| Message ID | 20191029230257.210897-1-bvanassche@acm.org (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | RDMA/srpt: Report the SCSI residual to the initiator | expand |
On Tue, Oct 29, 2019 at 04:02:57PM -0700, Bart Van Assche wrote: > The code added by this patch is similar to the code that already exists > in ibmvscsis_determine_resid(). This patch has been tested by running > the following command: > > strace -f sg_raw -r 1k /dev/sdb 12 00 00 00 60 00 -o inquiry.bin |& > grep resid= > > Cc: Honggang LI <honli@redhat.com> > Cc: Laurence Oberman <loberman@redhat.com> > Fixes: a42d985bd5b2 ("ib_srpt: Initial SRP Target merge for v3.3-rc1") > Signed-off-by: Bart Van Assche <bvanassche@acm.org> > --- > drivers/infiniband/ulp/srpt/ib_srpt.c | 24 ++++++++++++++++++++++++ > 1 file changed, 24 insertions(+) > > diff --git a/drivers/infiniband/ulp/srpt/ib_srpt.c b/drivers/infiniband/ulp/srpt/ib_srpt.c > index a278e76b9e02..c9972b686c27 100644 > --- a/drivers/infiniband/ulp/srpt/ib_srpt.c > +++ b/drivers/infiniband/ulp/srpt/ib_srpt.c > @@ -1362,9 +1362,11 @@ static int srpt_build_cmd_rsp(struct srpt_rdma_ch *ch, > struct srpt_send_ioctx *ioctx, u64 tag, > int status) > { > + struct se_cmd *cmd = &ioctx->cmd; > struct srp_rsp *srp_rsp; > const u8 *sense_data; > int sense_data_len, max_sense_len; > + int resid = cmd->residual_count; ^^^ Should be u32?
On 11/4/19 3:10 AM, Honggang LI wrote: > On Tue, Oct 29, 2019 at 04:02:57PM -0700, Bart Van Assche wrote: >> The code added by this patch is similar to the code that already exists >> in ibmvscsis_determine_resid(). This patch has been tested by running >> the following command: >> >> strace -f sg_raw -r 1k /dev/sdb 12 00 00 00 60 00 -o inquiry.bin |& >> grep resid= >> >> Cc: Honggang LI <honli@redhat.com> >> Cc: Laurence Oberman <loberman@redhat.com> >> Fixes: a42d985bd5b2 ("ib_srpt: Initial SRP Target merge for v3.3-rc1") >> Signed-off-by: Bart Van Assche <bvanassche@acm.org> >> --- >> drivers/infiniband/ulp/srpt/ib_srpt.c | 24 ++++++++++++++++++++++++ >> 1 file changed, 24 insertions(+) >> >> diff --git a/drivers/infiniband/ulp/srpt/ib_srpt.c b/drivers/infiniband/ulp/srpt/ib_srpt.c >> index a278e76b9e02..c9972b686c27 100644 >> --- a/drivers/infiniband/ulp/srpt/ib_srpt.c >> +++ b/drivers/infiniband/ulp/srpt/ib_srpt.c >> @@ -1362,9 +1362,11 @@ static int srpt_build_cmd_rsp(struct srpt_rdma_ch *ch, >> struct srpt_send_ioctx *ioctx, u64 tag, >> int status) >> { >> + struct se_cmd *cmd = &ioctx->cmd; >> struct srp_rsp *srp_rsp; >> const u8 *sense_data; >> int sense_data_len, max_sense_len; >> + int resid = cmd->residual_count; > ^^^ > Should be u32? Since se_cmd.residual_count has been declared as u32 it's probably better to declare this variable as u32 too. I will make that change and post a v2. Thanks, Bart.
diff --git a/drivers/infiniband/ulp/srpt/ib_srpt.c b/drivers/infiniband/ulp/srpt/ib_srpt.c index a278e76b9e02..c9972b686c27 100644 --- a/drivers/infiniband/ulp/srpt/ib_srpt.c +++ b/drivers/infiniband/ulp/srpt/ib_srpt.c @@ -1362,9 +1362,11 @@ static int srpt_build_cmd_rsp(struct srpt_rdma_ch *ch, struct srpt_send_ioctx *ioctx, u64 tag, int status) { + struct se_cmd *cmd = &ioctx->cmd; struct srp_rsp *srp_rsp; const u8 *sense_data; int sense_data_len, max_sense_len; + int resid = cmd->residual_count; /* * The lowest bit of all SAM-3 status codes is zero (see also @@ -1386,6 +1388,28 @@ static int srpt_build_cmd_rsp(struct srpt_rdma_ch *ch, srp_rsp->tag = tag; srp_rsp->status = status; + if (cmd->se_cmd_flags & SCF_UNDERFLOW_BIT) { + if (cmd->data_direction == DMA_TO_DEVICE) { + /* residual data from an underflow write */ + srp_rsp->flags = SRP_RSP_FLAG_DOUNDER; + srp_rsp->data_out_res_cnt = cpu_to_be32(resid); + } else if (cmd->data_direction == DMA_FROM_DEVICE) { + /* residual data from an underflow read */ + srp_rsp->flags = SRP_RSP_FLAG_DIUNDER; + srp_rsp->data_in_res_cnt = cpu_to_be32(resid); + } + } else if (cmd->se_cmd_flags & SCF_OVERFLOW_BIT) { + if (cmd->data_direction == DMA_TO_DEVICE) { + /* residual data from an overflow write */ + srp_rsp->flags = SRP_RSP_FLAG_DOOVER; + srp_rsp->data_out_res_cnt = cpu_to_be32(resid); + } else if (cmd->data_direction == DMA_FROM_DEVICE) { + /* residual data from an overflow read */ + srp_rsp->flags = SRP_RSP_FLAG_DIOVER; + srp_rsp->data_in_res_cnt = cpu_to_be32(resid); + } + } + if (sense_data_len) { BUILD_BUG_ON(MIN_MAX_RSP_SIZE <= sizeof(*srp_rsp)); max_sense_len = ch->max_ti_iu_len - sizeof(*srp_rsp);
The code added by this patch is similar to the code that already exists in ibmvscsis_determine_resid(). This patch has been tested by running the following command: strace -f sg_raw -r 1k /dev/sdb 12 00 00 00 60 00 -o inquiry.bin |& grep resid= Cc: Honggang LI <honli@redhat.com> Cc: Laurence Oberman <loberman@redhat.com> Fixes: a42d985bd5b2 ("ib_srpt: Initial SRP Target merge for v3.3-rc1") Signed-off-by: Bart Van Assche <bvanassche@acm.org> --- drivers/infiniband/ulp/srpt/ib_srpt.c | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+)