From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from out-170.mta1.migadu.com (out-170.mta1.migadu.com [95.215.58.170]) (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 4FD3342A82 for ; Thu, 5 Feb 2026 15:01:50 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=95.215.58.170 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1770303711; cv=none; b=NEtUQ0bLvZJhM2od8vafYD5hV+G4gN+LuG7RlO0iQ4I1AVan5kf2wawFCvVODkkzFECe9j7dxyis9w9qlo0r+fCf26mfQjj8O8G7aAbdQh3ZoRAR4/Wj7U74bN9ag9umo5Rqsn90qalefM8yAUZcPjvQU2eNl7eQFcANU0EIRZY= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1770303711; c=relaxed/simple; bh=z7DlHGRAWqmVCDGHT8bYe/0iOwwIf+tGe2De5QVYgmk=; h=Message-ID:Date:MIME-Version:Subject:To:Cc:References:From: In-Reply-To:Content-Type; b=cqa94PYNBHnAp5Hbd/w+4Egkh9FeJRjxK8EjbuL+5TFvvkC2rWLoLz5DwUVXFsm7xSJ/+dDK86vouAkX+99yXB9dwzCJonB9C1+FKmNolhhvhWy7vdavXsnvwTvMwie2UlcLx3mFQQuiVaBS1z0HbLvI+t60a2dZ6f60DyT/L4o= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.dev; spf=pass smtp.mailfrom=linux.dev; dkim=pass (1024-bit key) header.d=linux.dev header.i=@linux.dev header.b=R1OGaVme; arc=none smtp.client-ip=95.215.58.170 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.dev Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=linux.dev Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=linux.dev header.i=@linux.dev header.b="R1OGaVme" Message-ID: <3026ad8d-92ad-4683-8c3e-733d4070d033@linux.dev> DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linux.dev; s=key1; t=1770303707; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=oeEv7fnd0NPkK8yO5EsGkSZ8yT0L8dwlwYeRPZ0uSkk=; b=R1OGaVmeCDAWYZLObcFuiWzoPWWNt28v4igYWupPC//Zz1G0EDHuFGPWKTGAtHEUZLW2Fy 3v4fmgvc2ba3Usxdf4CZjRo+kqepU2FyfthjVzh8xwyIbi026D1Signk4cST/7CbkfmxUI 4G4+Ec/ufy2updDebH445fX0FG0OzlE= Date: Thu, 5 Feb 2026 23:01:11 +0800 Precedence: bulk X-Mailing-List: virtualization@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Subject: Re: [PATCH v4 0/3] targeted TLB sync IPIs for lockless page table Content-Language: en-US To: "David Hildenbrand (Arm)" , Peter Zijlstra , dave.hansen@intel.com Cc: Liam.Howlett@oracle.com, akpm@linux-foundation.org, aneesh.kumar@kernel.org, arnd@arndb.de, baohua@kernel.org, baolin.wang@linux.alibaba.com, boris.ostrovsky@oracle.com, bp@alien8.de, dave.hansen@linux.intel.com, dev.jain@arm.com, hpa@zytor.com, hughd@google.com, ioworker0@gmail.com, jannh@google.com, jgross@suse.com, kvm@vger.kernel.org, linux-arch@vger.kernel.org, linux-kernel@vger.kernel.org, linux-mm@kvack.org, lorenzo.stoakes@oracle.com, mingo@redhat.com, npache@redhat.com, npiggin@gmail.com, pbonzini@redhat.com, riel@surriel.com, ryan.roberts@arm.com, seanjc@google.com, shy828301@gmail.com, tglx@linutronix.de, virtualization@lists.linux.dev, will@kernel.org, x86@kernel.org, ypodemsk@redhat.com, ziy@nvidia.com References: <20260202095414.GE2995752@noisy.programming.kicks-ass.net> <20260202110329.74397-1-lance.yang@linux.dev> <20260202125030.GB1395266@noisy.programming.kicks-ass.net> <4700e7ba-8456-4a93-9e28-7e5a3ca2a1be@linux.dev> <20260202133713.GF1395266@noisy.programming.kicks-ass.net> <540adec9-c483-460a-a682-f2076cf015c2@linux.dev> <20260202150957.GD1282955@noisy.programming.kicks-ass.net> <06d48a52-e4ec-47cd-b3fb-0fccd4dc49f4@kernel.org> X-Report-Abuse: Please report any abuse attempt to abuse@migadu.com and include these headers. From: Lance Yang In-Reply-To: <06d48a52-e4ec-47cd-b3fb-0fccd4dc49f4@kernel.org> Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 8bit X-Migadu-Flow: FLOW_OUT On 2026/2/5 21:25, David Hildenbrand (Arm) wrote: > On 2/2/26 16:52, Lance Yang wrote: >> >> >> On 2026/2/2 23:09, Peter Zijlstra wrote: >>> On Mon, Feb 02, 2026 at 10:37:39PM +0800, Lance Yang wrote: >>>> >>>> >>>> >>>> PT_RECLAIM=y does have IPI for unshare/collapse — those paths call >>>> tlb_flush_unshared_tables() (for hugetlb unshare) and >>>> collapse_huge_page() >>>> (in khugepaged collapse), which already send IPIs today (broadcast >>>> to all >>>> CPUs via tlb_remove_table_sync_one()). >>>> >>>> What PT_RECLAIM=y doesn't need IPI for is table freeing ( >>>> __tlb_remove_table_one() uses call_rcu() instead). But table >>>> modification >>>> (unshare, collapse) still needs IPI to synchronize with lockless >>>> walkers, >>>> regardless of PT_RECLAIM. >>>> >>>> So PT_RECLAIM=y is not broken; it already has IPI where needed. This >>>> series >>>> just makes those IPIs targeted instead of broadcast. Does that clarify? >>> >>> Oh bah, reading is hard. I had missed they had more table_sync_one() >>> calls, >>> rather than remove_table_one(). >>> >>> So you *can* replace table_sync_one() with rcu_sync(), that will provide >>> the same guarantees. Its just a 'little' bit slower on the update side, >>> but does not incur the read side cost. >> >> Yep, we could replace the IPI with synchronize_rcu() on the sync side: >> >> - Currently: TLB flush → send IPI → wait for walkers to finish >> - With synchronize_rcu(): TLB flush → synchronize_rcu() -> waits for >> grace period >> >> Lockless walkers (e.g. GUP-fast) use local_irq_disable(); >> synchronize_rcu() also >> waits for regions with preemption/interrupts disabled, so it should >> work, IIUC. >> >> And then, the trade-off would be: >> - Read side: zero cost (no per-CPU tracking) >> - Write side: wait for RCU grace period (potentially slower) >> >> For collapse/unshare, that write-side latency might be acceptable :) >> >> @David, what do you think? > > Given that we just fixed the write-side latency from breaking Oracle's > databases completely, we have to be a bit careful here :) Yep, agreed. > > The thing is: on many x86 configs we don't need *any* TLB flushed or RCU > syncs. Right. Looks like that is low-hanging fruit. I'll send that out separately :) > > So "how much slower" are we talking about, especially on bigger/loaded > systems? Unfortunately the numbers are pretry bad. On an x86-64 64-core system under high load, each synchronize_rcu() is about *22.9* ms on average ... So for now, neither approach looks good: tracking on the read side adss cost to GUP-fast, and syncing on the write side e.g. synchronize_rcu() is too slow on large systems.