Cover Letter Detail
Show a cover letter.
GET /api/1.1/covers/2222318/?format=api
{ "id": 2222318, "url": "http://patchwork.ozlabs.org/api/1.1/covers/2222318/?format=api", "web_url": "http://patchwork.ozlabs.org/project/linux-pci/cover/20260411080148.471335-1-kwilczynski@kernel.org/", "project": { "id": 28, "url": "http://patchwork.ozlabs.org/api/1.1/projects/28/?format=api", "name": "Linux PCI development", "link_name": "linux-pci", "list_id": "linux-pci.vger.kernel.org", "list_email": "linux-pci@vger.kernel.org", "web_url": null, "scm_url": null, "webscm_url": null }, "msgid": "<20260411080148.471335-1-kwilczynski@kernel.org>", "date": "2026-04-11T08:01:24", "name": "[v4,00/24] PCI: Convert all dynamic sysfs attributes to static", "submitter": { "id": 86709, "url": "http://patchwork.ozlabs.org/api/1.1/people/86709/?format=api", "name": "Krzysztof Wilczyński", "email": "kwilczynski@kernel.org" }, "mbox": "http://patchwork.ozlabs.org/project/linux-pci/cover/20260411080148.471335-1-kwilczynski@kernel.org/mbox/", "series": [ { "id": 499539, "url": "http://patchwork.ozlabs.org/api/1.1/series/499539/?format=api", "web_url": "http://patchwork.ozlabs.org/project/linux-pci/list/?series=499539", "date": "2026-04-11T08:01:24", "name": "PCI: Convert all dynamic sysfs attributes to static", "version": 4, "mbox": "http://patchwork.ozlabs.org/series/499539/mbox/" } ], "comments": "http://patchwork.ozlabs.org/api/covers/2222318/comments/", "headers": { "Return-Path": "\n <linux-pci+bounces-52368-incoming=patchwork.ozlabs.org@vger.kernel.org>", "X-Original-To": [ "incoming@patchwork.ozlabs.org", "linux-pci@vger.kernel.org" ], "Delivered-To": "patchwork-incoming@legolas.ozlabs.org", "Authentication-Results": [ "legolas.ozlabs.org;\n\tdkim=pass (2048-bit key;\n unprotected) header.d=kernel.org header.i=@kernel.org header.a=rsa-sha256\n header.s=k20201202 header.b=NIsjok3Z;\n\tdkim-atps=neutral", "legolas.ozlabs.org;\n spf=pass (sender SPF authorized) smtp.mailfrom=vger.kernel.org\n (client-ip=104.64.211.4; helo=sin.lore.kernel.org;\n envelope-from=linux-pci+bounces-52368-incoming=patchwork.ozlabs.org@vger.kernel.org;\n receiver=patchwork.ozlabs.org)", "smtp.subspace.kernel.org;\n\tdkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org\n header.b=\"NIsjok3Z\"", "smtp.subspace.kernel.org;\n arc=none smtp.client-ip=10.30.226.201" ], "Received": [ "from sin.lore.kernel.org (sin.lore.kernel.org [104.64.211.4])\n\t(using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits)\n\t key-exchange x25519)\n\t(No client certificate requested)\n\tby legolas.ozlabs.org (Postfix) with ESMTPS id 4ft5jh67pGz1yCx\n\tfor <incoming@patchwork.ozlabs.org>; Sat, 11 Apr 2026 18:02:00 +1000 (AEST)", "from smtp.subspace.kernel.org (conduit.subspace.kernel.org\n [100.90.174.1])\n\tby sin.lore.kernel.org (Postfix) with ESMTP id DDC4C300E1B4\n\tfor <incoming@patchwork.ozlabs.org>; Sat, 11 Apr 2026 08:01:56 +0000 (UTC)", "from localhost.localdomain (localhost.localdomain [127.0.0.1])\n\tby smtp.subspace.kernel.org (Postfix) with ESMTP id 08291221F39;\n\tSat, 11 Apr 2026 08:01:54 +0000 (UTC)", "from smtp.kernel.org (aws-us-west-2-korg-mail-1.web.codeaurora.org\n [10.30.226.201])\n\t(using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits))\n\t(No client certificate requested)\n\tby smtp.subspace.kernel.org (Postfix) with ESMTPS id D67212459E5;\n\tSat, 11 Apr 2026 08:01:53 +0000 (UTC)", "by smtp.kernel.org (Postfix) with ESMTPSA id 54622C4CEF7;\n\tSat, 11 Apr 2026 08:01:50 +0000 (UTC)" ], "ARC-Seal": "i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116;\n\tt=1775894513; cv=none;\n b=IIt9lJKAEP95xFWJo7C2Rg4cdhGqbjpwZKFjmgLzYJOXOCTqwNPXejJb1vqxRcSgBXc5WosgjsBH2qCwfd+/Bm/VBYZ6IB14k1nSivXx8wlZDAfyEAUdrd2YnNvVY7g5ZH6Zq8WK2SvjWxge7ZdRYM7kturrCXkM7wl2Q8WlUt4=", "ARC-Message-Signature": "i=1; a=rsa-sha256; d=subspace.kernel.org;\n\ts=arc-20240116; t=1775894513; c=relaxed/simple;\n\tbh=Q899Jnv+KvcIpCetyXKm8aXkRrDGbgjIM8PWb00bmEk=;\n\th=From:To:Cc:Subject:Date:Message-ID:MIME-Version:Content-Type;\n b=EcxCPkD9psIPBQ5zIm5qo0LbLW4JN13pNRlvp42ijCFV8KRP6s1mw1/mR3ZenX41tkDhIeF6YwQWQkJj5bH0gsYfkqDqzbaH5TjeUAsD5rPtna8Hn/szZvsvu04yhcLiL3SgxKKeOODb8rN1YhYJ9K+GyB/uiEflAa6q5JtahDE=", "ARC-Authentication-Results": "i=1; smtp.subspace.kernel.org;\n dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org\n header.b=NIsjok3Z; arc=none smtp.client-ip=10.30.226.201", "DKIM-Signature": "v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org;\n\ts=k20201202; t=1775894513;\n\tbh=Q899Jnv+KvcIpCetyXKm8aXkRrDGbgjIM8PWb00bmEk=;\n\th=From:To:Cc:Subject:Date:From;\n\tb=NIsjok3Z/wMalNtn6zn5F0LWmfCVpJiqTF04jyxSBR6ClgweUhNuvsqP1kNNcCXl9\n\t yEZUhIW/nYDaf/3G75jf49DT1auLPESFIOZivrKuE+rSd8M0uZ3JiFa68B2a/5Ozzc\n\t EJPgqkzbPDgM5HdC7xCmfKIfnADHA9ctj3x9ciMTH5HQHKrtsEOlebrjhU2obeFA/E\n\t hQuIwBzl660nyDY6EHqSp0tkLJYBoTrNCA3JqL9JFIjnACheQeoixM1sYBx3JIZzJz\n\t wQb15nsMzq3ujH46MTLhP6Vwamf58U4A05RCF5olklkHGhcbrIBQlLJkdpqRQmJKpG\n\t ecpJx4gVSqUcA==", "From": "=?utf-8?q?Krzysztof_Wilczy=C5=84ski?= <kwilczynski@kernel.org>", "To": "Bjorn Helgaas <bhelgaas@google.com>", "Cc": "Bjorn Helgaas <helgaas@kernel.org>,\n Manivannan Sadhasivam <mani@kernel.org>,\n Lorenzo Pieralisi <lpieralisi@kernel.org>,\n Magnus Lindholm <linmag7@gmail.com>, Matt Turner <mattst88@gmail.com>,\n Richard Henderson <richard.henderson@linaro.org>,\n Christophe Leroy <chleroy@kernel.org>,\n Madhavan Srinivasan <maddy@linux.ibm.com>,\n Michael Ellerman <mpe@ellerman.id.au>, Nicholas Piggin <npiggin@gmail.com>,\n Dexuan Cui <decui@microsoft.com>,\n =?utf-8?q?Krzysztof_Ha=C5=82asa?= <khalasa@piap.pl>,\n Lukas Wunner <lukas@wunner.de>, \"Oliver O'Halloran\" <oohall@gmail.com>,\n Saurabh Singh Sengar <ssengar@microsoft.com>,\n Shuan He <heshuan@bytedance.com>, Srivatsa Bhat <srivatsabhat@microsoft.com>,\n\t=?utf-8?q?Ilpo_J=C3=A4rvinen?= <ilpo.jarvinen@linux.intel.com>,\n linux-pci@vger.kernel.org, linux-alpha@vger.kernel.org,\n linuxppc-dev@lists.ozlabs.org", "Subject": "[PATCH v4 00/24] PCI: Convert all dynamic sysfs attributes to static", "Date": "Sat, 11 Apr 2026 08:01:24 +0000", "Message-ID": "<20260411080148.471335-1-kwilczynski@kernel.org>", "X-Mailer": "git-send-email 2.53.0", "Precedence": "bulk", "X-Mailing-List": "linux-pci@vger.kernel.org", "List-Id": "<linux-pci.vger.kernel.org>", "List-Subscribe": "<mailto:linux-pci+subscribe@vger.kernel.org>", "List-Unsubscribe": "<mailto:linux-pci+unsubscribe@vger.kernel.org>", "MIME-Version": "1.0", "Content-Type": "text/plain; charset=UTF-8", "Content-Transfer-Encoding": "8bit" }, "content": "Hello,\n\nThis series converts every dynamically allocated PCI sysfs attribute to\na static const definition. After the full series, pci_sysfs_init() and\nsysfs_initialized are gone, and every sysfs file is created by the\ndriver model at device_add() time.\n\nCurrently, the PCI resource files (resourceN, resourceN_wc) and the\nlegacy bus files (legacy_io, legacy_mem) are created dynamically\nfrom two unsynchronised paths:\n\nPath A: late_initcall\n\n pci_sysfs_init (late_initcall)\n sysfs_initialized = 1\n for_each_pci_dev\n pci_create_sysfs_dev_files\n sysfs_create_bin_file (resourceN, resourceN_wc)\n pci_find_next_bus\n pci_create_legacy_files\n sysfs_create_bin_file (legacy_io, legacy_mem)\n\nPath B: device registration / hotplug\n\n pci_bus_add_devices\n pci_bus_add_device\n pci_create_sysfs_dev_files\n if (!sysfs_initialized) return <- only guard\n sysfs_create_bin_file (resourceN, resourceN_wc)\n\nOn most ACPI systems this does not race because PCI enumeration\ncompletes at subsys_initcall time, before pci_sysfs_init() runs:\n\n subsys_initcall (level 4):\n acpi_pci_root_add\n pci_bus_add_device\n pci_create_sysfs_dev_files\n if (!sysfs_initialized) <- not yet set\n return -EACCES\n\n late_initcall (level 7):\n pci_sysfs_init\n sysfs_initialized = 1\n for_each_pci_dev\n pci_create_sysfs_dev_files <- creates the files, no race\n\nOn Devicetree platforms the host controller is a platform driver that\nprobes via the driver model, often on a workqueue, and overlaps with the\nlate_initcall:\n\n CPU 0 (late_initcall) CPU 1 (driver probe)\n --------------------------- ----------------------------\n pci_sysfs_init()\n sysfs_initialized = 1\n for_each_pci_dev(pdev) pci_bus_add_device(pdev)\n pci_create_sysfs_dev_files() pci_create_sysfs_dev_files()\n sysfs_create_bin_file() sysfs_create_bin_file()\n -> \"duplicate filename\"\n\nThe same happens on ACPI when probing is asynchronous (hv_pci on\nAzure, RISC-V with ACPI).\n\nThe duplicate causes sysfs_create_bin_file() to fail with -EEXIST.\npci_create_resource_files() then calls pci_remove_resource_files() in\nits error unwind, tearing down files the other thread created and\nstill references through pdev->res_attr[]. This has caused kernel\npanics on i.MX6 and boot failures on other platforms.\n\nSeveral different fixes have been proposed over the years: reordering\nthe sysfs_initialized assignment, adding locks, checking\npci_dev_is_added(), setting pdev->res_attr[] to NULL after kfree\n(which only prevents a double-free on the teardown path, not the\nerror unwind removing the other thread's files). None would address the\nroot cause.\n\nThis has been reported a few times:\n\n - https://lore.kernel.org/linux-pci/20250702155112.40124-1-heshuan@bytedance.com/\n - https://lore.kernel.org/linux-pci/b51519d6-ce45-4b6d-8135-c70169bd110e@h-partners.com/\n - https://lore.kernel.org/linux-pci/1702093576-30405-1-git-send-email-ssengar@linux.microsoft.com/\n - https://lore.kernel.org/linux-pci/SY0P300MB04687548090B73E40AF97D8897B82@SY0P300MB0468.AUSP300.PROD.OUTLOOK.COM/\n - https://lore.kernel.org/linux-pci/20230105174736.GA1154719@bhelgaas/\n - https://lore.kernel.org/linux-pci/m3eebg9puj.fsf@t19.piap.pl/\n - https://lore.kernel.org/linux-pci/20200716110423.xtfyb3n6tn5ixedh@pali/\n - https://lore.kernel.org/linux-pci/1366196798-15929-1-git-send-email-artem.savkov@gmail.com/\n - https://bugzilla.kernel.org/show_bug.cgi?id=215515\n - https://bugzilla.kernel.org/show_bug.cgi?id=216888\n\nWith static attributes the driver model creates sysfs entries once per\ndevice at device_add() time, under the device lock, eliminating the\nlate_initcall iteration and the race along with it.\n\n\tKrzysztof\n\n---\nChanges in v4:\n https://lore.kernel.org/linux-pci/20260410055040.39233-1-kwilczynski@kernel.org/\n\n - Added new Reviewed-by tags.\n - Added pci_resource_is_io() and pci_resource_is_mem() helpers\n for resource type checks, replacing the open-coded bitwise\n flag tests in pci_mmap_resource(), pci_resource_io(), and\n Alpha's pci_mmap_resource(), as per Ilpo Järvinen's\n suggestion.\n - Split the __pci_mmap_fits() cleanup into two patches. An\n overflow fix for zero-length BARs, which now includes a\n Fixes: tag referencing the original Alpha PCI sysfs commit,\n and the WARN macro removal is a separate cleanup as per Ilpo\n Järvinen's suggestion.\n - Added a missing Fixes: tag to the Alpha lockdown check,\n referencing the commit that added the check to the generic\n path but missed Alpha's implementation.\n - Added PCI_LEGACY_IO_SIZE and PCI_LEGACY_MEM_SIZE macros to\n replace the raw literals used for legacy address space sizes.\n These are used in both Alpha's pci_mmap_legacy_page_range()\n and the static legacy attribute definitions, as per Ilpo\n Järvinen's suggestion.\n - Replaced sysfs_update_groups() in the BAR resize path with\n sysfs_remove_groups() before the resize and sysfs_create_groups()\n after, restoring the original teardown before BAR resize\n ordering. This was reported by Sashiko, see:\n https://sashiko.dev/#/patchset/20260410055040.39233-1-kwilczynski%40kernel.org?part=7\n - Defined pci_dev_resource_attr_groups as a NULL macro when\n HAVE_PCI_MMAP and ARCH_GENERIC_PCI_MMAP_RESOURCE are both\n absent, so the resize path compiles unconditionally without\n #ifdef guards in the function body. This was reported by\n Sashiko, see:\n https://sashiko.dev/#/patchset/20260410055040.39233-1-kwilczynski%40kernel.org?part=7\n - Moved the pci_legacy_has_sparse() prototype into the patch\n that introduces the function, alongside the existing\n pci_adjust_legacy_attr() declaration, to fix a bisection\n issue where Alpha would warn on -Wmissing-prototypes.\n This was reported by Sashiko, see:\n https://sashiko.dev/#/patchset/20260410055040.39233-1-kwilczynski%40kernel.org?part=18\n\nChanges in v3:\n https://lore.kernel.org/linux-pci/20210910202623.2293708-1-kw@linux.com/\n\n - Updated for modern kernel releases and expanded scope. The\n v2 only covered the generic resource files. This version\n also converts Alpha's sparse/dense resource files and the\n legacy bus attributes, removing pci_sysfs_init() entirely.\n - Split the single macro definition into three distinct ones\n (per I/O, UC, and WC), to make sure that each carries only\n the callbacks its resource type needs.\n - Updated to use the new .bin_size callback, as the attributes\n are const, to replace using a->size directly, which was not\n ideal. This required changes to pci_llseek_resource(), to\n ensure that it would work for device and bus-level attributes.\n - Updated the __resource_resize_store() to include CAP_SYS_ADMIN\n capabilities check.\n - Added the security_locked_down() check to Alpha's\n pci_mmap_resource(), to align with other architectures.\n\nChanges in v2:\n https://lore.kernel.org/linux-pci/20210825212255.878043-1-kw@linux.com/\n\n - Refactored code so that the macros, helpers and internal\n functions can be used to correctly leverage the read(),\n write() and mmap() callbacks rather than to use the\n .is_bin_visible() callback to set up sysfs objects\n internals as this is not supported.\n - Refactored some if-statements to check for a resource\n flag first, and then call either arch_can_pci_mmap_io()\n or arch_can_pci_mmap_wc(), plus store result of testing\n for IORESOURCE_MEM and IORESOURCE_PREFETCH flags into\n a boolean variable, as per Bjorn Helgaas' suggestion.\n - Renamed pci_read_resource_io() and pci_write_resource_io()\n callbacks so that these are not specifically tied to I/O\n BARs read() and write() operations also as per Bjorn\n Helgaas' suggestion.\n - Updated style for code handling bitwise operations to\n match the style that is preferred as per Bjorn Helgaas'\n suggestion.\n - Updated commit messages adding more details about the\n implementation as requested by Bjorn Helgaas.\n\nKrzysztof Wilczyński (24):\n PCI/sysfs: Use PCI resource accessor macros\n PCI: Add pci_resource_is_io() and pci_resource_is_mem() helpers\n PCI/sysfs: Only allow supported resource types in I/O and MMIO helpers\n PCI/sysfs: Use BAR length in pci_llseek_resource() when attr->size is\n zero\n PCI/sysfs: Add CAP_SYS_ADMIN check to __resource_resize_store()\n PCI/sysfs: Add static PCI resource attribute macros\n PCI/sysfs: Convert PCI resource files to static attributes\n PCI/sysfs: Convert __resource_resize_store() to use static attributes\n PCI/sysfs: Add stubs for pci_{create,remove}_sysfs_dev_files()\n PCI/sysfs: Limit pci_sysfs_init() late_initcall compile scope\n alpha/PCI: Add security_locked_down() check to pci_mmap_resource()\n alpha/PCI: Use BAR index in sysfs attr->private instead of resource\n pointer\n alpha/PCI: Use PCI resource accessor macros\n alpha/PCI: Clean up pci_mmap_resource()\n alpha/PCI: Fix __pci_mmap_fits() overflow for zero-length BARs\n alpha/PCI: Remove WARN from __pci_mmap_fits()\n alpha/PCI: Add static PCI resource attribute macros\n alpha/PCI: Convert resource files to static attributes\n PCI/sysfs: Remove pci_{create,remove}_sysfs_dev_files()\n PCI: Add macros for legacy I/O and memory address space sizes\n alpha/PCI: Compute legacy size in pci_mmap_legacy_page_range()\n PCI/sysfs: Add __weak pci_legacy_has_sparse() helper\n PCI/sysfs: Convert legacy I/O and memory attributes to static\n definitions\n PCI/sysfs: Remove pci_create_legacy_files() and pci_sysfs_init()\n\n arch/alpha/include/asm/pci.h | 13 +-\n arch/alpha/kernel/pci-sysfs.c | 373 +++++++++++----------\n arch/powerpc/include/asm/pci.h | 2 -\n drivers/pci/bus.c | 1 -\n drivers/pci/pci-sysfs.c | 575 +++++++++++++++++++--------------\n drivers/pci/pci.h | 16 +-\n drivers/pci/probe.c | 6 -\n drivers/pci/remove.c | 3 -\n include/linux/pci.h | 39 ++-\n 9 files changed, 578 insertions(+), 450 deletions(-)" }