From: Christoph Hellwig Date: Fri, 19 Jul 2019 06:00:53 +0000 Subject: Re: [PATCH 09/16] sparc64: use the generic get_user_pages_fast code Message-Id: <20190719060053.GA18491@lst.de> List-Id: References: <20190625143715.1689-1-hch@lst.de> <20190625143715.1689-10-hch@lst.de> <20190717215956.GA30369@altlinux.org> <20190718.141405.1070121094691581998.davem@davemloft.net> In-Reply-To: <20190718.141405.1070121094691581998.davem@davemloft.net> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: David Miller Cc: ldv@altlinux.org, hch@lst.de, khalid.aziz@oracle.com, torvalds@linux-foundation.org, akpm@linux-foundation.org, matorola@gmail.com, sparclinux@vger.kernel.org, linux-mm@kvack.org, linux-kernel@vger.kernel.org On Thu, Jul 18, 2019 at 02:14:05PM -0700, David Miller wrote: > From: "Dmitry V. Levin" > Date: Thu, 18 Jul 2019 00:59:56 +0300 > > > So this ended up as commit 7b9afb86b6328f10dc2cad9223d7def12d60e505 > > (thanks to Anatoly for bisecting) and introduced a regression: > > futex.test from the strace test suite now causes an Oops on sparc64 > > in futex syscall. > > > > Here is a heavily stripped down reproducer: > > Does not reproduce for me on a T4-2 machine. > > So this problem might depend on the type of system you are on, > I suspect it's one of those "pre-Niagara vs. Niagara and later" > situations because that's the dividing line between two set of > wildly different TLB and cache management methods. > > What kind of machine are you on? FYI, I'm pretty sure I tested this on Guenthers build test qemu setup in the end, which further speaks for the different machines issue.