[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20210430195232.30491-7-michel@lespinasse.org>
Date: Fri, 30 Apr 2021 12:52:07 -0700
From: Michel Lespinasse <michel@...pinasse.org>
To: Linux-MM <linux-mm@...ck.org>,
Linux-Kernel <linux-kernel@...r.kernel.org>
Cc: Laurent Dufour <ldufour@...ux.ibm.com>,
Peter Zijlstra <peterz@...radead.org>,
Michal Hocko <mhocko@...e.com>,
Matthew Wilcox <willy@...radead.org>,
Rik van Riel <riel@...riel.com>,
Paul McKenney <paulmck@...nel.org>,
Andrew Morton <akpm@...ux-foundation.org>,
Suren Baghdasaryan <surenb@...gle.com>,
Joel Fernandes <joelaf@...gle.com>,
Andy Lutomirski <luto@...nel.org>,
Michel Lespinasse <michel@...pinasse.org>
Subject: [PATCH 06/29] mm: introduce CONFIG_SPECULATIVE_PAGE_FAULT
This configuration variable will be used to build the code needed to
handle speculative page fault.
This is enabled by default on supported architectures with SMP and MMU set.
The architecture support is needed since the speculative page fault handler
is called from the architecture's page faulting code, and some code has to
be added there to try speculative fault handling first.
Signed-off-by: Michel Lespinasse <michel@...pinasse.org>
---
mm/Kconfig | 22 ++++++++++++++++++++++
1 file changed, 22 insertions(+)
diff --git a/mm/Kconfig b/mm/Kconfig
index 24c045b24b95..322bda319dea 100644
--- a/mm/Kconfig
+++ b/mm/Kconfig
@@ -872,4 +872,26 @@ config MAPPING_DIRTY_HELPERS
config KMAP_LOCAL
bool
+config ARCH_SUPPORTS_SPECULATIVE_PAGE_FAULT
+ def_bool n
+
+config SPECULATIVE_PAGE_FAULT
+ bool "Speculative page faults"
+ default y
+ depends on ARCH_SUPPORTS_SPECULATIVE_PAGE_FAULT && MMU && SMP
+ help
+ Try to handle user space page faults without holding the mmap lock.
+
+ Instead of blocking writers through the use of mmap lock,
+ the page fault handler merely verifies, at the end of the page
+ fault, that no writers have been running concurrently with it.
+
+ In high concurrency situations, the speculative fault handler
+ gains a throughput advantage by avoiding having to update the
+ mmap lock reader count.
+
+ If the check fails due to a concurrent writer, or due to hitting
+ an unsupported case, the fault handler falls back to classical
+ processing using the mmap read lock.
+
endmenu
--
2.20.1
Powered by blists - more mailing lists