linux-kernel - [PATCH 3/8] msm: scm: Flush the command buffer only instead of the entire cache

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Mon,  4 Aug 2014 18:31:45 -0700
From:	Stephen Boyd <sboyd@...eaurora.org>
To:	David Brown <davidb@...eaurora.org>,
	Kumar Gala <galak@...eaurora.org>
Cc:	Vikram Mulukutla <markivx@...eaurora.org>,
	linux-kernel@...r.kernel.org, linux-arm-msm@...r.kernel.org,
	linux-arm-kernel@...ts.infradead.org,
	Lina Iyer <lina.iyer@...aro.org>
Subject: [PATCH 3/8] msm: scm: Flush the command buffer only instead of the entire cache

From: Vikram Mulukutla <markivx@...eaurora.org>

scm_call flushes the entire cache before calling into the
secure world. This is both a performance penalty as well
as insufficient on SMP systems where the CPUs possess a
write-back L1 cache. Flush only the command and response
buffers instead, moving the responsibility of flushing any
other cached buffer (being passed to the secure world) to
callers.

Signed-off-by: Vikram Mulukutla <markivx@...eaurora.org>
Signed-off-by: Stephen Boyd <sboyd@...eaurora.org>
---
 arch/arm/mach-qcom/scm.c | 17 ++++++++++++-----
 1 file changed, 12 insertions(+), 5 deletions(-)

diff --git a/arch/arm/mach-qcom/scm.c b/arch/arm/mach-qcom/scm.c
index c08786ebf116..ec37b037e69c 100644
--- a/arch/arm/mach-qcom/scm.c
+++ b/arch/arm/mach-qcom/scm.c
@@ -196,12 +196,12 @@ static int __scm_call(const struct scm_command *cmd)
 	u32 cmd_addr = virt_to_phys(cmd);
 
 	/*
-	 * Flush the entire cache here so callers don't have to remember
-	 * to flush the cache when passing physical addresses to the secure
-	 * side in the buffer.
+	 * Flush the command buffer so that the secure world sees
+	 * the correct data.
 	 */
-	flush_cache_all();
-	outer_flush_all();
+	__cpuc_flush_dcache_area((void *)cmd, cmd->len);
+	outer_flush_range(cmd_addr, cmd_addr + cmd->len);
+
 	ret = smc(cmd_addr);
 	if (ret < 0)
 		ret = scm_remap_error(ret);
@@ -238,6 +238,13 @@ static void scm_inv_range(unsigned long start, unsigned long end)
  * @resp_len: length of the response buffer
  *
  * Sends a command to the SCM and waits for the command to finish processing.
+ *
+ * A note on cache maintenance:
+ * Note that any buffers that are expected to be accessed by the secure world
+ * must be flushed before invoking scm_call and invalidated in the cache
+ * immediately after scm_call returns. Cache maintenance on the command and
+ * response buffers is taken care of by scm_call; however, callers are
+ * responsible for any other cached buffers passed over to the secure world.
  */
 int scm_call(u32 svc_id, u32 cmd_id, const void *cmd_buf, size_t cmd_len,
 		void *resp_buf, size_t resp_len)
-- 
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
hosted by The Linux Foundation

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/