Re: [PATCH 2/2] interconnect: qcom: eliza: Add SDCC1 slave node

From: Mike Tipton

Date: Thu May 14 2026 - 09:10:13 EST


On Thu, May 14, 2026 at 01:39:24PM +0300, Abel Vesa wrote:
> On 26-05-14 12:52:21, Georgi Djakov wrote:
> > On 5/12/26 3:18 PM, Abel Vesa wrote:
> > > The Eliza interconnect provider is missing the SDCC1 CNOC CFG slave
> > > node. Add qhs_sdc1 to the provider node table and qsm_cfg links so
> > > SDCC1 interconnect paths can resolve to a provider node.
> > >
> > > Bump the qsm_cfg link count to match the new entry.
> > >
> > > Fixes: a39efc80ff50 ("interconnect: qcom: Add Eliza interconnect provider driver")
> > > Signed-off-by: Abel Vesa <abel.vesa@xxxxxxxxxxxxxxxx>
> > > ---
> > > drivers/interconnect/qcom/eliza.c | 11 +++++++++--
> > > 1 file changed, 9 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/drivers/interconnect/qcom/eliza.c b/drivers/interconnect/qcom/eliza.c
> > > index a4f7903f0524..6d12a1c3c034 100644
> > > --- a/drivers/interconnect/qcom/eliza.c
> > > +++ b/drivers/interconnect/qcom/eliza.c
> > > @@ -127,6 +127,12 @@ static struct qcom_icc_node qhs_qup2 = {
> > > .buswidth = 4,
> > > };
> > > +static struct qcom_icc_node qhs_sdc1 = {
> > > + .name = "qhs_sdc1",
> > > + .channels = 1,
> > > + .buswidth = 4,
> > > +};
> > > +
> > > static struct qcom_icc_node qhs_sdc2 = {
> > > .name = "qhs_sdc2",
> > > .channels = 1,
> > > @@ -383,7 +389,7 @@ static struct qcom_icc_node qsm_cfg = {
> > > .name = "qsm_cfg",
> > > .channels = 1,
> > > .buswidth = 4,
> > > - .num_links = 29,
> > > + .num_links = 30,
> > > .link_nodes = { &qhs_ahb2phy0, &qhs_ahb2phy1,
> > > &qhs_camera_cfg, &qhs_clk_ctl,
> > > &qhs_crypto0_cfg, &qhs_display_cfg,
> > > @@ -392,7 +398,7 @@ static struct qcom_icc_node qsm_cfg = {
> > > &qhs_mss_cfg, &qhs_pcie_0_cfg,
> > > &qhs_prng, &qhs_qdss_cfg,
> > > &qhs_qspi, &qhs_qup1,
> > > - &qhs_qup2, &qhs_sdc2,
> > > + &qhs_qup2, &qhs_sdc1, &qhs_sdc2,
> > > &qhs_tcsr, &qhs_tlmm,
> > > &qhs_ufs_mem_cfg, &qhs_usb3_0,
> > > &qhs_venus_cfg, &qhs_vsense_ctrl_cfg,
> >
> > Should we also add the sdc1 to some BCM?
>
> AFAICT, it should not. Even downstream doesn't add it.

It should be on CN0. Downstream not adding it to CN0 was also an
oversight. It shouldn't make much difference in this case since CN0 is
an on-off BCM and other BCMs along the path to qhs_sdc1 would indirectly
keep CNOC alive, but it should technically be on CN0 just like qhs_sdc2.

Thanks,
Mike