scsi: qla2xxx: Remove unnecessary call to memset
Subsystems (4)
PATCH 1/1 scsi: qla2xxx: Remove unnecessary call to memset
In qla25xx_set_els_cmds_supported(), a call is made to dma_alloc_coherent() followed by zeroing the memory with memset. This is unnecessary as dma_alloc_coherent() already zeros memory. Remove.
Issue identified with Coccinelle.
--- drivers/scsi/qla2xxx/qla_mbx.c | 2 -- 1 file changed, 2 deletions(-)
drivers/scsi/qla2xxx/qla_mbx.c +0 -2
@@ -4925,8 +4925,6 @@ qla25xx_set_els_cmds_supported(scsi_qla_host_t *vha)return QLA_MEMORY_ALLOC_FAILED ; } -memset (els_cmd_map,0 ,ELS_CMD_MAP_SIZE ); -/* List of Purex ELS */ cmd_opcode[0 ] =ELS_FPIN ; cmd_opcode[1 ] =ELS_RDP ;
Signature · 1 line
2.28.0
PATCH v2 1/1 scsi: Don't call memset after dma_alloc_coherent()
dma_alloc_coherent() already zeroes memory, so the extra call to memset() is unnecessary.
Issue identified with Coccinelle.
---
v2: I've noticed a few other places in the scsi subsystem with this pattern, so lets fix them all with one patch.
--- drivers/scsi/mpt3sas/mpt3sas_base.c | 1 - drivers/scsi/mvsas/mv_init.c | 4 ---- drivers/scsi/pmcraid.c | 1 - drivers/scsi/qla2xxx/qla_mbx.c | 2 -- 4 files changed, 8 deletions(-)
drivers/scsi/mpt3sas/mpt3sas_base.c +0 -1
@@ -5259,7 +5259,6 @@ _base_allocate_memory_pools(struct MPT3SAS_ADAPTER *ioc)_base_release_memory_pools (ioc); goto retry_allocation; } -memset (ioc->request,0 , sz);if (retry_sz) ioc_err(ioc, "request pool: dma_alloc_coherent succeed: hba_depth(%d), chains_per_io(%d), frame_sz(%d), total(%d kb)\n",
drivers/scsi/mvsas/mv_init.c +0 -4
@@ -246,19 +246,16 @@ static int mvs_alloc(struct mvs_info *mvi, struct Scsi_Host *shost) &mvi->tx_dma, GFP_KERNEL);if (!mvi->tx) goto err_out; -memset (mvi->tx,0 ,sizeof (*mvi->tx) *MVS_CHIP_SLOT_SZ ); mvi->rx_fis =dma_alloc_coherent (mvi->dev,MVS_RX_FISL_SZ , &mvi->rx_fis_dma,GFP_KERNEL );if (!mvi->rx_fis) goto err_out; -memset (mvi->rx_fis,0 ,MVS_RX_FISL_SZ ); mvi->rx =dma_alloc_coherent (mvi->dev,sizeof (*mvi->rx) * (MVS_RX_RING_SZ +1 ), &mvi->rx_dma,GFP_KERNEL );if (!mvi->rx) goto err_out; -memset (mvi->rx,0 ,sizeof (*mvi->rx) * (MVS_RX_RING_SZ +1 )); mvi->rx[0 ] =cpu_to_le32 (0xfff ); mvi->rx_cons =0xfff ;
@@ -267,7 +264,6 @@ static int mvs_alloc(struct mvs_info *mvi, struct Scsi_Host *shost) &mvi->slot_dma, GFP_KERNEL);if (!mvi->slot) goto err_out; -memset (mvi->slot,0 ,sizeof (*mvi->slot) * slot_nr); mvi->bulk_buffer = dma_alloc_coherent(mvi->dev, TRASH_BUCKET_SIZE,
drivers/scsi/pmcraid.c +0 -1
@@ -4716,7 +4716,6 @@ static int pmcraid_allocate_host_rrqs(struct pmcraid_instance *pinstance)return -ENOMEM ; } -memset (pinstance->hrrq_start[i],0 , buffer_size); pinstance->hrrq_curr[i] = pinstance->hrrq_start[i]; pinstance->hrrq_end[i] = pinstance->hrrq_start[i] +PMCRAID_MAX_CMD -1 ;
drivers/scsi/qla2xxx/qla_mbx.c +0 -2
@@ -4925,8 +4925,6 @@ qla25xx_set_els_cmds_supported(scsi_qla_host_t *vha)return QLA_MEMORY_ALLOC_FAILED ; } -memset (els_cmd_map,0 ,ELS_CMD_MAP_SIZE ); -/* List of Purex ELS */ cmd_opcode[0 ] =ELS_FPIN ; cmd_opcode[1 ] =ELS_RDP ;
Signature · 1 line
2.28.0
Re: [PATCH v2] scsi: Don't call memset after dma_alloc_coherent()
Quoted 6 lines · Alex Dewar
dma_alloc_coherent()already zeroes memory, so the extra call tomemset()is unnecessary.Issue identified with Coccinelle.
Signed-off-by Alex Dewar <[email protected]>
Gentle ping? I've just realised that this is the one we want, not the v1 I just replied to.
Quoted 79 lines
---
v2: I've noticed a few other places in the scsi subsystem with this pattern, so lets fix them all with one patch.--- drivers/scsi/mpt3sas/mpt3sas_base.c | 1 - drivers/scsi/mvsas/mv_init.c | 4 ---- drivers/scsi/pmcraid.c | 1 - drivers/scsi/qla2xxx/qla_mbx.c | 2 -- 4 files changed, 8 deletions(-)drivers/scsi/mpt3sas/mpt3sas_base.c +0
-1@@ -5259,7 +5259,6 @@ _base_allocate_memory_pools(struct MPT3SAS_ADAPTER *ioc)_base_release_memory_pools (ioc); goto retry_allocation; } -memset (ioc->request,0 , sz);if (retry_sz) ioc_err(ioc, "request pool: dma_alloc_coherent succeed: hba_depth(%d), chains_per_io(%d), frame_sz(%d), total(%d kb)\n",drivers/scsi/mvsas/mv_init.c +0
-4@@ -246,19 +246,16 @@ static int mvs_alloc(struct mvs_info *mvi, struct Scsi_Host *shost) &mvi->tx_dma, GFP_KERNEL);if (!mvi->tx) goto err_out; -memset (mvi->tx,0 ,sizeof (*mvi->tx) *MVS_CHIP_SLOT_SZ ); mvi->rx_fis =dma_alloc_coherent (mvi->dev,MVS_RX_FISL_SZ , &mvi->rx_fis_dma,GFP_KERNEL );if (!mvi->rx_fis) goto err_out; -memset (mvi->rx_fis,0 ,MVS_RX_FISL_SZ ); mvi->rx =dma_alloc_coherent (mvi->dev,sizeof (*mvi->rx) * (MVS_RX_RING_SZ +1 ), &mvi->rx_dma,GFP_KERNEL );if (!mvi->rx) goto err_out; -memset (mvi->rx,0 ,sizeof (*mvi->rx) * (MVS_RX_RING_SZ +1 )); mvi->rx[0 ] =cpu_to_le32 (0xfff ); mvi->rx_cons =0xfff ;@@ -267,7 +264,6 @@ static int mvs_alloc(struct mvs_info *mvi, struct Scsi_Host *shost) &mvi->slot_dma, GFP_KERNEL);if (!mvi->slot) goto err_out; -memset (mvi->slot,0 ,sizeof (*mvi->slot) * slot_nr); mvi->bulk_buffer = dma_alloc_coherent(mvi->dev, TRASH_BUCKET_SIZE,drivers/scsi/pmcraid.c +0
-1@@ -4716,7 +4716,6 @@ static int pmcraid_allocate_host_rrqs(struct pmcraid_instance *pinstance)return -ENOMEM ; } -memset (pinstance->hrrq_start[i],0 , buffer_size); pinstance->hrrq_curr[i] = pinstance->hrrq_start[i]; pinstance->hrrq_end[i] = pinstance->hrrq_start[i] +PMCRAID_MAX_CMD -1 ;drivers/scsi/qla2xxx/qla_mbx.c +0
-2@@ -4925,8 +4925,6 @@ qla25xx_set_els_cmds_supported(scsi_qla_host_t *vha)return QLA_MEMORY_ALLOC_FAILED ; } -memset (els_cmd_map,0 ,ELS_CMD_MAP_SIZE ); -/* List of Purex ELS */ cmd_opcode[0 ] =ELS_FPIN ; cmd_opcode[1 ] =ELS_RDP ;
Re: [PATCH v2] scsi: Don't call memset after dma_alloc_coherent()
In reply to Alex DewarAlex,
Quoted 2 lines
dma_alloc_coherent()already zeroes memory, so the extra call tomemset()is unnecessary.
One patch per driver, please.
Thanks!
Signature · 1 line
Martin K. Petersen Oracle Linux Engineering
Re: [PATCH v2] scsi: Don't call memset after dma_alloc_coherent()
Quoted 7 lines · Martin K. Petersen
Alex,
Quoted 2 lines
dma_alloc_coherent()already zeroes memory, so the extra call tomemset()is unnecessary.One patch per driver, please.
There's a single patch for QLA2XXX already submitted: https://lore.kernel.org/lkml/[email protected]/
I'll send separate patches for the other cases.
Quoted 5 lines
Thanks!
Signature · 1 line
Martin K. Petersen Oracle Linux Engineering
Re: [PATCH v2] scsi: Don't call memset after dma_alloc_coherent()
In reply to Martin K. PetersenQuoted 7 lines · Martin K. Petersen
Alex,
Quoted 2 lines
dma_alloc_coherent()already zeroes memory, so the extra call tomemset()is unnecessary.One patch per driver, please.
Nvm, someone's already beaten me to the punch!
Quoted 5 lines
Thanks!
Signature · 1 line
Martin K. Petersen Oracle Linux Engineering
Re: [PATCH v2] scsi: Don't call memset after dma_alloc_coherent()
Alex,
Quoted 1 line
Nvm, someone's already beaten me to the punch!
Yep, that one was already fixed up. Thanks!
Signature · 1 line
Martin K. Petersen Oracle Linux Engineering
Quoted 7 lines · Alex Dewar
In
qla25xx_set_els_cmds_supported(), a call is made todma_alloc_coherent()followed by zeroing the memory with memset. This is unnecessary asdma_alloc_coherent()already zeros memory. Remove.Issue identified with Coccinelle.
Signed-off-by Alex Dewar <[email protected]>
Gentle ping?
Quoted 17 lines
--- drivers/scsi/qla2xxx/qla_mbx.c | 2 -- 1 file changed, 2 deletions(-)drivers/scsi/qla2xxx/qla_mbx.c +0
-2@@ -4925,8 +4925,6 @@ qla25xx_set_els_cmds_supported(scsi_qla_host_t *vha)return QLA_MEMORY_ALLOC_FAILED ; } -memset (els_cmd_map,0 ,ELS_CMD_MAP_SIZE ); -/* List of Purex ELS */ cmd_opcode[0 ] =ELS_FPIN ; cmd_opcode[1 ] =ELS_RDP ;
Quoted 14 lines · Alex Dewar
Quoted 7 lines · Alex Dewar
In
qla25xx_set_els_cmds_supported(), a call is made todma_alloc_coherent()followed by zeroing the memory with memset. This is unnecessary asdma_alloc_coherent()already zeros memory. Remove.Issue identified with Coccinelle.
Signed-off-by Alex Dewar <[email protected]>Gentle ping?
Quoted 4 lines
--- drivers/scsi/qla2xxx/qla_mbx.c | 2 -- 1 file changed, 2 deletions(-)
Quoted 14 lines
drivers/scsi/qla2xxx/qla_mbx.c qla25xx_set_els_cmds_supported(scsi_qla_host_t *vha) · line 4930
Quoted 12 lines
drivers/scsi/qla2xxx/qla_mbx.c +0
-1@@ -4925,8 +4925,6 @@ qla25xx_set_els_cmds_supported(scsi_qla_host_t *vha)return QLA_MEMORY_ALLOC_FAILED ; } -memset (els_cmd_map,0 ,ELS_CMD_MAP_SIZE ); -/* List of Purex ELS */ cmd_opcode[0 ] =ELS_FPIN ; cmd_opcode[1 ] =ELS_RDP ;
Looks good.
Regards,
-Arun