[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH] hw/cxl/device: Add Flex Bus Port DVSEC
From: |
Jonathan Cameron |
Subject: |
Re: [PATCH] hw/cxl/device: Add Flex Bus Port DVSEC |
Date: |
Wed, 14 Dec 2022 14:01:35 +0000 |
On Wed, 14 Dec 2022 12:39:45 +0000
Jonathan Cameron via <qemu-devel@nongnu.org> wrote:
> On Tue, 13 Dec 2022 16:34:52 -0800
> Ira Weiny <ira.weiny@intel.com> wrote:
>
> > The Flex Bus Port DVSEC was missing on type 3 devices which was blocking
> > RAS checks.[1]
> >
> > Add the Flex Bus Port DVSEC to type 3 devices as per CXL 3.0 8.2.1.3.
> >
> > [1]
> > https://lore.kernel.org/linux-cxl/167096738875.2861540.11815053323626849940.stgit@djiang5-desk3.ch.intel.com/
> >
> > Cc: Dave Jiang <dave.jiang@intel.com>
> > Cc: Jonathan Cameron <Jonathan.Cameron@huawei.com>
> > Cc: Ben Widawsky <bwidawsk@kernel.org>
> > Cc: qemu-devel@nongnu.org
> > Cc: linux-cxl@vger.kernel.org
> > Signed-off-by: Ira Weiny <ira.weiny@intel.com>
>
> Not sure how we missed this one given the whole fun a while back with Mem
> Enable not
> being set by the kernel code. Ah well - I'm not going to go digging in the
> history
> for that.
Ah. I forgot the difference between the 'control' in here which isn't really a
control
in this case and the one in PCIe DVSEC for CXL devices which is the one were we
enable it.
I guess this calls for a more thorough check on what else we are missing
in the way of DVSEC - including this stuff which is all about link training but
does
provide some useful info the stuff that results from the training.
>
> Minor thing inline - unlike root ports, switches etc, type 3 device should not
> report CXL.cache support.
>
> Jonathan
>
>
> > ---
> > hw/mem/cxl_type3.c | 11 +++++++++++
> > 1 file changed, 11 insertions(+)
> >
> > diff --git a/hw/mem/cxl_type3.c b/hw/mem/cxl_type3.c
> > index 0317bd96a6fb..27f6ac0cb3c1 100644
> > --- a/hw/mem/cxl_type3.c
> > +++ b/hw/mem/cxl_type3.c
> > @@ -416,6 +416,17 @@ static void build_dvsecs(CXLType3Dev *ct3d)
> > cxl_component_create_dvsec(cxl_cstate, CXL2_TYPE3_DEVICE,
> > GPF_DEVICE_DVSEC_LENGTH, GPF_DEVICE_DVSEC,
> > GPF_DEVICE_DVSEC_REVID, dvsec);
> > +
> > + dvsec = (uint8_t *)&(CXLDVSECPortFlexBus){
> > + .cap = 0x27, /* Cache, IO, Mem, non-MLD */
>
> Type 3 typically wouldn't support cache. Probably want the comment to include
> bit 5 (68B Flit and VH capable) That should probably true in the other
> instances
> of this as well.
>
>
> > + .ctrl = 0x02, /* IO always enabled */
> > + .status = 0x27, /* same as capabilities */
> Again, not cache.
>
> > + .rcvd_mod_ts_data_phase1 = 0xef, /* WTF? */
> > + };
> > + cxl_component_create_dvsec(cxl_cstate, CXL2_TYPE3_DEVICE,
> > + PCIE_FLEXBUS_PORT_DVSEC_LENGTH_2_0,
> > + PCIE_FLEXBUS_PORT_DVSEC,
> > + PCIE_FLEXBUS_PORT_DVSEC_REVID_2_0, dvsec);
> > }
> >
> > static void hdm_decoder_commit(CXLType3Dev *ct3d, int which)
> >
> > ---
> > base-commit: e11b57108b0cb746bb9f3887054f34a2f818ed79
> > change-id: 20221213-ira-flexbus-port-ce526de8111d
> >
> > Best regards,
>
>