From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754712AbbGRIcL (ORCPT ); Sat, 18 Jul 2015 04:32:11 -0400 Received: from mail-lb0-f170.google.com ([209.85.217.170]:36325 "EHLO mail-lb0-f170.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753279AbbGRIbB (ORCPT ); Sat, 18 Jul 2015 04:31:01 -0400 From: Max Filippov To: linux-xtensa@linux-xtensa.org, linux-kernel@vger.kernel.org Cc: Chris Zankel , Marc Gauthier , Max Filippov , stable@vger.kernel.org Subject: [PATCH v2 10/13] xtensa: fix kernel register spilling Date: Sat, 18 Jul 2015 11:30:13 +0300 Message-Id: <1437208216-15729-11-git-send-email-jcmvbkbc@gmail.com> X-Mailer: git-send-email 1.8.1.4 In-Reply-To: <1437208216-15729-1-git-send-email-jcmvbkbc@gmail.com> References: <1437208216-15729-1-git-send-email-jcmvbkbc@gmail.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org call12 can't be safely used as the first call in the inline function, because the compiler does not extend the stack frame of the bounding function accordingly, which may result in corruption of local variables. If a call needs to be done, do call8 first followed by call12. For pure assembly code in _switch_to increase stack frame size of the bounding function. Cc: stable@vger.kernel.org Signed-off-by: Max Filippov --- arch/xtensa/include/asm/traps.h | 29 +++++++++++++++++++---------- arch/xtensa/kernel/entry.S | 4 ++-- 2 files changed, 21 insertions(+), 12 deletions(-) diff --git a/arch/xtensa/include/asm/traps.h b/arch/xtensa/include/asm/traps.h index 677bfcf..28f33a8 100644 --- a/arch/xtensa/include/asm/traps.h +++ b/arch/xtensa/include/asm/traps.h @@ -25,30 +25,39 @@ static inline void spill_registers(void) { #if XCHAL_NUM_AREGS > 16 __asm__ __volatile__ ( - " call12 1f\n" + " call8 1f\n" " _j 2f\n" " retw\n" " .align 4\n" "1:\n" +#if XCHAL_NUM_AREGS == 32 + " _entry a1, 32\n" + " addi a8, a0, 3\n" + " _entry a1, 16\n" + " mov a12, a12\n" + " retw\n" +#else " _entry a1, 48\n" - " addi a12, a0, 3\n" -#if XCHAL_NUM_AREGS > 32 - " .rept (" __stringify(XCHAL_NUM_AREGS) " - 32) / 12\n" + " call12 1f\n" + " retw\n" + " .align 4\n" + "1:\n" + " .rept (" __stringify(XCHAL_NUM_AREGS) " - 16) / 12\n" " _entry a1, 48\n" " mov a12, a0\n" " .endr\n" -#endif - " _entry a1, 48\n" + " _entry a1, 16\n" #if XCHAL_NUM_AREGS % 12 == 0 - " mov a8, a8\n" -#elif XCHAL_NUM_AREGS % 12 == 4 " mov a12, a12\n" -#elif XCHAL_NUM_AREGS % 12 == 8 +#elif XCHAL_NUM_AREGS % 12 == 4 " mov a4, a4\n" +#elif XCHAL_NUM_AREGS % 12 == 8 + " mov a8, a8\n" #endif " retw\n" +#endif "2:\n" - : : : "a12", "a13", "memory"); + : : : "a8", "a9", "memory"); #else __asm__ __volatile__ ( " mov a12, a12\n" diff --git a/arch/xtensa/kernel/entry.S b/arch/xtensa/kernel/entry.S index d060fc2..5771d5c 100644 --- a/arch/xtensa/kernel/entry.S +++ b/arch/xtensa/kernel/entry.S @@ -1838,7 +1838,7 @@ ENDPROC(system_call) mov a12, a0 .endr #endif - _entry a1, 48 + _entry a1, 16 #if XCHAL_NUM_AREGS % 12 == 0 mov a8, a8 #elif XCHAL_NUM_AREGS % 12 == 4 @@ -1862,7 +1862,7 @@ ENDPROC(system_call) ENTRY(_switch_to) - entry a1, 16 + entry a1, 48 mov a11, a3 # and 'next' (a3) -- 1.8.1.4