From patchwork Sun Mar 6 17:41:09 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: Jonathan Cameron X-Patchwork-Id: 12770790 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id C1495C433EF for ; Sun, 6 Mar 2022 17:50:43 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S231596AbiCFRvf (ORCPT ); Sun, 6 Mar 2022 12:51:35 -0500 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:40416 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S230335AbiCFRve (ORCPT ); Sun, 6 Mar 2022 12:51:34 -0500 Received: from frasgout.his.huawei.com (frasgout.his.huawei.com [185.176.79.56]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 9F6D03E5DB for ; Sun, 6 Mar 2022 09:50:41 -0800 (PST) Received: from fraeml738-chm.china.huawei.com (unknown [172.18.147.201]) by frasgout.his.huawei.com (SkyGuard) with ESMTP id 4KBTcJ5JVLz67N6g; Mon, 7 Mar 2022 01:49:20 +0800 (CST) Received: from lhreml710-chm.china.huawei.com (10.201.108.61) by fraeml738-chm.china.huawei.com (10.206.15.219) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.1.2308.21; Sun, 6 Mar 2022 18:50:39 +0100 Received: from SecurePC-101-06.china.huawei.com (10.122.247.231) by lhreml710-chm.china.huawei.com (10.201.108.61) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256) id 15.1.2308.21; Sun, 6 Mar 2022 17:50:38 +0000 From: Jonathan Cameron To: , , =?utf-8?q?Alex_Benn?= =?utf-8?q?=C3=A9e?= , Marcel Apfelbaum , "Michael S . Tsirkin" , Igor Mammedov , Markus Armbruster CC: , Ben Widawsky , "Peter Maydell" , Shameerali Kolothum Thodi , =?utf-8?q?Philippe_Mathieu-Daud=C3=A9?= , Saransh Gupta1 , Shreyas Shah , Chris Browy , Samarth Saxena , "Dan Williams" Subject: [PATCH v7 18/46] hw/cxl/device: Implement MMIO HDM decoding (8.2.5.12) Date: Sun, 6 Mar 2022 17:41:09 +0000 Message-ID: <20220306174137.5707-19-Jonathan.Cameron@huawei.com> X-Mailer: git-send-email 2.32.0 In-Reply-To: <20220306174137.5707-1-Jonathan.Cameron@huawei.com> References: <20220306174137.5707-1-Jonathan.Cameron@huawei.com> MIME-Version: 1.0 X-Originating-IP: [10.122.247.231] X-ClientProxiedBy: lhreml706-chm.china.huawei.com (10.201.108.55) To lhreml710-chm.china.huawei.com (10.201.108.61) X-CFilter-Loop: Reflected Precedence: bulk List-ID: X-Mailing-List: linux-cxl@vger.kernel.org From: Ben Widawsky A device's volatile and persistent memory are known Host Defined Memory (HDM) regions. The mechanism by which the device is programmed to claim the addresses associated with those regions is through dedicated logic known as the HDM decoder. In order to allow the OS to properly program the HDMs, the HDM decoders must be modeled. There are two ways the HDM decoders can be implemented, the legacy mechanism is through the PCIe DVSEC programming from CXL 1.1 (8.1.3.8), and MMIO is found in 8.2.5.12 of the spec. For now, 8.1.3.8 is not implemented. Much of CXL device logic is implemented in cxl-utils. The HDM decoder however is implemented directly by the device implementation. Whilst the implementation currently does no validity checks on the encoder set up, future work will add sanity checking specific to the type of cxl component. Signed-off-by: Ben Widawsky Co-developed-by: Jonathan Cameron Signed-off-by: Jonathan Cameron Reviewed-by: Alex Bennée --- v7: * Drop pointless void * cast. (Alex) * Add assertion but without the suggested >> 2 as unit is bytes here (Alex) hw/mem/cxl_type3.c | 55 ++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 55 insertions(+) diff --git a/hw/mem/cxl_type3.c b/hw/mem/cxl_type3.c index a8d7cfcc81..16b113d5ed 100644 --- a/hw/mem/cxl_type3.c +++ b/hw/mem/cxl_type3.c @@ -46,6 +46,57 @@ static void build_dvsecs(CXLType3Dev *ct3d) REG_LOC_DVSEC_REVID, dvsec); } +static void hdm_decoder_commit(CXLType3Dev *ct3d, int which) +{ + ComponentRegisters *cregs = &ct3d->cxl_cstate.crb; + uint32_t *cache_mem = cregs->cache_mem_registers; + + assert(which == 0); + + /* TODO: Sanity checks that the decoder is possible */ + ARRAY_FIELD_DP32(cache_mem, CXL_HDM_DECODER0_CTRL, COMMIT, 0); + ARRAY_FIELD_DP32(cache_mem, CXL_HDM_DECODER0_CTRL, ERR, 0); + + ARRAY_FIELD_DP32(cache_mem, CXL_HDM_DECODER0_CTRL, COMMITTED, 1); +} + +static void ct3d_reg_write(void *opaque, hwaddr offset, uint64_t value, + unsigned size) +{ + CXLComponentState *cxl_cstate = opaque; + ComponentRegisters *cregs = &cxl_cstate->crb; + CXLType3Dev *ct3d = container_of(cxl_cstate, CXLType3Dev, cxl_cstate); + uint32_t *cache_mem = cregs->cache_mem_registers; + bool should_commit = false; + int which_hdm = -1; + + assert(size == 4); + g_assert(offset <= CXL2_COMPONENT_CM_REGION_SIZE); + + switch (offset) { + case A_CXL_HDM_DECODER0_CTRL: + should_commit = FIELD_EX32(value, CXL_HDM_DECODER0_CTRL, COMMIT); + which_hdm = 0; + break; + default: + break; + } + + stl_le_p((uint8_t *)cache_mem + offset, value); + if (should_commit) { + hdm_decoder_commit(ct3d, which_hdm); + } +} + +static void ct3_finalize(Object *obj) +{ + CXLType3Dev *ct3d = CT3(obj); + CXLComponentState *cxl_cstate = &ct3d->cxl_cstate; + ComponentRegisters *regs = &cxl_cstate->crb; + + g_free(regs->special_ops); +} + static void cxl_setup_memory(CXLType3Dev *ct3d, Error **errp) { MemoryRegion *mr; @@ -86,6 +137,9 @@ static void ct3_realize(PCIDevice *pci_dev, Error **errp) ct3d->cxl_cstate.pdev = pci_dev; build_dvsecs(ct3d); + regs->special_ops = g_new0(MemoryRegionOps, 1); + regs->special_ops->write = ct3d_reg_write; + cxl_component_register_block_init(OBJECT(pci_dev), cxl_cstate, TYPE_CXL_TYPE3_DEV); @@ -138,6 +192,7 @@ static const TypeInfo ct3d_info = { .parent = TYPE_PCI_DEVICE, .class_init = ct3_class_init, .instance_size = sizeof(CXLType3Dev), + .instance_finalize = ct3_finalize, .interfaces = (InterfaceInfo[]) { { INTERFACE_CXL_DEVICE }, { INTERFACE_PCIE_DEVICE },