From nobody Mon Dec 1 22:36:52 2025 Received: from bali.collaboradmins.com (bali.collaboradmins.com [148.251.105.195]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 7DAF32D028A; Fri, 28 Nov 2025 18:52:59 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=148.251.105.195 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1764355982; cv=none; b=ix4BaTMCvVoLndsguZ2AdXrY4KBA40OIdSdd8BjG+cO5EL9S8tKUToNGiSsdxyxmRU3HY6LRCwuvmjTn4U8Ab4gS5npZ5SOnQ/eaOLDYXrWu0TrM2munTRsJrqXWu9Rz/nYwqgqskvVTj5MYAFNsrXiCj1QIp8qL3f855nZUd/0= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1764355982; c=relaxed/simple; bh=PO8gWy+x68NFPFgtgxVVBpssyG+zZ/+aShO0x1+KKAs=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version:Content-Type; b=LbdYQ16HY7zQgU4PgdIkYUAtM5v4rXEZWvRIVp9tM3OvIOmomplwin2ANXcXi5jvMmLObRheeCv8VxknmVsmLnHcXV1oF1hP5pkEEJkUFBy4O9wrcra8sr89z56M6dBE8852ooXmQnd0585xlIRB6FEyc9arOHuCTuIf9ybeOII= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=collabora.com; spf=pass smtp.mailfrom=collabora.com; dkim=pass (2048-bit key) header.d=collabora.com header.i=@collabora.com header.b=oADkikIm; arc=none smtp.client-ip=148.251.105.195 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=collabora.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=collabora.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=collabora.com header.i=@collabora.com header.b="oADkikIm" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=collabora.com; s=mail; t=1764355977; bh=PO8gWy+x68NFPFgtgxVVBpssyG+zZ/+aShO0x1+KKAs=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=oADkikIm+tNB++DIyc3gm46mNvrtmcw5jOQA4gm4Ok4xQu2/yhsrQiQY/XrlZZLr/ NnNbGS9U784YiFDFZje/cLi8Ce6ayHb9WbusJLlrt+3BNM/htKAbZR/9btYx2oqUze 7CxaTMCwc9bqrRYQkCsIiTO4RPHb4Q2/91N71nG7sdakSXbo1fdAeWzZBSRO3uuNXe hOiW+79f5Mk5T+uAAw2IMt2g3kK7IQwcJBtlsXhLIL15KMq/4fdTPYdOnNK9ruflBD uCT9LmCDEAaVQS6YdGv5Ib6vQ9qPDEfOFdjDVwiyKxO+8+LofbX8O/sGQNyVdPCgxH gkjR+dw+nq/8Q== Received: from debian-rockchip-rock5b-rk3588.. (unknown [IPv6:2a01:e0a:5e3:6100:2e0:4cff:fe03:d8c]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits) server-digest SHA256) (No client certificate requested) (Authenticated sender: loicmolinari) by bali.collaboradmins.com (Postfix) with ESMTPSA id E526217E1146; Fri, 28 Nov 2025 19:52:56 +0100 (CET) From: =?UTF-8?q?Lo=C3=AFc=20Molinari?= To: Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann , David Airlie , Simona Vetter , Jani Nikula , Joonas Lahtinen , Rodrigo Vivi , Tvrtko Ursulin , Boris Brezillon , Rob Herring , Steven Price , Liviu Dudau , Melissa Wen , =?UTF-8?q?Ma=C3=ADra=20Canal?= , Hugh Dickins , Baolin Wang , Andrew Morton , =?UTF-8?q?Lo=C3=AFc=20Molinari?= , Al Viro , =?UTF-8?q?Miko=C5=82aj=20Wasiak?= , Christian Brauner , Nitin Gote , Andi Shyti , Jonathan Corbet , Christopher Healy , Matthew Wilcox , Bagas Sanjaya Cc: linux-kernel@vger.kernel.org, dri-devel@lists.freedesktop.org, intel-gfx@lists.freedesktop.org, linux-mm@kvack.org, linux-doc@vger.kernel.org, kernel@collabora.com Subject: [PATCH v10 02/10] drm/shmem-helper: Map huge pages in fault handler Date: Fri, 28 Nov 2025 19:52:44 +0100 Message-ID: <20251128185252.3092-3-loic.molinari@collabora.com> X-Mailer: git-send-email 2.47.3 In-Reply-To: <20251128185252.3092-1-loic.molinari@collabora.com> References: <20251128185252.3092-1-loic.molinari@collabora.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Attempt a PMD sized PFN insertion into the VMA if the faulty address of the fault handler is part of a huge page. On builds with CONFIG_TRANSPARENT_HUGEPAGE enabled, if the mmap() user address is PMD size aligned, if the GEM object is backed by shmem buffers on mountpoints setting the 'huge=3D' option and if the shmem backing store manages to allocate a huge folio, CPU mapping would then benefit from significantly increased memcpy() performance. When these conditions are met on a system with 2 MiB huge pages, an aligned copy of 2 MiB would raise a single page fault instead of 4096. v4: - implement map_pages instead of huge_fault v6: - get rid of map_pages handler for now (keep it for another series along with arm64 contpte support) Signed-off-by: Lo=C3=AFc Molinari Reviewed-by: Boris Brezillon --- drivers/gpu/drm/drm_gem_shmem_helper.c | 55 +++++++++++++++++++++----- 1 file changed, 46 insertions(+), 9 deletions(-) diff --git a/drivers/gpu/drm/drm_gem_shmem_helper.c b/drivers/gpu/drm/drm_g= em_shmem_helper.c index be89be1c804c..81f4ac7cb8f6 100644 --- a/drivers/gpu/drm/drm_gem_shmem_helper.c +++ b/drivers/gpu/drm/drm_gem_shmem_helper.c @@ -567,31 +567,68 @@ int drm_gem_shmem_dumb_create(struct drm_file *file, = struct drm_device *dev, } EXPORT_SYMBOL_GPL(drm_gem_shmem_dumb_create); =20 +static bool drm_gem_shmem_fault_is_valid(struct drm_gem_object *obj, + pgoff_t pgoff) +{ + struct drm_gem_shmem_object *shmem =3D to_drm_gem_shmem_obj(obj); + + if (drm_WARN_ON_ONCE(obj->dev, !shmem->pages) || + pgoff >=3D (obj->size >> PAGE_SHIFT) || + shmem->madv < 0) + return false; + + return true; +} + +static bool drm_gem_shmem_map_pmd(struct vm_fault *vmf, unsigned long addr, + struct page *page) +{ +#ifdef CONFIG_ARCH_SUPPORTS_PMD_PFNMAP + unsigned long pfn =3D page_to_pfn(page); + unsigned long paddr =3D pfn << PAGE_SHIFT; + bool aligned =3D (addr & ~PMD_MASK) =3D=3D (paddr & ~PMD_MASK); + + if (aligned && + pmd_none(*vmf->pmd) && + folio_test_pmd_mappable(page_folio(page))) { + pfn &=3D PMD_MASK >> PAGE_SHIFT; + if (vmf_insert_pfn_pmd(vmf, pfn, false) =3D=3D VM_FAULT_NOPAGE) + return true; + } +#endif + + return false; +} + static vm_fault_t drm_gem_shmem_fault(struct vm_fault *vmf) { struct vm_area_struct *vma =3D vmf->vma; struct drm_gem_object *obj =3D vma->vm_private_data; struct drm_gem_shmem_object *shmem =3D to_drm_gem_shmem_obj(obj); - loff_t num_pages =3D obj->size >> PAGE_SHIFT; - vm_fault_t ret; - struct page *page; + struct page **pages =3D shmem->pages; pgoff_t page_offset; + unsigned long pfn; + vm_fault_t ret; =20 /* Offset to faulty address in the VMA (without the fake offset). */ page_offset =3D vmf->pgoff - vma->vm_pgoff; =20 dma_resv_lock(shmem->base.resv, NULL); =20 - if (page_offset >=3D num_pages || - drm_WARN_ON_ONCE(obj->dev, !shmem->pages) || - shmem->madv < 0) { + if (unlikely(!drm_gem_shmem_fault_is_valid(obj, page_offset))) { ret =3D VM_FAULT_SIGBUS; - } else { - page =3D shmem->pages[page_offset]; + goto out; + } =20 - ret =3D vmf_insert_pfn(vma, vmf->address, page_to_pfn(page)); + if (drm_gem_shmem_map_pmd(vmf, vmf->address, pages[page_offset])) { + ret =3D VM_FAULT_NOPAGE; + goto out; } =20 + pfn =3D page_to_pfn(pages[page_offset]); + ret =3D vmf_insert_pfn(vma, vmf->address, pfn); + + out: dma_resv_unlock(shmem->base.resv); =20 return ret; --=20 2.47.3