On Tue, 2019-10-08 at 08:21 +0200, Hannes Reinecke wrote:
> On 10/7/19 10:45 PM, Ewan D. Milne wrote:
> >
> > The patch itself looks OK, but I was wondering about a couple of things:
> >
> > - There are other places in scsi_dh_alua where the ASC/ASCQ 04 0A is
>
See below.
On Mon, 2019-10-07 at 15:57 +0200, Hannes Reinecke wrote:
> From: Hannes Reinecke
>
> Some arrays are not capable of returning RTPG data during state
> transitioning, but rather return an 'LUN not accessible, asymmetric
> access state transition' sense code. In these cases we
> can se
N "10.01.00.19-k"
>
> #define QLA_DRIVER_MAJOR_VER 10
> #define QLA_DRIVER_MINOR_VER 1
Reviewed-by: Ewan D. Milne
= rp->id.b24 ||
> + fcport->scan_needed) {
> qlt_schedule_sess_for_deletion(fcport);
> }
> fcport->d_id.b24 = rp->id.b24;
> + fcport->scan_needed = 0;
> break;
> }
>
Reviewed-by: Ewan D. Milne
ort->plogi_nack_done_deadline)) {
> - set_bit(RELOGIN_NEEDED, &vha->dpc_flags);
> - return;
> - }
> - }
> -
> if (fcport->last_rscn_gen != fcport->rscn_gen) {
> ql_dbg(ql_dbg_disc, vha, 0x20e9, "%s %d %8phC post gnl\n",
> __func__, __LINE__, fcport->port_name);
Reviewed-by: Ewan D. Milne
rvdata(ha->pdev);
>
> - qla82xx_mbx_intr_disable(vha);
> + if (ha->interrupts_on)
> + qla82xx_mbx_intr_disable(vha);
>
> spin_lock_irq(&ha->hardware_lock);
> if (IS_QLA8044(ha))
Reviewed-by: Ewan D. Milne
ptr = (void *)req->ring;
> - struct qla_fdt_layout *fdt = (void *)req->ring;
> + struct qla_fdt_layout *fdt = (struct qla_fdt_layout *)req->ring;
> uint8_t man_id, flash_id;
> uint16_t mid = 0, fid = 0;
>
> - qla24xx_read_flash_data(vha, (void *)fdt, ha->flt_region_fdt,
> + ha->isp_ops->read_optrom(vha, fdt, ha->flt_region_fdt << 2,
> OPTROM_BURST_DWORDS);
> if (le16_to_cpu(*wptr) == 0x)
> goto no_flash_data;
Reviewed-by: Ewan D. Milne
"with %d vectors, using %d vectors.\n",
> - ha->msix_count, ret);
> + ql_log(ql_log_info, vha, 0x00c6,
> + "MSI-X: Using %d vectors\n", ret);
> ha->msix_count = ret;
> /* Recalculate queue values */
> if (ha->mqiobase && (ql2xmqsupport || ql2xnvmeenable)) {
Reviewed-by: Ewan D. Milne
n-1;
> diff --git a/drivers/scsi/lpfc/lpfc_sli4.h b/drivers/scsi/lpfc/lpfc_sli4.h
> index 3aeca387b22a..329f7aa7e169 100644
> --- a/drivers/scsi/lpfc/lpfc_sli4.h
> +++ b/drivers/scsi/lpfc/lpfc_sli4.h
> @@ -44,6 +44,11 @@
> #define LPFC_HBA_HDWQ_MAX128
> #define LPFC_HBA_HDWQ_DEF0
>
> +/* FCP MQ queue count limiting */
> +#define LPFC_FCP_MQ_THRESHOLD_MIN0
> +#define LPFC_FCP_MQ_THRESHOLD_MAX128
> +#define LPFC_FCP_MQ_THRESHOLD_DEF8
> +
> /* Common buffer size to accomidate SCSI and NVME IO buffers */
> #define LPFC_COMMON_IO_BUF_SZ768
>
Looks good.
Reviewed-by: Ewan D. Milne
{
> - sdev_printk(KERN_ERR, cmd->device,
> + scmd_printk(KERN_ERR, cmd,
> "timing out command, waited %lus\n",
> wait_for/HZ);
> disposition = SUCCESS;
Reviewed-by: Ewan D. Milne
p=%x\n", cmd, ctx, cmd->cmnd[0]);
> + "queued cmd %p, ctx %p, op=%x\n", cmd, ctx, op);
>
> spin_unlock_irqrestore(&adapter->hw_lock, flags);
>
> - pvscsi_kick_io(adapter, cmd->cmnd[0]);
> + pvscsi_kick_io(adapter, op);
>
> return 0;
> }
Reviewed-by: Ewan D. Milne
On Tue, 2019-06-18 at 12:51 +0200, Hannes Reinecke wrote:
> On 6/15/19 12:10 AM, Himanshu Madhani wrote:
> > From: Arun Easi
> >
> > BUG: unable to handle kernel NULL pointer dereference at (null)
> > IP: [] qla_nvme_unregister_remote_port+0x6c/0xf0 [qla2xxx]
> > PGD 80084cf41067 PU
>
> static struct nvme_fc_port_template qla_nvme_fc_transport = {
> @@ -603,7 +608,8 @@ static void qla_nvme_unregister_remote_port(struct
> work_struct *work)
> return;
>
> ql_log(ql_log_warn, NULL, 0x2112,
> - "%s: unregister remoteport on %p\n",__func__, fcport);
> + "%s: unregister remoteport on %p %8phN\n",
> + __func__, fcport, fcport->port_name);
>
> nvme_fc_set_remoteport_devloss(fcport->nvme_remote_port, 0);
> init_completion(&fcport->nvme_del_done);
Reviewed-by: Ewan D. Milne
See below.
On Fri, 2019-06-14 at 15:10 -0700, Himanshu Madhani wrote:
> From: Arun Easi
>
> BUG: unable to handle kernel NULL pointer dereference at (null)
> IP: [] qla_nvme_unregister_remote_port+0x6c/0xf0 [qla2xxx]
> PGD 80084cf41067 PUD 84d288067 PMD 0
> Oops: [#1] SMP
> Ca
On Mon, 2019-06-10 at 11:40 -0700, James Bottomley wrote:
> On Mon, 2019-06-10 at 23:03 +0800, Ming Lei wrote:
> > The current way isn't safe for chained sgl, so use sgl helper to
> > operate sgl.
>
> This also isn't a chained driver. However, this driver seems to
> achieve this by magic number m
On Mon, 2019-06-10 at 11:37 -0700, James Bottomley wrote:
> On Mon, 2019-06-10 at 23:03 +0800, Ming Lei wrote:
> > The current way isn't safe for chained sgl, so use sgl helper to
> > operate sgl.
>
> The advansys driver doesn't currently use a chained scatterlist. In
> theory it could; the
>
>
r_32_bits(busaddr));
> m_sg->baseaddr_h = cpu_to_le32(upper_32_bits(busaddr));
> m_sg->flags = 0;
> - sgd_setsz(mhba, m_sg, cpu_to_le32(sg_dma_len(&sg[i])));
> + sgd_setsz(mhba, m_sg, cpu_to_le32(sg_dma_len(sg)));
> if ((i + 1) == *sg_count)
> m_sg->flags |= 1U << mhba->eot_flag;
>
Reviewed-by: Ewan D. Milne
= sg_dma_address(sgel);
> cnt = sg_dma_len(sgel);
> sgl->addr_hi = putPaddrHigh(physaddr);
Looks right for rsp->sg
Reviewed by: Ewan D. Milne
erlist[i]));
> + ioadl64[i].data_len = cpu_to_be32(sg_dma_len(sg));
> + ioadl64[i].address = cpu_to_be64(sg_dma_address(sg));
> }
>
> ioadl64[i-1].flags |= cpu_to_be32(IPR_IOADL_FLAGS_LAST);
Reviewed-by: Ewan D. Milne
lock->sg_ptr = 0L; /* Last ADV_SG_BLOCK in
> list. */
> return ADV_SUCCESS;
> }
> - slp++;
> + slp = sg_next(slp);
> }
> sg_block->sg_cnt = NO_OF_SG_PER_BLOCK;
> prev_sg_block = sg_block;
Reviewed-by: Ewan D. Milne
UM_SG_ENTRIES_PER_SEGMENT);
>
> sge = &ctx->sgl->sge[0];
> - for (i = 0; i < count; i++, sg++) {
> + for (i = 0; i < count; i++, sg = sg_next(sg)) {
> sge[i].addr = sg_dma_address(sg);
> sge[i].length = sg_dma_len(sg);
> sge[i].flags = 0;
Reviewed-by: Ewan D. Milne
On Fri, 2019-06-07 at 07:52 +0200, Hannes Reinecke wrote:
> On 6/6/19 7:26 PM, Bart Van Assche wrote:
> > On 6/5/19 10:46 PM, Hannes Reinecke wrote:
> > > Why not simply '-EPERM' ?
> > > Translating a state into something else seems counterproductive.
> >
> > Personally I'm OK with returning -EPER
10;
> - if (phba->sli3_options & LPFC_SLI3_BG_ENABLED)
> - lpfc_setup_bg(phba, shost);
>
> lpfc_host_attrib_init(shost);
>
Reviewed-by: Ewan D. Milne
On Thu, 2019-05-09 at 15:55 -0700, Bart Van Assche wrote:
> On 5/9/19 11:58 AM, Ewan D. Milne wrote:
> > On Thu, 2019-05-09 at 07:06 -0700, Bart Van Assche wrote:
> > > On 5/9/19 6:18 AM, Hannes Reinecke wrote:
> > > > The 'qla_tgt_wq' workqueue is used for
On Thu, 2019-05-09 at 07:06 -0700, Bart Van Assche wrote:
> On 5/9/19 6:18 AM, Hannes Reinecke wrote:
> > The 'qla_tgt_wq' workqueue is used for generic command aborts,
> > not just target-related functions. So allocate the workqueue
> > always to avoid a kernel crash when aborting commands.
>
> H
hani
> ---
> drivers/scsi/qla2xxx/qla_os.c | 221
> --
> 1 file changed, 82 insertions(+), 139 deletions(-)
>
>
Reviewed-by: Ewan D. Milne
See below.
On Mon, 2019-05-06 at 13:52 -0700, Himanshu Madhani wrote:
> From: Quinn Tran
>
> commit c7702b8c2271 ("scsi: qla2xxx: Get mutex lock before checking
> optrom_state") fixed crash while reading optrom data by adding mutex
> locking. However, there can be still case where previous WRITE
e_it_fail;
> @@ -194,6 +195,7 @@ struct gendisk {
>*/
> struct disk_part_tbl __rcu *part_tbl;
> struct hd_struct part0;
> + DECLARE_BITMAP(user_ro_bitmap, DISK_MAX_PARTS);
>
> const struct block_device_operations *fops;
> struct request_queue *queue;
> @@ -431,12 +433,13 @@ extern void del_gendisk(struct gendisk *gp);
> extern struct gendisk *get_gendisk(dev_t dev, int *partno);
> extern struct block_device *bdget_disk(struct gendisk *disk, int partno);
>
> -extern void set_device_ro(struct block_device *bdev, int flag);
> -extern void set_disk_ro(struct gendisk *disk, int flag);
> +extern void set_device_ro(struct block_device *bdev, bool state);
> +extern void set_disk_ro(struct gendisk *disk, bool state);
> +extern bool get_user_ro(struct gendisk *disk, unsigned int partno);
>
> static inline int get_disk_ro(struct gendisk *disk)
> {
> - return disk->part0.policy;
> + return disk->part0.read_only;
> }
>
> extern void disk_block_events(struct gendisk *disk);
Looks good.
Reviewed-by: Ewan D. Milne
"ELS IOCB Done -%s error hdl=%x comp_status=0x%x error
> subcode 1=0x%x error subcode 2=0x%x total_byte=0x%x\n",
> type, sp->handle, comp_status, fw_status[1], fw_status[2],
> le16_to_cpu(((struct els_sts_entry_24xx *)
Reviewed-by: Ewan D. Milne
> @@ -1163,8 +1165,6 @@ void qlt_unreg_sess(struct fc_port *sess)
> if (sess->se_sess)
> vha->hw->tgt.tgt_ops->clear_nacl_from_fcport_map(sess);
>
> - qla2x00_mark_device_lost(vha, sess, 0, 0);
> -
> sess->deleted = QLA_SESS_DELETION_IN_PROGRESS;
> sess->disc_state = DSC_DELETE_PEND;
> sess->last_rscn_gen = sess->rscn_gen;
Reviewed-by: Ewan D. Milne
lude/scsi/scsi_transport_fc.h | 3 +++
> 8 files changed, 37 insertions(+), 15 deletions(-)
>
Reviewed-by: Ewan D. Milne
gt; #define ELS_RSP_MASK 0xff
> @@ -643,6 +646,7 @@ struct fc_vft_header {
> #define ELS_CMD_RNID 0x78
> #define ELS_CMD_LIRR 0x7A
> #define ELS_CMD_LCB0x81
> +#define ELS_CMD_FPIN ELS_FPIN
> #endif
>
> /*
Reviewed-by: Ewan D. Milne
ost_fc_event(struct Scsi_Host *shost, u32
> event_number,
>* Note: when calling fc_host_post_fc_event(), vendor_id may be
>* specified as 0.
>*/
> +void fc_host_fpin_rcv(struct Scsi_Host *shost, u32 fpin_len, char *fpin_buf);
> struct fc_vport *fc_vport_create(struct Scsi_Host *shost, int channel,
> struct fc_vport_identifiers *);
> int fc_vport_terminate(struct fc_vport *vport);
Reviewed-by: Ewan D. Milne
ta_buf, u64 vendor_id);
> + u32 data_len, char *data_buf, u64 vendor_id);
> +void fc_host_post_fc_event(struct Scsi_Host *shost, u32 event_number,
> + enum fc_host_event_code event_code,
> + u32 data_len, char *data_buf, u64 vendor_id);
> /* Note: when specifying vendor_id to fc_host_post_vendor_event()
> - * be sure to read the Vendor Type and ID formatting requirements
> - * specified in scsi_netlink.h
> + * or fc_host_post_fc_event(), be sure to read the Vendor Type
> + * and ID formatting requirements specified in scsi_netlink.h
> + * Note: when calling fc_host_post_fc_event(), vendor_id may be
> + * specified as 0.
>*/
> struct fc_vport *fc_vport_create(struct Scsi_Host *shost, int channel,
> struct fc_vport_identifiers *);
Reviewed-by: Ewan D. Milne
| 8 +++
> drivers/scsi/lpfc/lpfc_hw.h | 4 ++
> drivers/scsi/scsi_transport_fc.c | 116
> +++
> include/scsi/scsi_transport_fc.h | 13 -
> include/uapi/scsi/fc/fc_els.h| 33 +++++++
> 5 files changed, 113 insertions(+), 61 deletions(-)
>
Reviewed-by: Ewan D. Milne
> + __u8fpin_cmd; /* command (0x16) */
> + __u8 fpin_zero[3]; /* specified as zero - part of cmd */
> + __be32 fpin_desc_cnt; /* count of descriptors */
> + struct fc_fn_desc fpin_desc[0]; /* Descriptor list */
> +};
> +
> #endif /* _FC_ELS_H_ */
Reviewed-by: Ewan D. Milne
P_ID)
> continue;
Would be better I think to test (ha->current_topology == ISP_CFG_NL) first
in the if statement but logically looks fine.
Reviewed-by: Ewan D. Milne
On Fri, 2019-03-15 at 15:04 -0700, Himanshu Madhani wrote:
> This patch fixes crash due to NULL pointer derefrence because
> CPU pointer is not set and used by driver. Instead, driver is
> passes CPU as tag via ha->isp_ops->{lun_reset|target_reset}
>
...
>
> Fixes: 9cf2bab630765 ("block: kill re
d be removed from the patch description.
>
> Fixes: e4db40e7a1a2 ("scsi: hisi_sas: use dma_set_mask_and_coherent")
> Cc:
> Suggested-by: Ewan D. Milne
> Signed-off-by: Hannes Reinecke
> ---
> drivers/scsi/hisi_sas/hisi_sas_main.c | 8 ++--
> 1 file chan
om the patch description.
>
> Fixes: a69b080025ea ("scsi: bfa: use dma_set_mask_and_coherent")
> Cc:
> Suggested-by: Ewan D. Milne
> Signed-off-by: Hannes Reinecke
> ---
> drivers/scsi/bfa/bfad.c | 10 +++---
> 1 file changed, 7 insertions(+), 3 deletions(
"..the first DMA mask...".
>
> Fixes: 453cd3700ca3 ("scsi: hptiop: use dma_set_mask")
> Cc:
> Suggested-by: Ewan D. Milne
> Signed-off-by: Hannes Reinecke
> ---
> drivers/scsi/hptiop.c | 10 +++---
> 1 file changed, 7 insertions(+), 3 deletions(-
r SCSI/FCP I/O errors.
>
> Fixes: f30e1bfd6154 ("scsi: lpfc: use dma_set_mask_and_coherent")
> Cc:
> Suggested-by: Don Dutile
> Signed-off-by: Ewan D. Milne
> Signed-off-by: Hannes Reinecke
> ---
> drivers/scsi/lpfc/lpfc_init.c | 9 ++---
> 1 file c
On Wed, 2019-02-13 at 12:42 +0100, Hannes Reinecke wrote:
> The recent patchset to use dma_set_mask_and_coherent() introduced
> a regression where a call to set a 64-bit DMA mask was followed
> by a call to set a 32-bit DMA mask, leading to I/O errors and
> data corruption.
>
> Patchset is based o
On Tue, 2019-02-12 at 10:34 -0800, Christoph Hellwig wrote:
> On Tue, Feb 12, 2019 at 10:37:45AM -0500, Ewan D. Milne wrote:
> >
> > It looks like this would introduce a different problem. "error" is set to
> > -ENODEV earlier in the two functions so that the vario
On Tue, 2019-02-12 at 00:06 -0800, Christoph Hellwig wrote:
> On Mon, Feb 11, 2019 at 10:05:02AM -0500, Ewan D. Milne wrote:
> > The change to use dma_set_mask_and_coherent() incorrectly made a second
> > call with the 32 bit DMA mask value when the call with the 64 bit DMA
"size 0x%x\n", phba->sli4_hba.cq_max);
> + rc = -ENOMEM;
> goto out_destroy;
> }
> lpfc_setup_cq_lookup(phba);
Reviewed-by: Ewan D. Milne
: f30e1bfd6154 ("scsi: lpfc: use dma_set_mask_and_coherent")
Cc:
Suggested-by: Don Dutile
Signed-off-by: Ewan D. Milne
---
drivers/scsi/lpfc/lpfc_init.c | 12 ++--
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_in
h & 0x400) && ql2xnvmeenable) {
> + if ((ha->fw_attributes_h &
> + (FW_ATTR_H_NVME | FW_ATTR_H_NVME_UPDATED)) &&
> + ql2xnvmeenable) {
> vha->flags.nvme_enabled = 1;
> ql_log(ql_log_info, vha, 0xd302,
> "%s: FC-NVMe is Enabled (0x%x)\n",
Reviewed-by: Ewan D. Milne
has been freed.
Ewan D. Milne (2):
lpfc: nvme: avoid hang / use-after-free when destroying localport
lpfc: nvmet: avoid hang / use-after-free when destroying targetport
drivers/scsi/lpfc/lpfc_nvme.c | 16 +---
drivers/scsi/lpfc/lpfc_nvme.h | 2 +-
drivers/scsi/lpfc/lpfc_nvmet.c
.
Fix this by putting the completion on the stack.
Signed-off-by: Ewan D. Milne
---
drivers/scsi/lpfc/lpfc_nvmet.c | 8 +---
drivers/scsi/lpfc/lpfc_nvmet.h | 2 +-
2 files changed, 6 insertions(+), 4 deletions(-)
diff --git a/drivers/scsi/lpfc/lpfc_nvmet.c b/drivers/scsi/lpfc/lpfc_nvmet.c
this by putting the completion on the stack.
Signed-off-by: Ewan D. Milne
---
drivers/scsi/lpfc/lpfc_nvme.c | 16 +---
drivers/scsi/lpfc/lpfc_nvme.h | 2 +-
2 files changed, 10 insertions(+), 8 deletions(-)
diff --git a/drivers/scsi/lpfc/lpfc_nvme.c b/drivers/scsi/lpfc/lpfc_nvme.c
.c
> @@ -6939,7 +6939,7 @@ static int qla2xxx_map_queues(struct Scsi_Host *shost)
> if (USER_CTRL_IRQ(vha->hw))
> rc = blk_mq_map_queues(qmap);
> else
> - rc = blk_mq_pci_map_queues(qmap, vha->hw->pdev, 0);
> + rc = blk_mq_pci_map_queues(qmap, vha->hw->pdev,
> vha->irq_offset);
> return rc;
> }
>
I have such a machine, SCSI-MQ doesn't work with its 8Gb adapter without this.
Reviewed-by: Ewan D. Milne
On Thu, 2019-01-03 at 16:00 -0500, Douglas Gilbert wrote:
> On 2019-01-03 4:47 a.m., Randall Huang wrote:
> > On Wed, Jan 02, 2019 at 11:51:33PM -0800, Christoph Hellwig wrote:
> > > On Wed, Dec 26, 2018 at 12:15:04PM +0800, Randall Huang wrote:
> > > > In SPC-3, WRITE(10)/(16) support grouping fun
uot;)
Cc: sta...@vger.kernel.org # 4.11+
Signed-off-by: Ewan D. Milne
---
drivers/scsi/lpfc/lpfc_sli.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/scsi/lpfc/lpfc_sli.c b/drivers/scsi/lpfc/lpfc_sli.c
index b9e5cd7..462ed4a 100644
--- a/drivers/scsi/lpfc/lpfc_sli.c
_rev == LPFC_SLI_REV4 &&
> (!(vport->load_flag & FC_UNLOADING)) &&
> (bf_get(lpfc_sli_intf_if_type,
> - &phba->sli4_hba.sli_intf) ==
> + &phba->sli4_hba.sli_intf) >=
> LPFC_SLI_INTF_IF_TYPE_2) &&
> (kref_read(&ndlp->kref) > 0)) {
> mbox->ctx_ndlp = lpfc_nlp_get(ndlp);
Reviewed-by: Ewan D. Milne
On Tue, 2018-11-27 at 17:55 +0800, chenxiang (M) wrote:
> [ 629.210506] Unable to handle kernel paging request at virtual address
> 8027e048
> [ 629.210506] Unable to handle kernel paging request at virtual address
> 8027e048
> ...
> [ 629.380218] pc : deadline_remove_request+0
On Thu, 2018-11-15 at 13:52 +0100, Hannes Reinecke wrote:
> On 11/14/18 6:13 PM, Ewan D. Milne wrote:
> > On Tue, 2018-11-13 at 09:49 -0800, Bart Van Assche wrote:
> > > On Tue, 2018-11-13 at 17:38 +, Madhani, Himanshu wrote:
> > > > On Nov 13, 2018, at 6
On Tue, 2018-11-13 at 09:49 -0800, Bart Van Assche wrote:
> On Tue, 2018-11-13 at 17:38 +, Madhani, Himanshu wrote:
> > On Nov 13, 2018, at 6:23 AM, Bart Van Assche wrote:
> > > On Tue, 2018-11-13 at 01:02 +, Madhani, Himanshu wrote:
> > > > I see other drivers also use similar information
t->ct_desc.ct_sns = dma_alloc_coherent(&vha->hw->pdev->dev,
> sizeof(struct ct_sns_pkt), &fcport->ct_desc.ct_sns_dma,
Thanks Laurence for the bisect that found this...
Reviewed-by: Ewan D. Milne
On Fri, 2018-10-19 at 12:12 +, YueHaibing wrote:
> Fixes gcc '-Wunused-but-set-variable' warning:
>
> drivers/scsi/qedf/qedf_main.c: In function 'qedf_eh_abort':
> drivers/scsi/qedf/qedf_main.c:619:21: warning:
> variable 'fcport' set but not used [-Wunused-but-set-variable]
> struct qedf_r
So in v2 you moved the #ifdef CONFIG_BLK_DEV_ZONED
so that nr_zones is always present? It was previously
changed to keep the request_queue size smaller.
Would make more sense to make the nr_zones sysfs
code here conditional on CONFIG_BLK_DEV_ZONED?
-Ewan
On Thu, 2018-10-11 at 16:09 +0900, Dami
On Tue, 2018-10-09 at 08:30 +0200, Hannes Reinecke wrote:
> lpfc_nlp_get() in lpfc_els_unsol_buffer() is not running under a
> lock, so there is a chance that it might actually fail. But as we
> never check the return value we'll get a crash in lpfc_nlp_put()
> later on trying to free an invalid bu
This change would permit a device that was OFFLINE (for any reason)
to go back to RUNNING via: OFFLINE -> BLOCKED -> RUNNING
Obviously that was your intention, but e.g. if a device was put
OFFLINE due to exceeding max_medium_access_timeouts a fabric
event and recovery would put it back online.
S
++---
> drivers/scsi/qla2xxx/qla_mbx.c| 2 +-
> drivers/scsi/qla2xxx/qla_nvme.c | 5 +--
> drivers/scsi/qla2xxx/qla_os.c | 4 +-
> drivers/scsi/qla2xxx/qla_target.c | 3 +-
> 6 files changed, 62 insertions(+), 57 deletions(-)
>
All patches in series Reviewed-by: Ewan D. Milne
define LS_MDS_LOOPBACK 0x10/* MDS Diagnostics Link Up (Loopback) */
>
> uint32_t hba_flag; /* hba generic flags */
> #define HBA_ERATT_HANDLED0x1 /* This flag is set when eratt handled */
Fixes driver crash when lpfc_enable_nvmet used.
Tested-by: Ewan D. Milne
kfree(sfp);
> return ERR_PTR(-ENODEV);
> }
> list_add_tail(&sfp->sfd_siblings, &sdp->sfds);
>
Reviewed-by: Ewan D. Milne
> dev->hba_map[bus][target].scan_counter = dev->scan_counter;
>
> aac_set_safw_target_qd(dev, bus, target);
> -
> -update_devtype:
> - dev->hba_map[bus][target].devtype = devtype;
> }
> }
>
>
The "Fixes:" tag above does not look correct to me, I've put in
what I see in Martin's tree.
Fixes a very noticeable performance regression.
Reviewed-by: Ewan D. Milne
+ ql_dbg(ql_dbg_io, sp->vha, 0x3033,
> + "%s: scope=%#x retry_delay=%#x\n", __func__,
> + sts24->retry_delay >> 14, retry_delay);
> + }
> } else {
> if (scsi_status & SS_SENSE_LEN_VALID)
> sense_len = le16_to_cpu(sts->req_sense_length);
Reviewed-by: Ewan D. Milne
gt; ha->link_data_rate)
> + fcport->fp_speed > ha->link_data_rate ||
> + !ha->flags.gpsc_supported)
> return;
>
> rval = qla2x00_set_idma_speed(vha, fcport->loop_id, fcport->fp_speed,
Martin,
This patch fixes the issue Eda found in our test environment.
Reported-by: Eda Zhou
Reviewed-by: Ewan D. Milne
Tested-by: Ewan D. Milne
On Mon, 2018-06-04 at 08:28 +, Bart Van Assche wrote:
> On Sun, 2018-06-03 at 22:09 -0700, Himanshu Madhani wrote:
> > This patch prevents driver from setting lower default speed
> > of 1 GB/sec, if the switch does not support Get Port Speed
> > Capabilities (GPSC) command. Setting this default
On Fri, 2018-05-04 at 20:37 -0700, James Smart wrote:
> A race condition is being seen. I/O is being issued to scsi targets
> simultaneous to a short connectivity loss. The rport block/unblock is
> occurring, but the unblock is allowing pending I/O commands to intermix
> with the commands being use
On Thu, 2018-02-15 at 14:00 -0500, Ewan D. Milne wrote:
> On Wed, 2017-02-22 at 08:20 -0800, James Smart wrote:
> > add missing Kconfig NVME dependencies
> >
> > Can't believe I missed posting this
> >
> > -- james
> >
> > Signed-off-by:
On Wed, 2017-02-22 at 08:20 -0800, James Smart wrote:
> add missing Kconfig NVME dependencies
>
> Can't believe I missed posting this
>
> -- james
>
> Signed-off-by: James Smart
> ---
> drivers/scsi/Kconfig | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/drivers/scsi/Kconfig b/driver
> +
> stat = qla4xxx_reset_target(ha, ddb_entry);
> if (stat != QLA_SUCCESS) {
> starget_printk(KERN_INFO, scsi_target(cmd->device),
> @@ -9376,9 +9415,16 @@ static int qla4xxx_eh_host_reset(struct scsi_cmnd *cmd)
> {
> int return_status = FAILED;
> struct scsi_qla_host *ha;
> + int rval;
>
> ha = to_qla_host(cmd->device->host);
>
> + rval = qla4xxx_isp_check_reg(ha);
> + if (rval != QLA_SUCCESS) {
> + ql4_printk(KERN_INFO, ha, "PCI/Register disconnect,
> exiting.\n");
> + return FAILED;
> + }
> +
> if ((is_qla8032(ha) || is_qla8042(ha)) && ql4xdontresethba)
> qla4_83xx_set_idc_dontreset(ha);
>
Reviewed-by: Ewan D. Milne
On Tue, 2018-01-30 at 16:07 +0100, Hannes Reinecke wrote:
> On 01/29/2018 10:08 PM, Mike Snitzer wrote:
> > We currently don't restack the queue_limits if the lowest, or
> > intermediate, layer of an IO stack changes.
> >
> > This is particularly unfortunate in the case of FLUSH/FUA which may
> >
On Tue, 2017-12-19 at 22:56 -0800, Himanshu Madhani wrote:
> @@ -4915,12 +4941,32 @@ static int qlt_24xx_handle_els(struct scsi_qla_host
> *vha,
> }
>
> if (sess != NULL) {
> + bool delete = false;
> spin_lock_irqsave(&tgt->ha
t fix your issue. It should.
commit 81b6c999897919d5a16fedc018fe375dbab091c5
Author: Hannes Reinecke
Date: Wed Dec 13 14:21:37 2017 +0100
scsi: core: check for device state in __scsi_remove_target()
As it turned out device_get() doesn't use kref_get_unless_zero(), so
On Thu, 2017-12-14 at 17:10 -0500, Ewan D. Milne wrote:
> On Thu, 2017-12-14 at 10:02 +0100, Hannes Reinecke wrote:
> > On 12/14/2017 09:05 AM, Jason Yan wrote:
> > >
> > > On 2017/12/14 6:23, Bart Van Assche wrote:
> > >> On Wed, 2017-12-13 at 14:21 +010
er to have the fix in the meantime.
>
> Cheers,
>
> Hannes
We have 2 reproducible test cases, this patch fixes one of them,
which was a continually oscillating FC target port w/short dev_loss_tmo.
I'm still waiting for a report on the iSCSI test. The code looks good.
We need to get some kind of fix for this sooner rather than later.
Reviewed-by: Ewan D. Milne
On Fri, 2017-12-01 at 11:00 -0600, Steve Wise wrote:
> Hey,
>
> I'm seeing this null pointer dereference with linux-4.15.0-rc1. To reproduce
> it, I connect two ram disks via iscsi/TCP, and start an fio:
>
> iscsiadm -m discovery --op update --type sendtargets -p 172.16.1.10:3260
> iscsiadm -m
We have another test case that demonstrates this issue involving
duplicate invocations of scsi_device_dev_release() on the same
device. This other test case involves repeated log in / log out
of an iSCSI target. (The first test case I mentioned in an earlier
mail was an oscillating FC port with a
On Wed, 2017-11-29 at 19:11 +, Bart Van Assche wrote:
> On Wed, 2017-11-29 at 13:49 -0500, Ewan D. Milne wrote:
> > because a get inside a destructor would *always* be wrong, no?
>
> Hello Ewan,
>
> That's not what we are discussing. What can happen with the SCSI
get
> *starget)
>*/
> if (sdev->channel != starget->channel ||
> sdev->id != starget->id ||
> - !get_device(&sdev->sdev_gendev))
> + scsi_device_get_not_deleted(sdev))
> continue;
> spin_unlock_irqrestore(shost->host_lock, flags);
> scsi_remove_device(sdev);
See subsequent discussion, however, we have a reproducible case here
and the patch does appear to fix the issue (500+ iterations).
Reviewed-by: Ewan D. Milne
On Wed, 2017-11-29 at 17:39 +, gre...@linuxfoundation.org wrote:
> On Wed, Nov 29, 2017 at 05:20:50PM +0100, h...@lst.de wrote:
> > On Wed, Nov 29, 2017 at 04:18:30PM +, Bart Van Assche wrote:
> > > As the above patch description shows it can happen that the SCSI core
> > > calls
> > > get
lock_irqrestore(&pring->ring_lock, iflag);
> + else
> + spin_unlock_irqrestore(&phba->hbalock, iflag);
>
> if (cmdiocbp) {
> if (cmdiocbp->iocb_cmpl) {
The other callers of lpfc_sli_iocbq_lookup() use the 2 different locks,
depending upon the SLI-3/SLI-4 case.
Reviewed-by: Ewan D. Milne
> set_bit(QPAIR_ONLINE_CHECK_NEEDED, &vha->dpc_flags);
> qla2xxx_wake_dpc(vha);
> }
> @@ -6326,7 +6335,7 @@ qla2xxx_pci_error_detected(struct pci_dev *pdev,
> pci_channel_state_t state)
> case pci_channel_io_perm_failure:
> ha->flags.pci_channel_io_perm_failure = 1;
> qla2x00_abort_all_cmds(vha, DID_NO_CONNECT << 16);
> - if (ql2xmqsupport) {
> + if (ql2xmqsupport || ql2xnvmeenable) {
> set_bit(QPAIR_ONLINE_CHECK_NEEDED, &vha->dpc_flags);
> qla2xxx_wake_dpc(vha);
> }
Reviewed-by: Ewan D. Milne
ql_dbg(ql_dbg_async, vha, 0x5075,
> +"Format 1: Remote WWPN %8phC.\n",
> +rptid_entry->u.f1.port_name);
> +
> + ql_dbg(ql_dbg_async, vha, 0x5075,
> +"Format 1: WWPN %8phC.\n",
> + vha->port_name);
> +
> + /* N2N. direct connect */
> + if (IS_QLA27XX(ha) &&
> + ((rptid_entry->u.f1.flags>>1) & 0x7) == 2) {
> + /* if our portname is higher then initiate N2N login */
> + if (wwn_to_u64(vha->port_name) >
> + wwn_to_u64(rptid_entry->u.f1.port_name)) {
> + // ??? qlt_update_host_map(vha, id);
> + vha->n2n_id = 0x1;
> + ql_dbg(ql_dbg_async, vha, 0x5075,
> + "Format 1: Setting n2n_update_needed for id
> %d\n",
> + vha->n2n_id);
> + } else {
> + ql_dbg(ql_dbg_async, vha, 0x5075,
> + "Format 1: Remote login - Waiting for WWPN
> %8phC.\n",
> + rptid_entry->u.f1.port_name);
> + }
> +
> + memcpy(vha->n2n_port_name, rptid_entry->u.f1.port_name,
> + WWN_SIZE);
> + set_bit(N2N_LOGIN_NEEDED, &vha->dpc_flags);
> + set_bit(REGISTER_FC4_NEEDED, &vha->dpc_flags);
> + set_bit(REGISTER_FDMI_NEEDED, &vha->dpc_flags);
> + return;
> + }
>
> /* buffer to buffer credit flag */
> vha->flags.bbcr_enable = (rptid_entry->u.f1.bbcr & 0xf) != 0;
> @@ -4599,6 +4633,48 @@ qla25xx_set_driver_version(scsi_qla_host_t *vha, char
> *version)
> return rval;
> }
>
> +int
> +qla24xx_get_port_login_templ(scsi_qla_host_t *vha, dma_addr_t buf_dma,
> +void *buf, uint16_t bufsiz)
> +{
> + int rval, i;
> + mbx_cmd_t mc;
> + mbx_cmd_t *mcp = &mc;
> + uint32_t*bp;
> +
> + if (!IS_FWI2_CAPABLE(vha->hw))
> + return QLA_FUNCTION_FAILED;
> +
> + ql_dbg(ql_dbg_mbx + ql_dbg_verbose, vha, 0x1159,
> + "Entered %s.\n", __func__);
> +
> + mcp->mb[0] = MBC_GET_RNID_PARAMS;
> + mcp->mb[1] = RNID_TYPE_PORT_LOGIN << 8;
> + mcp->mb[2] = MSW(buf_dma);
> + mcp->mb[3] = LSW(buf_dma);
> + mcp->mb[6] = MSW(MSD(buf_dma));
> + mcp->mb[7] = LSW(MSD(buf_dma));
> + mcp->mb[8] = bufsiz/4;
> + mcp->out_mb = MBX_8|MBX_7|MBX_6|MBX_5|MBX_4|MBX_3|MBX_2|MBX_1|MBX_0;
> + mcp->in_mb = MBX_1|MBX_0;
> + mcp->tov = MBX_TOV_SECONDS;
> + mcp->flags = 0;
> + rval = qla2x00_mailbox_command(vha, mcp);
> +
> + if (rval != QLA_SUCCESS) {
> + ql_dbg(ql_dbg_mbx, vha, 0x115a,
> + "Failed=%x mb[0]=%x,%x.\n", rval, mcp->mb[0], mcp->mb[1]);
> + } else {
> + ql_dbg(ql_dbg_mbx + ql_dbg_verbose, vha, 0x115b,
> + "Done %s.\n", __func__);
> + bp = (uint32_t *) buf;
> + for (i = 0; i < (bufsiz-4)/4; i++, bp++)
> + *bp = cpu_to_be32(*bp);
> + }
> +
> + return rval;
> +}
> +
> static int
> qla2x00_read_asic_temperature(scsi_qla_host_t *vha, uint16_t *temp)
> {
Thanks.
Tested-by: Ewan D. Milne
On Thu, 2017-10-12 at 23:08 -0700, Madhani, Madhani wrote:
> From: Himanshu Madhani
>
> Fixes following warning reported by 0-day kernel test build
>
> drivers/scsi/qla2xxx/qla_iocb.c: In function 'qla2x00_els_dcmd2_iocb_timeout':
> drivers/scsi/qla2xxx/qla_iocb.c:2611:50: warning: format '%x' e
On Mon, 2017-09-25 at 15:28 -0400, Martin K. Petersen wrote:
> Xin,
>
> > ChunYu found a kernel crash by syzkaller:
>
> [...]
>
> > It's caused by skb_shared_info at the end of sk_buff was overwritten by
> > ISCSI_KEVENT_IF_ERROR when parsing nlmsg info from skb in iscsi_if_rx.
> >
> > During th
commands with the UNMAP bit set to be limited to the
> value reported in MAXIMUM UNMAP LBA COUNT in the Block Limits VPD.
>
> Implement a blacklist option that can be used to accommodate devices
> with this behavior.
>
> Reported-by: Bill Kuzeja
> Reported-by: Ewan D. Milne
>
On Mon, 2017-09-25 at 21:46 -0400, Martin K. Petersen wrote:
> Ewan,
>
> > Some devices do not support a WRITE SAME / WRITE SAME(16) with the
> > UNMAP bit set up to the length specified in the MAXIMUM WRITE SAME
> > LENGTH field in the block limits VPD page (or, the field is zero,
> > indicating
From: "Ewan D. Milne"
Some devices do not support a WRITE SAME / WRITE SAME(16) with the UNMAP
bit set up to the length specified in the MAXIMUM WRITE SAME LENGTH field
in the block limits VPD page (or, the field is zero, indicating there is
no limit). Limit the length by the MAXIMUM
On Tue, 2017-09-19 at 09:02 -0400, Bill Kuzeja wrote:
> I encountered this issue putting XFS on several brands of SSDs on my
> system. During initialization, I would see a bunch of timeouts on
> WRITE_SAME_16 commands, which would get aborted, reissued, and complete.
> The logs look like this:
>
>
[cc's snipped to linux-scsi ]
On Thu, 2017-08-10 at 17:05 +, man...@openmail.cc wrote:
> Hello,
>
> I'd like to report this rare panic I experienced today. I've been on
> 4.12.3 since it was released and got this panic totally unexpected,
> probably when terminating my WL compositor. I at
k_get_queue(SDp->request_queue))
> goto out_put_disk;
> + disk->queue = SDp->request_queue;
> tpnt->driver = &st_template;
>
> tpnt->device = SDp;
Fixes: 2b5bebccd282 ("st: Take additional queue ref in st_probe")
Reviewed-by: Ewan D. Milne
On Fri, 2017-07-14 at 10:19 -0400, Mike Snitzer wrote:
>
> Do you see a benefit to extracting that portion of your WIP patch
> (removing the ->complete handler entirely)?
>
> Or leave well enough alone and just continue to disable dm-mq's ability
> to stack on .request_fn paths?
>
> Given SCSI's
[ removed cc: stable from discussion ]
On Wed, 2017-06-28 at 09:38 +0200, Johannes Thumshirn wrote:
>
> Looks good,
> Reviewed-by: Johannes Thumshirn
>
> Although we've been tampering with the target removal code for quite some
> time now, so I really have the gut feeling we haven't really fixe
On Tue, 2017-06-27 at 14:02 +0300, Andrei Borzenkov wrote:
> I'm facing storage system that while generally advertising support for
> WRITE_SAME_16 with UNMAP in reality fails this request depending on
> exact volume configuration. This configuration is done on storage side
> and may happen after h
From: "Ewan D. Milne"
The addition of the STARGET_REMOVE state had the side effect of
introducing a race condition that can cause a crash.
scsi_target_reap_ref_release() checks the starget->state to
see if it still in STARGET_CREATED, and if so, skips calling
transport_remove
On Thu, 2017-05-18 at 11:14 -0700, Song Liu wrote:
> This patch adds rate limits to SCSI sense code uevets. Currently,
> the rate limit is hard-coded to 16 events per second.
>
> The code tracks nano second time of latest 16 events in a circular
> buffer. When a new event arrives, the time is comp
1 - 100 of 273 matches
Mail list logo
