mrioc->current_event;
+ if (fwevt) {
+ /*
+ * Take a reference to ensure the event is not freed by the
+ * worker thread while we are evaluating or cancelling it.
+ */
+ mpi3mr_fwevt_get(fwevt);
+ }
+ spin_unlock_irqrestore(&mrioc->fwevt_lock, flags);
+
+ if (fwevt) {
/*
* Don't call cancel_work_sync() API for the
* fwevt work if the controller reset is
@@ -399,6 +397,7 @@ void mpi3mr_cleanup_fwevt_list(struct mpi3mr_ioc *mrioc)
*/
if (current_work() == &fwevt->work || fwevt->pending_at_sml) {
fwevt->discard = 1;
+ mpi3mr_fwevt_put(fwevt);
return;
}
@@ -2129,9 +2128,19 @@ static void mpi3mr_fwevt_bh(struct mpi3mr_ioc *mrioc,
u16 perst_id, handle, dev_info;
struct mpi3_device0_sas_sata_format *sasinf = NULL;
unsigned int timeout;
+ unsigned long flags;
- mpi3mr_fwevt_del_from_list(mrioc, fwevt);
+ spin_lock_irqsave(&mrioc->fwevt_lock, flags);
+ if (!list_empty(&fwevt->list)) {
+ list_del_init(&fwevt->list);
+ /*
+ * Put fwevt reference count after
+ * removing it from fwevt_list
+ */
+ mpi3mr_fwevt_put(fwevt);
+ }
mrioc->current_event = fwevt;
+ spin_unlock_irqrestore(&mrioc->fwevt_lock, flags);
if (mrioc->stop_drv_processing) {
dprint_event_bh(mrioc, "ignoring event(0x%02x) in the bottom half handler\n"
@@ -2264,9 +2273,12 @@ static void mpi3mr_fwevt_bh(struct mpi3mr_ioc *mrioc,
mpi3mr_process_event_ack(mrioc, fwevt->event_id,
fwevt->evt_ctx);
out:
+ spin_lock_irqsave(&mrioc->fwevt_lock, flags);
+ mrioc->current_event = NULL;
+ spin_unlock_irqrestore(&mrioc->fwevt_lock, flags);
+
/* Put fwevt reference count to neutralize kref_init increment */
mpi3mr_fwevt_put(fwevt);
- mrioc->current_event = NULL;
}
/**
--
2.47.3
^ permalink raw reply related [flat|nested] 14+ messages in thread* [PATCH v3 08/10] mpi3mr: Fix SAS port allocation and registration error handling
2026-07-24 10:24 [PATCH v3 00/10] mpi3mr: Few Enhancements and minor fixes Ranjan Kumar
` (6 preceding siblings ...)
2026-07-24 10:25 ` [PATCH v3 07/10] mpi3mr: Fix firmware event reference leak during cleanup Ranjan Kumar
@ 2026-07-24 10:25 ` Ranjan Kumar
2026-07-24 10:25 ` [PATCH v3 09/10] mpi3mr: Fix SAS PHY cleanup in host addition error paths Ranjan Kumar
2026-07-24 10:25 ` [PATCH v3 10/10] mpi3mr: Driver version update to 8.18.0.8.50 Ranjan Kumar
9 siblings, 0 replies; 14+ messages in thread
From: Ranjan Kumar @ 2026-07-24 10:25 UTC (permalink / raw)
To: linux-scsi, martin.petersen
Cc: sathya.prakash, chandrakanth.patil, vishakhavc, ipylypiv,
Ranjan Kumar, Sashiko
During SAS port creation, the driver does not verify successful port
allocation before attempting registration, which can lead to a NULL
pointer dereference. Additionally, if registration fails, the allocated
port is not freed, resulting in a memory leak.
Fix this by adding a NULL check after allocation and freeing the port
when registration fails.
Additional fixes in the error handling path include:
1. Fixing similar missing NULL checks for rphy allocations.
2. Ensuring the target device reference is released in the error path
to prevent resource leaks.
3. Fixing a double-free vulnerability in the sas_rphy_add() error path.
Calling sas_port_delete() automatically cleans up the attached rphy,
so explicitly calling sas_rphy_free() immediately afterward resulted
in a double-free. Remove the redundant sas_rphy_free() call.
Reported-by: Sashiko <sashiko-bot@kernel.org>
Closes: https://sashiko.dev/#/patchset/20260626114109.43685-1-ranjan.kumar@broadcom.com?part=8
Closes: https://sashiko.dev/#/patchset/20260708183305.244485-1-ranjan.kumar@broadcom.com?part=8
Signed-off-by: Chandrakanth Patil <chandrakanth.patil@broadcom.com>
Signed-off-by: Ranjan Kumar <ranjan.kumar@broadcom.com>
---
drivers/scsi/mpi3mr/mpi3mr_transport.c | 23 +++++++++++++++++++++++
1 file changed, 23 insertions(+)
diff --git a/drivers/scsi/mpi3mr/mpi3mr_transport.c b/drivers/scsi/mpi3mr/mpi3mr_transport.c
index 240f67a8e2e3..00f115ef3a47 100644
--- a/drivers/scsi/mpi3mr/mpi3mr_transport.c
+++ b/drivers/scsi/mpi3mr/mpi3mr_transport.c
@@ -1428,9 +1428,15 @@ static struct mpi3mr_sas_port *mpi3mr_sas_port_add(struct mpi3mr_ioc *mrioc,
}
port = sas_port_alloc_num(mr_sas_node->parent_dev);
+ if (!port) {
+ ioc_err(mrioc, "failure at %s:%d/%s() (sas_port_alloc)!\n",
+ __FILE__, __LINE__, __func__);
+ goto out_fail;
+ }
if ((sas_port_add(port))) {
ioc_err(mrioc, "failure at %s:%d/%s()!\n",
__FILE__, __LINE__, __func__);
+ sas_port_free(port);
goto out_fail;
}
@@ -1450,10 +1456,22 @@ static struct mpi3mr_sas_port *mpi3mr_sas_port_add(struct mpi3mr_ioc *mrioc,
mr_sas_port->port = port;
if (mr_sas_port->remote_identify.device_type == SAS_END_DEVICE) {
rphy = sas_end_device_alloc(port);
+ if (!rphy) {
+ ioc_err(mrioc, "failure at %s:%d/%s() (sas_end_device_alloc)!\n",
+ __FILE__, __LINE__, __func__);
+ sas_port_delete(port);
+ goto out_fail;
+ }
tgtdev->dev_spec.sas_sata_inf.rphy = rphy;
} else {
rphy = sas_expander_alloc(port,
mr_sas_port->remote_identify.device_type);
+ if (!rphy) {
+ ioc_err(mrioc, "failure at %s:%d/%s() (sas_expander_alloc)!\n",
+ __FILE__, __LINE__, __func__);
+ sas_port_delete(port);
+ goto out_fail;
+ }
}
rphy->identify = mr_sas_port->remote_identify;
@@ -1463,6 +1481,8 @@ static struct mpi3mr_sas_port *mpi3mr_sas_port_add(struct mpi3mr_ioc *mrioc,
if ((sas_rphy_add(rphy))) {
ioc_err(mrioc, "failure at %s:%d/%s()!\n",
__FILE__, __LINE__, __func__);
+ sas_port_delete(port);
+ goto out_fail;
}
if (mr_sas_port->remote_identify.device_type == SAS_END_DEVICE) {
tgtdev->dev_spec.sas_sata_inf.pend_sas_rphy_add = 0;
@@ -1498,6 +1518,9 @@ static struct mpi3mr_sas_port *mpi3mr_sas_port_add(struct mpi3mr_ioc *mrioc,
return mr_sas_port;
out_fail:
+ if (tgtdev)
+ mpi3mr_tgtdev_put(tgtdev);
+
list_for_each_entry_safe(mr_sas_phy, next, &mr_sas_port->phy_list,
port_siblings)
list_del(&mr_sas_phy->port_siblings);
--
2.47.3
^ permalink raw reply related [flat|nested] 14+ messages in thread* [PATCH v3 09/10] mpi3mr: Fix SAS PHY cleanup in host addition error paths
2026-07-24 10:24 [PATCH v3 00/10] mpi3mr: Few Enhancements and minor fixes Ranjan Kumar
` (7 preceding siblings ...)
2026-07-24 10:25 ` [PATCH v3 08/10] mpi3mr: Fix SAS port allocation and registration error handling Ranjan Kumar
@ 2026-07-24 10:25 ` Ranjan Kumar
2026-07-24 10:25 ` [PATCH v3 10/10] mpi3mr: Driver version update to 8.18.0.8.50 Ranjan Kumar
9 siblings, 0 replies; 14+ messages in thread
From: Ranjan Kumar @ 2026-07-24 10:25 UTC (permalink / raw)
To: linux-scsi, martin.petersen
Cc: sathya.prakash, chandrakanth.patil, vishakhavc, ipylypiv,
Ranjan Kumar, Sashiko
When adding a SAS host, the driver allocates a PHY array and subsequently
creates individual SAS PHYs. If a later step fails, the error path exits
without cleaning up previously allocated resources, resulting in leaks of
both the PHY array and any registered SAS PHYs.
Additionally, the return value of mpi3mr_add_host_phy() was being ignored.
If it failed, mr_sas_phy->phy would be left as NULL, which could later
lead to a NULL pointer dereference in mpi3mr_sas_port_add() when the
attached device triggers a device addition event.
Add a dedicated cleanup path that deletes any successfully created SAS
PHYs and frees the PHY array before returning from initialization
failure paths. Also, check the return value of mpi3mr_add_host_phy()
and jump to the cleanup path on failure.
Reported-by: Sashiko <sashiko-bot@kernel.org>
Closes: https://sashiko.dev/#/patchset/20260626114109.43685-1-ranjan.kumar@broadcom.com?part=9
Signed-off-by: Chandrakanth Patil <chandrakanth.patil@broadcom.com>
Signed-off-by: Ranjan Kumar <ranjan.kumar@broadcom.com>
---
drivers/scsi/mpi3mr/mpi3mr_transport.c | 33 ++++++++++++++++++--------
1 file changed, 23 insertions(+), 10 deletions(-)
diff --git a/drivers/scsi/mpi3mr/mpi3mr_transport.c b/drivers/scsi/mpi3mr/mpi3mr_transport.c
index 00f115ef3a47..b44e36f84e0e 100644
--- a/drivers/scsi/mpi3mr/mpi3mr_transport.c
+++ b/drivers/scsi/mpi3mr/mpi3mr_transport.c
@@ -1216,13 +1216,14 @@ void mpi3mr_sas_host_add(struct mpi3mr_ioc *mrioc)
}
num_phys = sas_io_unit_pg0->num_phys;
kfree(sas_io_unit_pg0);
+ sas_io_unit_pg0 = NULL;
mrioc->sas_hba.host_node = 1;
INIT_LIST_HEAD(&mrioc->sas_hba.sas_port_list);
mrioc->sas_hba.parent_dev = &mrioc->shost->shost_gendev;
mrioc->sas_hba.phy = kzalloc_objs(struct mpi3mr_sas_phy, num_phys);
if (!mrioc->sas_hba.phy)
- return;
+ goto out;
mrioc->sas_hba.num_phys = num_phys;
@@ -1230,12 +1231,12 @@ void mpi3mr_sas_host_add(struct mpi3mr_ioc *mrioc)
(num_phys * sizeof(struct mpi3_sas_io_unit0_phy_data));
sas_io_unit_pg0 = kzalloc(sz, GFP_KERNEL);
if (!sas_io_unit_pg0)
- return;
+ goto out_free_phy;
if (mpi3mr_cfg_get_sas_io_unit_pg0(mrioc, sas_io_unit_pg0, sz)) {
ioc_err(mrioc, "failure at %s:%d/%s()!\n",
__FILE__, __LINE__, __func__);
- goto out;
+ goto out_free_phy;
}
mrioc->sas_hba.handle = 0;
@@ -1249,12 +1250,12 @@ void mpi3mr_sas_host_add(struct mpi3mr_ioc *mrioc)
MPI3_SAS_PHY_PGAD_FORM_PHY_NUMBER, i)) {
ioc_err(mrioc, "failure at %s:%d/%s()!\n",
__FILE__, __LINE__, __func__);
- goto out;
+ goto out_free_phy;
}
if (ioc_status != MPI3_IOCSTATUS_SUCCESS) {
ioc_err(mrioc, "failure at %s:%d/%s()!\n",
__FILE__, __LINE__, __func__);
- goto out;
+ goto out_free_phy;
}
if (!mrioc->sas_hba.handle)
@@ -1264,26 +1265,27 @@ void mpi3mr_sas_host_add(struct mpi3mr_ioc *mrioc)
if (!(mpi3mr_get_hba_port_by_id(mrioc, port_id)))
if (!mpi3mr_alloc_hba_port(mrioc, port_id))
- goto out;
+ goto out_free_phy;
mrioc->sas_hba.phy[i].handle = mrioc->sas_hba.handle;
mrioc->sas_hba.phy[i].phy_id = i;
mrioc->sas_hba.phy[i].hba_port =
mpi3mr_get_hba_port_by_id(mrioc, port_id);
- mpi3mr_add_host_phy(mrioc, &mrioc->sas_hba.phy[i],
- phy_pg0, mrioc->sas_hba.parent_dev);
+ if (mpi3mr_add_host_phy(mrioc, &mrioc->sas_hba.phy[i],
+ phy_pg0, mrioc->sas_hba.parent_dev))
+ goto out_free_phy;
}
if ((mpi3mr_cfg_get_dev_pg0(mrioc, &ioc_status, &dev_pg0,
sizeof(dev_pg0), MPI3_DEVICE_PGAD_FORM_HANDLE,
mrioc->sas_hba.handle))) {
ioc_err(mrioc, "%s: device page0 read failed\n", __func__);
- goto out;
+ goto out_free_phy;
}
if (ioc_status != MPI3_IOCSTATUS_SUCCESS) {
ioc_err(mrioc, "device page read failed for handle(0x%04x), with ioc_status(0x%04x) failure at %s:%d/%s()!\n",
mrioc->sas_hba.handle, ioc_status, __FILE__, __LINE__,
__func__);
- goto out;
+ goto out_free_phy;
}
mrioc->sas_hba.enclosure_handle =
le16_to_cpu(dev_pg0.enclosure_handle);
@@ -1306,6 +1308,17 @@ void mpi3mr_sas_host_add(struct mpi3mr_ioc *mrioc)
le64_to_cpu(encl_pg0.enclosure_logical_id);
}
+ goto out;
+
+out_free_phy:
+ for (i = 0; i < mrioc->sas_hba.num_phys; i++) {
+ if (mrioc->sas_hba.phy[i].phy)
+ sas_phy_delete(mrioc->sas_hba.phy[i].phy);
+ }
+ kfree(mrioc->sas_hba.phy);
+ mrioc->sas_hba.phy = NULL;
+ mrioc->sas_hba.num_phys = 0;
+
out:
kfree(sas_io_unit_pg0);
}
--
2.47.3
^ permalink raw reply related [flat|nested] 14+ messages in thread* [PATCH v3 10/10] mpi3mr: Driver version update to 8.18.0.8.50
2026-07-24 10:24 [PATCH v3 00/10] mpi3mr: Few Enhancements and minor fixes Ranjan Kumar
` (8 preceding siblings ...)
2026-07-24 10:25 ` [PATCH v3 09/10] mpi3mr: Fix SAS PHY cleanup in host addition error paths Ranjan Kumar
@ 2026-07-24 10:25 ` Ranjan Kumar
9 siblings, 0 replies; 14+ messages in thread
From: Ranjan Kumar @ 2026-07-24 10:25 UTC (permalink / raw)
To: linux-scsi, martin.petersen
Cc: sathya.prakash, chandrakanth.patil, vishakhavc, ipylypiv,
Ranjan Kumar
Update driver version to 8.18.0.8.50
Signed-off-by: Ranjan Kumar <ranjan.kumar@broadcom.com>
---
drivers/scsi/mpi3mr/mpi3mr.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/scsi/mpi3mr/mpi3mr.h b/drivers/scsi/mpi3mr/mpi3mr.h
index 1d11d7c69536..c6bbd6b33cfe 100644
--- a/drivers/scsi/mpi3mr/mpi3mr.h
+++ b/drivers/scsi/mpi3mr/mpi3mr.h
@@ -56,8 +56,8 @@ extern struct list_head mrioc_list;
extern int prot_mask;
extern atomic64_t event_counter;
-#define MPI3MR_DRIVER_VERSION "8.17.0.3.50"
-#define MPI3MR_DRIVER_RELDATE "09-January-2026"
+#define MPI3MR_DRIVER_VERSION "8.18.0.8.50"
+#define MPI3MR_DRIVER_RELDATE "26-June-2026"
#define MPI3MR_DRIVER_NAME "mpi3mr"
#define MPI3MR_DRIVER_LICENSE "GPL"
--
2.47.3
^ permalink raw reply related [flat|nested] 14+ messages in thread