linux-kernel - [PATCH] ipmi: ssif: Add msleep in multipart test

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-next>] [day] [month] [year] [list]

Message-Id: <20211026025834.82766-1-likunkun@bytedance.com>
Date:   Tue, 26 Oct 2021 10:58:34 +0800
From:   Kunkun Li <likunkun@...edance.com>
To:     minyard@....org, openipmi-developer@...ts.sourceforge.net
Cc:     linux-kernel@...r.kernel.org, Kunkun Li <likunkun@...edance.com>
Subject: [PATCH] ipmi: ssif: Add msleep in multipart test

During multipart test, cmd(6,7,8) or cmd(6,7,7) will
be sent continuously.

The pressure test found some BMC systems cannot process
messages in time, resulting in read_response continues to receive
error messages from i2c.
Retry mechanism will takes 10s, and finally set not support
multipart transmit.

So, to work around this，add msleep after sending cmd 6 and
cmd 7 respectively. The problem did not appear again in
pressure test.

Signed-off-by: Kunkun Li <likunkun@...edance.com>
---
 drivers/char/ipmi/ipmi_ssif.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/char/ipmi/ipmi_ssif.c b/drivers/char/ipmi/ipmi_ssif.c
index 20d5af92966d..65841798fafe 100644
--- a/drivers/char/ipmi/ipmi_ssif.c
+++ b/drivers/char/ipmi/ipmi_ssif.c
@@ -1453,6 +1453,7 @@ static int start_multipart_test(struct i2c_client *client,
 	ret = i2c_smbus_write_block_data(client,
 					 SSIF_IPMI_MULTI_PART_REQUEST_START,
 					 32, msg);
+	msleep(SSIF_MSG_MSEC);
 	if (ret) {
 		retry_cnt--;
 		if (retry_cnt > 0)
@@ -1467,6 +1468,7 @@ static int start_multipart_test(struct i2c_client *client,
 	ret = i2c_smbus_write_block_data(client,
 					 SSIF_IPMI_MULTI_PART_REQUEST_MIDDLE,
 					 32, msg + 32);
+	msleep(SSIF_MSG_MSEC);
 	if (ret) {
 		dev_err(&client->dev, "Could not write multi-part middle, though the BMC said it could handle it.  Just limit sends to one part.\n");
 		return ret;
-- 
2.11.0