From patchwork Sat Sep 10 19:50:07 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: Serge Semin X-Patchwork-Id: 12972693 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 bombadil.infradead.org (bombadil.infradead.org [198.137.202.133]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id 5353AECAAD3 for ; Sat, 10 Sep 2022 20:12:12 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=bombadil.20210309; h=Sender: Content-Transfer-Encoding:Content-Type:List-Subscribe:List-Help:List-Post: List-Archive:List-Unsubscribe:List-Id:MIME-Version:References:In-Reply-To: Message-ID:Date:Subject:CC:To:From:Reply-To:Content-ID:Content-Description: Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID: List-Owner; bh=4hqWexscK1Xc0b5AnXH0ZaaHf9d7VR6ZNF5XYd3UMWs=; b=m3RNCF7686WYsM /PQzYWhaJOHxsVvLz8QqJ8NApZX6RCSsPFszqbsLF6+Zftr7H7+41z3MhntI1kMsauHHtDEvpidew Gp2KVBTy0LfxflR9gAWhijfszp5vBKzFB4ZKYJ9svUtojdDeafYk/9cxJEMK4A24+vzO2obIv1Wf9 gjT/yUMOWM1HOqNek2iZ/4xXSSTTiVOfv8Uz0CIJqqj0N0o5fk3cqk9CxH7e11W8RdDTrGf3S4Cq4 +AaXE9LvSxK0ILWn5pnsuaADh62hSCj0CcB5pgetJUk6Bjf3WqgEOY0RaQgp4R+pfGs0IK4v847ZL PMPS7+qWCrPZG+iwtNPw==; Received: from localhost ([::1] helo=bombadil.infradead.org) by bombadil.infradead.org with esmtp (Exim 4.94.2 #2 (Red Hat Linux)) id 1oX6oi-00CvZ8-4m; Sat, 10 Sep 2022 20:11:04 +0000 Received: from mail.baikalelectronics.com ([87.245.175.230]) by bombadil.infradead.org with esmtp (Exim 4.94.2 #2 (Red Hat Linux)) id 1oX6VA-00Cjbr-8X for linux-arm-kernel@lists.infradead.org; Sat, 10 Sep 2022 19:50:54 +0000 Received: from mail (mail.baikal.int [192.168.51.25]) by mail.baikalelectronics.com (Postfix) with ESMTP id 4812DDB9; Sat, 10 Sep 2022 22:54:21 +0300 (MSK) DKIM-Filter: OpenDKIM Filter v2.11.0 mail.baikalelectronics.com 4812DDB9 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=baikalelectronics.ru; s=mail; t=1662839661; bh=HPHDfyJrEYpaEFhSpc1k6K86wG14xYplLp+NbPkZqWo=; h=From:To:CC:Subject:Date:In-Reply-To:References:From; b=pUIXAd5vciQNU61b15y4AebY0eGmCtt/hUN5McYf4fLf2AYXyn3woR7eFcLpwcX5X adQnm5E+CkFBb1j7PZjaKShx3W7zn7hN6Q8Kz8Hm+gS1OpOrBzmUPBPW84vjTTRq9b XQ0st63bn4VIvDkizEjCh9T8S+bDXf/rIMte/BUs= Received: from localhost (192.168.168.10) by mail (192.168.51.25) with Microsoft SMTP Server (TLS) id 15.0.1395.4; Sat, 10 Sep 2022 22:50:30 +0300 From: Serge Semin To: Michal Simek , Borislav Petkov , Mauro Carvalho Chehab , Tony Luck , James Morse , Robert Richter CC: Serge Semin , Serge Semin , Alexey Malahov , Michail Ivanov , Pavel Parkhomenko , Punnaiah Choudary Kalluri , Manish Narani , Dinh Nguyen , , , Subject: [PATCH RESEND v2 18/18] EDAC/synopsys: Add mapping-based memory size calculation Date: Sat, 10 Sep 2022 22:50:07 +0300 Message-ID: <20220910195007.11027-19-Sergey.Semin@baikalelectronics.ru> In-Reply-To: <20220910195007.11027-1-Sergey.Semin@baikalelectronics.ru> References: <20220910195007.11027-1-Sergey.Semin@baikalelectronics.ru> MIME-Version: 1.0 X-ClientProxiedBy: MAIL.baikal.int (192.168.51.25) To mail (192.168.51.25) X-CRM114-Version: 20100106-BlameMichelson ( TRE 0.8.0 (BSD) ) MR-646709E3 X-CRM114-CacheID: sfid-20220910_125052_763776_51933C3A X-CRM114-Status: GOOD ( 18.33 ) X-BeenThere: linux-arm-kernel@lists.infradead.org X-Mailman-Version: 2.1.34 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: "linux-arm-kernel" Errors-To: linux-arm-kernel-bounces+linux-arm-kernel=archiver.kernel.org@lists.infradead.org Currently the size of the memory attached to the controller is retrieved by means of the si_meminfo() method. It isn't quite correct because the system may have more than one memory controller. There is a better and more portable approach available to find out the attached memory size. Since now we have the full HIF/SDRAM mapping table available right in the device probe procedure and the DQ-bus width is detected at that stage too, that info can be used to calculate the total memory size accessible over the corresponding DW uMCTL2 DDR controller. It can be done since the hardware reference manual demands that none two SDRAM bits are mapped to the same HIF bit [1] and that the unused SDRAM address bits mapping must be disabled [2]. Note the size calculation procedure takes the ranks mapping into account. That part will be removed after we add the multi-ranked MC registration. [1] DesignWare® Cores Enhanced Universal DDR Memory Controller (uMCTL2) Databook, Version 3.91a, October 2020, p.108 [2] DesignWare® Cores Enhanced Universal DDR Memory Controller (uMCTL2) Databook, Version 3.91a, October 2020, p.109 Signed-off-by: Serge Semin --- drivers/edac/synopsys_edac.c | 66 ++++++++++++++++++++++++++---------- 1 file changed, 49 insertions(+), 17 deletions(-) diff --git a/drivers/edac/synopsys_edac.c b/drivers/edac/synopsys_edac.c index 84bede647947..e1e5468acbd2 100644 --- a/drivers/edac/synopsys_edac.c +++ b/drivers/edac/synopsys_edac.c @@ -960,20 +960,6 @@ static inline enum dev_type snps_get_dtype(u32 mstr) return DEV_UNKNOWN; } -/** - * snps_get_memsize - Read the size of the attached memory device. - * - * Return: the memory size in bytes. - */ -static u32 snps_get_memsize(void) -{ - struct sysinfo inf; - - si_meminfo(&inf); - - return inf.totalram * inf.mem_unit; -} - /** * snps_get_mtype - Returns controller memory type. * @mstr: Master CSR value. @@ -1388,6 +1374,51 @@ static void snps_get_addr_map(struct snps_edac_priv *priv) snps_get_hif_rank_map(priv, regval); } +/** + * snps_get_sdram_size - Calculate SDRAM size. + * @priv: DDR memory controller private data. + * + * The total size of the attached memory is calculated based on the HIF/SDRAM + * mapping table. It can be done since the hardware reference manual demands + * that none two SDRAM bits should be mapped to the same HIF bit and that the + * unused SDRAM address bits mapping must be disabled. + * + * Return: the memory size in bytes. + */ +static u64 snps_get_sdram_size(struct snps_edac_priv *priv) +{ + struct snps_hif_sdram_map *map = &priv->hif_sdram_map; + u64 size = 0; + int i; + + for (i = 0; i < DDR_MAX_ROW_WIDTH; i++) { + if (map->row[i] != DDR_ADDRMAP_UNUSED) + size++; + } + + for (i = 0; i < DDR_MAX_COL_WIDTH; i++) { + if (map->col[i] != DDR_ADDRMAP_UNUSED) + size++; + } + + for (i = 0; i < DDR_MAX_BANK_WIDTH; i++) { + if (map->bank[i] != DDR_ADDRMAP_UNUSED) + size++; + } + + for (i = 0; i < DDR_MAX_BANKGRP_WIDTH; i++) { + if (map->bankgrp[i] != DDR_ADDRMAP_UNUSED) + size++; + } + + for (i = 0; i < DDR_MAX_RANK_WIDTH; i++) { + if (map->rank[i] != DDR_ADDRMAP_UNUSED) + size++; + } + + return 1ULL << (size + priv->info.dq_width); +} + /** * snps_init_csrows - Initialize the csrow data. * @mci: EDAC memory controller instance. @@ -1400,7 +1431,8 @@ static void snps_init_csrows(struct mem_ctl_info *mci) struct snps_edac_priv *priv = mci->pvt_info; struct csrow_info *csi; struct dimm_info *dimm; - u32 size, row, width; + u32 row, width; + u64 size; int j; /* Actual SDRAM-word width for which ECC is calculated */ @@ -1408,13 +1440,13 @@ static void snps_init_csrows(struct mem_ctl_info *mci) for (row = 0; row < mci->nr_csrows; row++) { csi = mci->csrows[row]; - size = snps_get_memsize(); + size = snps_get_sdram_size(priv); for (j = 0; j < csi->nr_channels; j++) { dimm = csi->channels[j]->dimm; dimm->edac_mode = EDAC_SECDED; dimm->mtype = priv->info.sdram_mode; - dimm->nr_pages = (size >> PAGE_SHIFT) / csi->nr_channels; + dimm->nr_pages = PHYS_PFN(size) / csi->nr_channels; dimm->grain = width; dimm->dtype = priv->info.dev_cfg; }