From mboxrd@z Thu Jan 1 00:00:00 1970 From: Dan Carpenter Date: Wed, 27 Jun 2012 09:02:58 +0000 Subject: [patch -resend] mmc: ushc: fix an endianness conversion in ushc_request() Message-Id: <20120627090258.GH31212@elgon.mountain> List-Id: References: <20120627085800.GA3007@mwanda> In-Reply-To: <20120627085800.GA3007@mwanda> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Chris Ball Cc: Greg Kroah-Hartman , linux-mmc@vger.kernel.org, David Vrabel , linux-kernel@vger.kernel.org, kernel-janitors@vger.kernel.org The ->cmd_idx field is 8 bits, not 16 so the call to cpu_to_le16() will probably set cmd_idx to zero here on big endian systems. My guess is that this works because it has been tested on little endian systems where the conversion to le16 is a nop. Signed-off-by: Dan Carpenter --- I don't have a way to test this. I have added a sentence to the changelog. I have added David Vrabel to the CC list. This was originally sent to the mailing list on Mon, 21 Nov 2011 and there was no response. diff --git a/drivers/mmc/host/ushc.c b/drivers/mmc/host/ushc.c index c0105a2..25932f5 100644 --- a/drivers/mmc/host/ushc.c +++ b/drivers/mmc/host/ushc.c @@ -278,7 +278,7 @@ static void ushc_request(struct mmc_host *mmc, struct mmc_request *req) ushc->current_req = req; /* Start cmd with CBW. */ - ushc->cbw->cmd_idx = cpu_to_le16(req->cmd->opcode); + ushc->cbw->cmd_idx = req->cmd->opcode; if (req->data) ushc->cbw->block_size = cpu_to_le16(req->data->blksz); else