linux-kernel - [PATCH] mmc: sdio: Fix sdio wait busy implement limitation

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]

Message-ID: <1492416571-28843-2-git-send-email-jiajie.hao@mediatek.com>
Date:   Mon, 17 Apr 2017 16:09:31 +0800
From:   <jiajie.hao@...iatek.com>
To:     <ulf.hansson@...aro.org>
CC:     <linus.walleij@...aro.org>, <chaotian.jing@...iatek.com>,
        <yong.mao@...iatek.com>, <eddie.huang@...iatek.com>,
        <srv_heupstream@...iatek.com>, <linux-mmc@...r.kernel.org>,
        <linux-kernel@...r.kernel.org>,
        <linux-mediatek@...ts.infradead.org>,
        <linux-arm-kernel@...iatek.com>,
        Jiajie Hao <jiajie.hao@...iatek.com>
Subject: [PATCH] mmc: sdio: Fix sdio wait busy implement limitation

From: Jiajie Hao <jiajie.hao@...iatek.com>

The host may issue an I/O abort by writing to the CCCR at any time
during I/O read operation via CMD52. And host may need suspend
transcation during write busy stage in SDIO suspend/resume scenario.
>From other side, a card may accept CMD52 during data transfer phase.

Previous implement would block issuing above command in busy stage.
It cause function driver can't implement as proper way and has no
opportunity to do some coverage in error case via I/O abort etc.

We need bypass some necessary operation during busy check stage.

Change-Id: Ib5a4599a2a98cbcf441b564f5f6fbb10c8a59e1b
Signed-off-by: Jiajie Hao <jiajie.hao@...iatek.com>
---
 drivers/mmc/core/core.c     |    4 +++-
 drivers/mmc/core/sdio_ops.h |   10 ++++++++++
 2 files changed, 13 insertions(+), 1 deletion(-)

diff --git a/drivers/mmc/core/core.c b/drivers/mmc/core/core.c
index 926e0fd..9c9a4bc 100644
--- a/drivers/mmc/core/core.c
+++ b/drivers/mmc/core/core.c
@@ -234,8 +234,10 @@ static void __mmc_start_request(struct mmc_host *host, struct mmc_request *mrq)
 	/*
 	 * For sdio rw commands we must wait for card busy otherwise some
 	 * sdio devices won't work properly.
+	 * And bypass I/O abort, reset and bus suspend operations.
 	 */
-	if (mmc_is_io_op(mrq->cmd->opcode) && host->ops->card_busy) {
+	if (mmc_is_io_op(mrq->cmd->opcode) && host->ops->card_busy &&
+	    !mmc_is_io_bypass(mrq->cmd->opcode, mrq->cmd->arg)) {
 		int tries = 500; /* Wait aprox 500ms at maximum */
 
 		while (host->ops->card_busy(host) && --tries)
diff --git a/drivers/mmc/core/sdio_ops.h b/drivers/mmc/core/sdio_ops.h
index bed8a83..9896696 100644
--- a/drivers/mmc/core/sdio_ops.h
+++ b/drivers/mmc/core/sdio_ops.h
@@ -31,5 +31,15 @@ static inline bool mmc_is_io_op(u32 opcode)
 	return opcode == SD_IO_RW_DIRECT || opcode == SD_IO_RW_EXTENDED;
 }
 
+static inline bool mmc_is_io_bypass(u32 opcode, u32 arg)
+{
+	u32 addr;
+
+	addr = (arg >> 9) & 0x1FFFF;
+
+	return ((opcode == SD_IO_RW_DIRECT) &&
+		(addr == SDIO_CCCR_ABORT || addr == SDIO_CCCR_SUSPEND));
+}
+
 #endif
 
-- 
1.7.9.5