From mboxrd@z Thu Jan 1 00:00:00 1970 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 BAA38CD5BD4 for ; Thu, 5 Sep 2024 16:24:13 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=bombadil.20210309; h=Sender:List-Subscribe:List-Help :List-Post:List-Archive:List-Unsubscribe:List-Id:Content-Transfer-Encoding: Content-Type:In-Reply-To:From:References:Cc:To:Subject:MIME-Version:Date: Message-ID:Reply-To:Content-ID:Content-Description:Resent-Date:Resent-From: Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:List-Owner; bh=BNRyKjH1muO1XoWdooXBVBh3UVWl9K/c06REWgMdetA=; b=PLvdXtevFyQx5FdwpX1Se+ENpx 4jDV8ljlWmJByO2+osLPw4+YDDvvHW4L6PK5+HCoA8ur44RNYcZfyOJZJukyr7QCpvRCEqWGNFCMG WYFx+6CcMAQB/tSzl6XwpIjWl7WM0yWXwLFvmQmW/SNF1Yy8B/WJJzAp8AYkyq5pEwK5GMWCNLk/i nZv+mpn5Ipsng1QesE69J0nLLdRNl/LeQcBcHbnuFqc4BUWGgR+FnJWY7Bb4bL2M3Jm2tuwg/TWu6 48tSvPyVUFFK9Csgsa77Wd1PqSoC/lF8+kb3+Nl5o9ItKNZ5Q0EiNXKSfmzcHRN8qqIsSH2LnSr7z AMMuLHug==; Received: from localhost ([::1] helo=bombadil.infradead.org) by bombadil.infradead.org with esmtp (Exim 4.97.1 #2 (Red Hat Linux)) id 1smFH6-000000096Ba-1wkD; Thu, 05 Sep 2024 16:24:00 +0000 Received: from foss.arm.com ([217.140.110.172]) by bombadil.infradead.org with esmtp (Exim 4.97.1 #2 (Red Hat Linux)) id 1smFG6-000000095x9-3aCv for linux-arm-kernel@lists.infradead.org; Thu, 05 Sep 2024 16:23:00 +0000 Received: from usa-sjc-imap-foss1.foss.arm.com (unknown [10.121.207.14]) by usa-sjc-mx-foss1.foss.arm.com (Postfix) with ESMTP id 81016FEC; Thu, 5 Sep 2024 09:23:24 -0700 (PDT) Received: from [10.1.196.40] (e121345-lin.cambridge.arm.com [10.1.196.40]) by usa-sjc-imap-foss1.foss.arm.com (Postfix) with ESMTPSA id 342573F73B; Thu, 5 Sep 2024 09:22:53 -0700 (PDT) Message-ID: <7f4ee8d7-0e17-4363-bf91-d67846157c67@arm.com> Date: Thu, 5 Sep 2024 17:22:51 +0100 MIME-Version: 1.0 User-Agent: Mozilla Thunderbird Subject: Re: [PATCH] Revert "iommu/io-pgtable-arm: Optimise non-coherent unmap" To: Rob Clark Cc: iommu@lists.linux.dev, linux-arm-msm@vger.kernel.org, freedreno@lists.freedesktop.org, Ashish Mhetre , Rob Clark , Will Deacon , Joerg Roedel , "moderated list:ARM SMMU DRIVERS" , open list References: <20240905124956.84932-1-robdclark@gmail.com> From: Robin Murphy Content-Language: en-GB In-Reply-To: Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 8bit X-CRM114-Version: 20100106-BlameMichelson ( TRE 0.8.0 (BSD) ) MR-646709E3 X-CRM114-CacheID: sfid-20240905_092259_006911_998D2DB0 X-CRM114-Status: GOOD ( 28.65 ) 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 On 05/09/2024 2:57 pm, Rob Clark wrote: > On Thu, Sep 5, 2024 at 6:24 AM Robin Murphy wrote: >> >> On 05/09/2024 1:49 pm, Rob Clark wrote: >>> From: Rob Clark >>> >>> This reverts commit 85b715a334583488ad7fbd3001fe6fd617b7d4c0. >>> >>> It was causing gpu smmu faults on x1e80100. >>> >>> I _think_ what is causing this is the change in ordering of >>> __arm_lpae_clear_pte() (dma_sync_single_for_device() on the pgtable >>> memory) and io_pgtable_tlb_flush_walk(). >> >> As I just commented, how do you believe the order of operations between: >> >> __arm_lpae_clear_pte(); >> if (!iopte_leaf()) { >> io_pgtable_tlb_flush_walk(); >> >> and: >> >> if (!iopte_leaf()) { >> __arm_lpae_clear_pte(); >> io_pgtable_tlb_flush_walk(); >> >> fundamentally differs? > > from my reading of the original patch, the ordering is the same for > non-leaf nodes, but not for leaf nodes But tlb_flush_walk is never called for leaf entries either way, so no such ordering exists... And the non-leaf path still calls __arm_lpae_clear_pte() and io_pgtable_tlb_add_page() in the same order relative to each other too, it's just now done for the whole range in one go, after any non-leaf entries have already been dealt with. Thanks, Robin. > >> I'm not saying there couldn't be some subtle bug in the implementation >> which we've all missed, but I still can't see an issue with the intended >> logic. >> >>> I'm not entirely sure how >>> this patch is supposed to work correctly in the face of other >>> concurrent translations (to buffers unrelated to the one being >>> unmapped(), because after the io_pgtable_tlb_flush_walk() we can have >>> stale data read back into the tlb. >> >> Read back from where? The ex-table PTE which was already set to zero >> before tlb_flush_walk was called? >> >> And isn't the hilariously overcomplicated TBU driver supposed to be >> telling you exactly what happened here? Otherwise I'm going to continue >> to seriously question the purpose of shoehorning that upstream at all... > > I guess I could try the TBU driver. But I already had my patchset to > extract the pgtable walk for gpu devcore dump, and that is telling me > that the CPU view of the pgtable is fine. Which I think just leaves a > tlbinv problem. If that is the case, swapping the order of leaf node > cpu cache ops and tlbinv ops seems like the cause. But maybe I'm > missing something. > > BR, > -R > >> Thanks, >> Robin. >> >>> Signed-off-by: Rob Clark >>> --- >>> drivers/iommu/io-pgtable-arm.c | 31 ++++++++++++++----------------- >>> 1 file changed, 14 insertions(+), 17 deletions(-) >>> >>> diff --git a/drivers/iommu/io-pgtable-arm.c b/drivers/iommu/io-pgtable-arm.c >>> index 16e51528772d..85261baa3a04 100644 >>> --- a/drivers/iommu/io-pgtable-arm.c >>> +++ b/drivers/iommu/io-pgtable-arm.c >>> @@ -274,13 +274,13 @@ static void __arm_lpae_sync_pte(arm_lpae_iopte *ptep, int num_entries, >>> sizeof(*ptep) * num_entries, DMA_TO_DEVICE); >>> } >>> >>> -static void __arm_lpae_clear_pte(arm_lpae_iopte *ptep, struct io_pgtable_cfg *cfg, int num_entries) >>> +static void __arm_lpae_clear_pte(arm_lpae_iopte *ptep, struct io_pgtable_cfg *cfg) >>> { >>> - for (int i = 0; i < num_entries; i++) >>> - ptep[i] = 0; >>> >>> - if (!cfg->coherent_walk && num_entries) >>> - __arm_lpae_sync_pte(ptep, num_entries, cfg); >>> + *ptep = 0; >>> + >>> + if (!cfg->coherent_walk) >>> + __arm_lpae_sync_pte(ptep, 1, cfg); >>> } >>> >>> static size_t __arm_lpae_unmap(struct arm_lpae_io_pgtable *data, >>> @@ -653,28 +653,25 @@ static size_t __arm_lpae_unmap(struct arm_lpae_io_pgtable *data, >>> max_entries = ARM_LPAE_PTES_PER_TABLE(data) - unmap_idx_start; >>> num_entries = min_t(int, pgcount, max_entries); >>> >>> - /* Find and handle non-leaf entries */ >>> - for (i = 0; i < num_entries; i++) { >>> - pte = READ_ONCE(ptep[i]); >>> + while (i < num_entries) { >>> + pte = READ_ONCE(*ptep); >>> if (WARN_ON(!pte)) >>> break; >>> >>> - if (!iopte_leaf(pte, lvl, iop->fmt)) { >>> - __arm_lpae_clear_pte(&ptep[i], &iop->cfg, 1); >>> + __arm_lpae_clear_pte(ptep, &iop->cfg); >>> >>> + if (!iopte_leaf(pte, lvl, iop->fmt)) { >>> /* Also flush any partial walks */ >>> io_pgtable_tlb_flush_walk(iop, iova + i * size, size, >>> ARM_LPAE_GRANULE(data)); >>> __arm_lpae_free_pgtable(data, lvl + 1, iopte_deref(pte, data)); >>> + } else if (!iommu_iotlb_gather_queued(gather)) { >>> + io_pgtable_tlb_add_page(iop, gather, iova + i * size, size); >>> } >>> - } >>> >>> - /* Clear the remaining entries */ >>> - __arm_lpae_clear_pte(ptep, &iop->cfg, i); >>> - >>> - if (gather && !iommu_iotlb_gather_queued(gather)) >>> - for (int j = 0; j < i; j++) >>> - io_pgtable_tlb_add_page(iop, gather, iova + j * size, size); >>> + ptep++; >>> + i++; >>> + } >>> >>> return i * size; >>> } else if (iopte_leaf(pte, lvl, iop->fmt)) {