From: Arun Easi <aeasi@marvell.com>
To: "Martin K. Petersen" <martin.petersen@oracle.com>
Cc: <linux-scsi@vger.kernel.org>
Subject: Re: [PATCH 03/15] scsi: qla2xxx: Use the proper SCSI midlayer interfaces for PI
Date: Wed, 16 Jun 2021 17:15:39 -0700 [thread overview]
Message-ID: <alpine.LRH.2.21.9999.2106161711000.17918@irv1user01.caveonetworks.com> (raw)
In-Reply-To: <20210609033929.3815-4-martin.petersen@oracle.com>
On Tue, 8 Jun 2021, 8:39pm, Martin K. Petersen wrote:
> Use the SCSI midlayer interfaces to query protection interval,
> reference tag, and per-command DIX flags.
>
> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
> ---
> drivers/scsi/qla2xxx/qla_iocb.c | 84 ++++++---------------------------
> 1 file changed, 15 insertions(+), 69 deletions(-)
>
> diff --git a/drivers/scsi/qla2xxx/qla_iocb.c b/drivers/scsi/qla2xxx/qla_iocb.c
> index 38b5bdde2405..42a6fbba7529 100644
> --- a/drivers/scsi/qla2xxx/qla_iocb.c
> +++ b/drivers/scsi/qla2xxx/qla_iocb.c
> @@ -145,7 +145,6 @@ inline int
> qla24xx_configure_prot_mode(srb_t *sp, uint16_t *fw_prot_opts)
> {
> struct scsi_cmnd *cmd = GET_CMD_SP(sp);
> - uint8_t guard = scsi_host_get_guard(cmd->device->host);
>
> /* We always use DIFF Bundling for best performance */
> *fw_prot_opts = 0;
> @@ -166,7 +165,7 @@ qla24xx_configure_prot_mode(srb_t *sp, uint16_t *fw_prot_opts)
> break;
> case SCSI_PROT_READ_PASS:
> case SCSI_PROT_WRITE_PASS:
> - if (guard & SHOST_DIX_GUARD_IP)
> + if (cmd->prot_flags & SCSI_PROT_IP_CHECKSUM)
> *fw_prot_opts |= PO_MODE_DIF_TCP_CKSUM;
> else
> *fw_prot_opts |= PO_MODE_DIF_PASS;
> @@ -176,6 +175,9 @@ qla24xx_configure_prot_mode(srb_t *sp, uint16_t *fw_prot_opts)
> break;
> }
>
> + if (!(cmd->prot_flags & SCSI_PROT_GUARD_CHECK))
> + *fw_prot_opts |= PO_DISABLE_GUARD_CHECK;
> +
> return scsi_prot_sg_count(cmd);
> }
>
> @@ -772,74 +774,18 @@ qla24xx_set_t10dif_tags(srb_t *sp, struct fw_dif_context *pkt,
> {
> struct scsi_cmnd *cmd = GET_CMD_SP(sp);
>
> - switch (scsi_get_prot_type(cmd)) {
> - case SCSI_PROT_DIF_TYPE0:
> - /*
> - * No check for ql2xenablehba_err_chk, as it would be an
> - * I/O error if hba tag generation is not done.
> - */
> - pkt->ref_tag = cpu_to_le32((uint32_t)
> - (0xffffffff & scsi_get_lba(cmd)));
> -
> - if (!qla2x00_hba_err_chk_enabled(sp))
> - break;
> -
> - pkt->ref_tag_mask[0] = 0xff;
> - pkt->ref_tag_mask[1] = 0xff;
> - pkt->ref_tag_mask[2] = 0xff;
> - pkt->ref_tag_mask[3] = 0xff;
> - break;
> -
> - /*
> - * For TYPE 2 protection: 16 bit GUARD + 32 bit REF tag has to
> - * match LBA in CDB + N
> - */
> - case SCSI_PROT_DIF_TYPE2:
> - pkt->app_tag = cpu_to_le16(0);
> - pkt->app_tag_mask[0] = 0x0;
> - pkt->app_tag_mask[1] = 0x0;
> -
> - pkt->ref_tag = cpu_to_le32((uint32_t)
> - (0xffffffff & scsi_get_lba(cmd)));
> + pkt->ref_tag = cpu_to_le32(scsi_prot_ref_tag(cmd));
>
> - if (!qla2x00_hba_err_chk_enabled(sp))
> - break;
> -
> - /* enable ALL bytes of the ref tag */
> - pkt->ref_tag_mask[0] = 0xff;
> - pkt->ref_tag_mask[1] = 0xff;
> - pkt->ref_tag_mask[2] = 0xff;
> - pkt->ref_tag_mask[3] = 0xff;
> - break;
> -
> - /* For Type 3 protection: 16 bit GUARD only */
> - case SCSI_PROT_DIF_TYPE3:
> - pkt->ref_tag_mask[0] = pkt->ref_tag_mask[1] =
> - pkt->ref_tag_mask[2] = pkt->ref_tag_mask[3] =
> - 0x00;
> - break;
> -
> - /*
> - * For TYpe 1 protection: 16 bit GUARD tag, 32 bit REF tag, and
> - * 16 bit app tag.
> - */
> - case SCSI_PROT_DIF_TYPE1:
> - pkt->ref_tag = cpu_to_le32((uint32_t)
> - (0xffffffff & scsi_get_lba(cmd)));
> - pkt->app_tag = cpu_to_le16(0);
> - pkt->app_tag_mask[0] = 0x0;
> - pkt->app_tag_mask[1] = 0x0;
> -
> - if (!qla2x00_hba_err_chk_enabled(sp))
> - break;
> -
> - /* enable ALL bytes of the ref tag */
> - pkt->ref_tag_mask[0] = 0xff;
> - pkt->ref_tag_mask[1] = 0xff;
> - pkt->ref_tag_mask[2] = 0xff;
> - pkt->ref_tag_mask[3] = 0xff;
> - break;
> + if (cmd->prot_flags & SCSI_PROT_REF_CHECK) {
I would also add a "&& qla2x00_hba_err_chk_enabled(sp)" to the above check
to preserve the old semantics (which was helpful during testing).
With that:
Reviewed-by: Arun Easi <aeasi@marvell.com>
Regards,
-Arun
> + pkt->ref_tag_mask[0] = 0xff;
> + pkt->ref_tag_mask[1] = 0xff;
> + pkt->ref_tag_mask[2] = 0xff;
> + pkt->ref_tag_mask[3] = 0xff;
> }
> +
> + pkt->app_tag = __constant_cpu_to_le16(0);
> + pkt->app_tag_mask[0] = 0x0;
> + pkt->app_tag_mask[1] = 0x0;
> }
>
> int
> @@ -905,7 +851,7 @@ qla24xx_walk_and_build_sglist_no_difb(struct qla_hw_data *ha, srb_t *sp,
> memset(&sgx, 0, sizeof(struct qla2_sgx));
> if (sp) {
> cmd = GET_CMD_SP(sp);
> - prot_int = cmd->device->sector_size;
> + prot_int = scsi_prot_interval(cmd);
>
> sgx.tot_bytes = scsi_bufflen(cmd);
> sgx.cur_sg = scsi_sglist(cmd);
>
next prev parent reply other threads:[~2021-06-17 0:15 UTC|newest]
Thread overview: 39+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-06-09 3:39 [PATCH 00/15] Subject: Protection information and block size cleanup Martin K. Petersen
2021-06-09 3:39 ` [PATCH 01/15] scsi: core: Add scsi_prot_ref_tag() helper Martin K. Petersen
2021-06-09 19:55 ` Bart Van Assche
2021-06-14 14:33 ` Benjamin Block
2021-06-09 3:39 ` [PATCH 02/15] scsi: lpfc: Use the proper SCSI midlayer interfaces for PI Martin K. Petersen
2021-06-09 3:39 ` [PATCH 03/15] scsi: qla2xxx: " Martin K. Petersen
2021-06-17 0:15 ` Arun Easi [this message]
2021-06-09 3:39 ` [PATCH 04/15] scsi: mpt3sas: " Martin K. Petersen
2021-06-09 3:39 ` [PATCH 05/15] scsi: mpi3mr: " Martin K. Petersen
2021-06-09 3:39 ` [PATCH 06/15] scsi: zfcp: " Martin K. Petersen
2021-06-14 14:33 ` Benjamin Block
2021-06-15 2:27 ` Martin K. Petersen
2021-06-15 9:31 ` Benjamin Block
2021-06-09 3:39 ` [PATCH 07/15] scsi: isci: " Martin K. Petersen
2021-06-09 3:39 ` [PATCH 08/15] scsi: scsi_debug: Remove dump_sector() Martin K. Petersen
2021-06-09 19:55 ` Bart Van Assche
2021-06-14 22:06 ` Douglas Gilbert
2021-06-09 3:39 ` [PATCH 09/15] scsi: scsi_debug: Improve RDPROTECT/WRPROTECT handling Martin K. Petersen
2021-06-14 22:07 ` Douglas Gilbert
2021-06-09 3:39 ` [PATCH 10/15] scsi: core: Introduce scsi_get_sector() Martin K. Petersen
2021-06-14 14:48 ` Benjamin Block
2021-06-15 2:28 ` Martin K. Petersen
2021-06-15 9:33 ` Benjamin Block
2021-06-09 3:39 ` [PATCH 11/15] scsi: iser: Use scsi_get_sector() instead of scsi_get_lba() Martin K. Petersen
2021-06-09 3:39 ` [PATCH 12/15] scsi: core: Make scsi_get_lba() return the LBA Martin K. Petersen
2021-06-09 19:57 ` Bart Van Assche
2021-06-14 14:50 ` Benjamin Block
2021-06-15 9:34 ` Benjamin Block
2021-06-09 3:39 ` [PATCH 13/15] scsi: core: Add helper to return number of logical blocks in a request Martin K. Petersen
2021-06-09 19:58 ` Bart Van Assche
2021-06-10 3:19 ` Martin K. Petersen
2021-06-10 3:40 ` Bart Van Assche
2021-06-09 3:39 ` [PATCH 14/15] scsi: lpfc: Switch to scsi_get_block_count() Martin K. Petersen
2021-06-09 3:39 ` [PATCH 15/15] scsi: ufs: core: Use scsi_get_lba() to get LBA Martin K. Petersen
2021-06-09 19:59 ` Bart Van Assche
2021-07-16 21:27 ` [PATCH 00/15] Subject: Protection information and block size cleanup Bart Van Assche
2021-07-19 1:43 ` Martin K. Petersen
2021-07-21 3:53 ` Bart Van Assche
2021-07-24 2:13 ` Martin K. Petersen
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=alpine.LRH.2.21.9999.2106161711000.17918@irv1user01.caveonetworks.com \
--to=aeasi@marvell.com \
--cc=linux-scsi@vger.kernel.org \
--cc=martin.petersen@oracle.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).