From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:50265) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1bLSGv-00036o-Pb for qemu-devel@nongnu.org; Fri, 08 Jul 2016 05:40:34 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1bLSGs-0003Zo-E4 for qemu-devel@nongnu.org; Fri, 08 Jul 2016 05:40:33 -0400 Received: from mail-lf0-x241.google.com ([2a00:1450:4010:c07::241]:36127) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1bLSGs-0003ZB-59 for qemu-devel@nongnu.org; Fri, 08 Jul 2016 05:40:30 -0400 Received: by mail-lf0-x241.google.com with SMTP id a2so5058707lfe.3 for ; Fri, 08 Jul 2016 02:40:30 -0700 (PDT) References: <1467735496-16256-1-git-send-email-alex.bennee@linaro.org> <1467735496-16256-3-git-send-email-alex.bennee@linaro.org> <577E6252.5060204@linaro.org> From: Sergey Fedorov Message-ID: <577F750B.5040902@gmail.com> Date: Fri, 8 Jul 2016 12:40:27 +0300 MIME-Version: 1.0 In-Reply-To: <577E6252.5060204@linaro.org> Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 8bit Subject: Re: [Qemu-devel] [PATCH v2 2/6] tcg: set up tb->page_addr before insertion List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Sergey Fedorov , =?UTF-8?Q?Alex_Benn=c3=a9e?= , mttcg@listserver.greensocs.com, qemu-devel@nongnu.org, fred.konrad@greensocs.com, a.rigo@virtualopensystems.com, cota@braap.org, bobby.prani@gmail.com, rth@twiddle.net Cc: mark.burton@greensocs.com, pbonzini@redhat.com, jan.kiszka@siemens.com, peter.maydell@linaro.org, claudio.fontana@huawei.com, Peter Crosthwaite On 07/07/16 17:08, Sergey Fedorov wrote: > On 05/07/16 19:18, Alex Bennée wrote: >> This ensures that if we find the TB on the slow path that tb->page_addr >> is correctly set before being tested. >> >> Signed-off-by: Alex Bennée > Reveiwed-by: Sergey Fedorov However, we may need to use smp_wmb() here paired with smp_rmb() in tb_find_fast(). Alternatively, QHT might provide some explicit guarantee about memory ordering around qht_insert() and qht_lookup(). Actually, there is already right ordering thanks to seqlock operations in QHT implementation. This means we could elide explicit memory barrier from the first patch as well. Kind regards, Sergey > >> --- >> translate-all.c | 8 ++++---- >> 1 file changed, 4 insertions(+), 4 deletions(-) >> >> diff --git a/translate-all.c b/translate-all.c >> index 96efe48..97e834a 100644 >> --- a/translate-all.c >> +++ b/translate-all.c >> @@ -1126,10 +1126,6 @@ static void tb_link_page(TranslationBlock *tb, tb_page_addr_t phys_pc, >> { >> uint32_t h; >> >> - /* add in the hash table */ >> - h = tb_hash_func(phys_pc, tb->pc, tb->flags); >> - qht_insert(&tcg_ctx.tb_ctx.htable, tb, h); >> - >> /* add in the page list */ >> tb_alloc_page(tb, 0, phys_pc & TARGET_PAGE_MASK); >> if (phys_page2 != -1) { >> @@ -1138,6 +1134,10 @@ static void tb_link_page(TranslationBlock *tb, tb_page_addr_t phys_pc, >> tb->page_addr[1] = -1; >> } >> >> + /* add in the hash table */ >> + h = tb_hash_func(phys_pc, tb->pc, tb->flags); >> + qht_insert(&tcg_ctx.tb_ctx.htable, tb, h); >> + >> #ifdef DEBUG_TB_CHECK >> tb_page_check(); >> #endif