linux-kernel - [PATCH V2 RFC] mtd: spi-nor: intel: provide a range for poll

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: <1486973622-1072-1-git-send-email-der.herr@hofr.at>
Date:   Mon, 13 Feb 2017 09:13:42 +0100
From:   Nicholas Mc Guire <der.herr@...r.at>
To:     Cyrille Pitchen <cyrille.pitchen@...el.com>
Cc:     Marek Vasut <marek.vasut@...il.com>,
        David Woodhouse <dwmw2@...radead.org>,
        Brian Norris <computersforpeace@...il.com>,
        Boris Brezillon <boris.brezillon@...e-electrons.com>,
        Richard Weinberger <richard@....at>,
        Mika Westerberg <mika.westerberg@...ux.intel.com>,
        linux-mtd@...ts.infradead.org, linux-kernel@...r.kernel.org,
        Nicholas Mc Guire <der.herr@...r.at>
Subject: [PATCH V2 RFC] mtd: spi-nor: intel: provide a range for poll_timout

The overall poll time here is INTEL_SPI_TIMEOUT * 1000 which is
5000 * 1000 - so 5seconds and it is coded as a tight loop here delay_us
to readl_poll_timeout() is set to 0. As this is never called in an atomic
context sleeping should be no issue and there is no reasons for the
tight-loop here.

Signed-off-by: Nicholas Mc Guire <der.herr@...r.at>
---

V2: fixed prefix as pointed out by Boris Brezillon
    <boris.brezillon@...e-electrons.com>

Problem located by experimental coccinelle script:
./drivers/mtd/spi-nor/intel-spi.c:265:8-26: WARNING: usleep_range min=0 for delay INTEL_SPI_TIMEOUT * 1000
./drivers/mtd/spi-nor/intel-spi.c:274:8-26: WARNING: usleep_range min=0 for delay INTEL_SPI_TIMEOUT * 1000

The rational for setting the delay_us here to 40 is that readx_poll_timeout()
will take delay_us >> 2 + 1 as min value and that should be at least 10us (see
Documentation/timers/timers-howto.txt). Ideally the delay would be made
even larger to keep the load on the hrtimer subsystem low as these delays
here do not seem to be critical. Someone that knows the details of this device
would need to check if a larger delay would be ok here.

Patch was compile tested with: multi_v7_defconfig (implies CONFIG_MTD_SPI_NOR=y)
one coccicheck finding reported and one spars finding (in separate patches)

Patch is against 4.10-rc7 (localversion-next is next-20170213)

 drivers/mtd/spi-nor/intel-spi.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/mtd/spi-nor/intel-spi.c b/drivers/mtd/spi-nor/intel-spi.c
index a10f602..4630716 100644
--- a/drivers/mtd/spi-nor/intel-spi.c
+++ b/drivers/mtd/spi-nor/intel-spi.c
@@ -263,7 +263,7 @@ static int intel_spi_wait_hw_busy(struct intel_spi *ispi)
 	u32 val;
 
 	return readl_poll_timeout(ispi->base + HSFSTS_CTL, val,
-				  !(val & HSFSTS_CTL_SCIP), 0,
+				  !(val & HSFSTS_CTL_SCIP), 40,
 				  INTEL_SPI_TIMEOUT * 1000);
 }
 
@@ -272,7 +272,7 @@ static int intel_spi_wait_sw_busy(struct intel_spi *ispi)
 	u32 val;
 
 	return readl_poll_timeout(ispi->sregs + SSFSTS_CTL, val,
-				  !(val & SSFSTS_CTL_SCIP), 0,
+				  !(val & SSFSTS_CTL_SCIP), 40,
 				  INTEL_SPI_TIMEOUT * 1000);
 }
 
-- 
2.1.4