[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <201804040917.NjWkgZGY%fengguang.wu@intel.com>
Date: Wed, 4 Apr 2018 09:32:53 +0800
From: kbuild test robot <lkp@...el.com>
To: Sergey Suloev <ssuloev@...altech.com>
Cc: kbuild-all@...org, Mark Brown <broonie@...nel.org>,
Maxime Ripard <maxime.ripard@...tlin.com>,
Chen-Yu Tsai <wens@...e.org>, linux-spi@...r.kernel.org,
linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
Sergey Suloev <ssuloev@...altech.com>
Subject: Re: [PATCH v3 5/6] spi: sun6i: introduce register set/unset helpers
Hi Sergey,
Thank you for the patch! Yet something to improve:
[auto build test ERROR on spi/for-next]
[also build test ERROR on v4.16 next-20180403]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]
url: https://github.com/0day-ci/linux/commits/Sergey-Suloev/spi-Add-support-for-DMA-transfers-in-sun6i-SPI-driver/20180404-053231
base: https://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi.git for-next
config: alpha-allyesconfig (attached as .config)
compiler: alpha-linux-gnu-gcc (Debian 7.2.0-11) 7.2.0
reproduce:
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# save the attached .config to linux build tree
make.cross ARCH=alpha
Note: the linux-review/Sergey-Suloev/spi-Add-support-for-DMA-transfers-in-sun6i-SPI-driver/20180404-053231 HEAD a0c010a285d830f07bb81ea59eaea8773d78b74c builds fine.
It only hurts bisectibility.
All errors (new ones prefixed by >>):
drivers//spi/spi-sun6i.c: In function 'sun6i_spi_transfer_one':
>> drivers//spi/spi-sun6i.c:371:2: error: implicit declaration of function 'sun6i_spi_enable_interrupt'; did you mean 'sun6i_spi_transfer_one'? [-Werror=implicit-function-declaration]
sun6i_spi_enable_interrupt(sspi, SUN6I_INT_CTL_TC |
^~~~~~~~~~~~~~~~~~~~~~~~~~
sun6i_spi_transfer_one
drivers//spi/spi-sun6i.c:277:27: warning: unused variable 'tx_time' [-Wunused-variable]
unsigned int start, end, tx_time;
^~~~~~~
drivers//spi/spi-sun6i.c:277:22: warning: unused variable 'end' [-Wunused-variable]
unsigned int start, end, tx_time;
^~~
drivers//spi/spi-sun6i.c:277:15: warning: unused variable 'start' [-Wunused-variable]
unsigned int start, end, tx_time;
^~~~~
drivers//spi/spi-sun6i.c:276:31: warning: unused variable 'timeout' [-Wunused-variable]
unsigned int mclk_rate, div, timeout;
^~~~~~~
cc1: some warnings being treated as errors
vim +371 drivers//spi/spi-sun6i.c
43836daab Sergey Suloev 2018-04-03 270
3558fe900 Maxime Ripard 2014-02-05 271 static int sun6i_spi_transfer_one(struct spi_master *master,
3558fe900 Maxime Ripard 2014-02-05 272 struct spi_device *spi,
3558fe900 Maxime Ripard 2014-02-05 273 struct spi_transfer *tfr)
3558fe900 Maxime Ripard 2014-02-05 274 {
3558fe900 Maxime Ripard 2014-02-05 275 struct sun6i_spi *sspi = spi_master_get_devdata(master);
3558fe900 Maxime Ripard 2014-02-05 276 unsigned int mclk_rate, div, timeout;
719bd6542 Michal Suchanek 2016-06-13 277 unsigned int start, end, tx_time;
913f536c6 Icenowy Zheng 2017-03-06 278 unsigned int trig_level;
3558fe900 Maxime Ripard 2014-02-05 279 unsigned int tx_len = 0;
3558fe900 Maxime Ripard 2014-02-05 280 int ret = 0;
3558fe900 Maxime Ripard 2014-02-05 281 u32 reg;
3558fe900 Maxime Ripard 2014-02-05 282
e31cf0250 Sergey Suloev 2018-04-03 283 /* A zero length transfer never finishes if programmed
e31cf0250 Sergey Suloev 2018-04-03 284 in the hardware */
e31cf0250 Sergey Suloev 2018-04-03 285 if (!tfr->len)
e31cf0250 Sergey Suloev 2018-04-03 286 return 0;
e31cf0250 Sergey Suloev 2018-04-03 287
e31cf0250 Sergey Suloev 2018-04-03 288 /* Don't support transfer larger than the FIFO */
e31cf0250 Sergey Suloev 2018-04-03 289 if (tfr->len > sspi->fifo_depth)
e31cf0250 Sergey Suloev 2018-04-03 290 return -EMSGSIZE;
3558fe900 Maxime Ripard 2014-02-05 291
3558fe900 Maxime Ripard 2014-02-05 292 sspi->tx_buf = tfr->tx_buf;
3558fe900 Maxime Ripard 2014-02-05 293 sspi->rx_buf = tfr->rx_buf;
3558fe900 Maxime Ripard 2014-02-05 294 sspi->len = tfr->len;
3558fe900 Maxime Ripard 2014-02-05 295
3558fe900 Maxime Ripard 2014-02-05 296 /* Clear pending interrupts */
3558fe900 Maxime Ripard 2014-02-05 297 sun6i_spi_write(sspi, SUN6I_INT_STA_REG, ~0);
3558fe900 Maxime Ripard 2014-02-05 298
3558fe900 Maxime Ripard 2014-02-05 299 /* Reset FIFO */
3558fe900 Maxime Ripard 2014-02-05 300 sun6i_spi_write(sspi, SUN6I_FIFO_CTL_REG,
3558fe900 Maxime Ripard 2014-02-05 301 SUN6I_FIFO_CTL_RF_RST | SUN6I_FIFO_CTL_TF_RST);
3558fe900 Maxime Ripard 2014-02-05 302
3558fe900 Maxime Ripard 2014-02-05 303 /*
913f536c6 Icenowy Zheng 2017-03-06 304 * Setup FIFO interrupt trigger level
913f536c6 Icenowy Zheng 2017-03-06 305 * Here we choose 3/4 of the full fifo depth, as it's the hardcoded
913f536c6 Icenowy Zheng 2017-03-06 306 * value used in old generation of Allwinner SPI controller.
913f536c6 Icenowy Zheng 2017-03-06 307 * (See spi-sun4i.c)
913f536c6 Icenowy Zheng 2017-03-06 308 */
913f536c6 Icenowy Zheng 2017-03-06 309 trig_level = sspi->fifo_depth / 4 * 3;
913f536c6 Icenowy Zheng 2017-03-06 310 sun6i_spi_write(sspi, SUN6I_FIFO_CTL_REG,
e31cf0250 Sergey Suloev 2018-04-03 311 (trig_level << SUN6I_FIFO_CTL_RF_RDY_TRIG_LEVEL_BITS));
913f536c6 Icenowy Zheng 2017-03-06 312
3558fe900 Maxime Ripard 2014-02-05 313 /*
3558fe900 Maxime Ripard 2014-02-05 314 * If it's a TX only transfer, we don't want to fill the RX
3558fe900 Maxime Ripard 2014-02-05 315 * FIFO with bogus data
3558fe900 Maxime Ripard 2014-02-05 316 */
3558fe900 Maxime Ripard 2014-02-05 317 if (sspi->rx_buf)
2c98d976c Sergey Suloev 2018-04-03 318 sun6i_spi_unset(sspi, SUN6I_TFR_CTL_REG, SUN6I_TFR_CTL_DHB);
3558fe900 Maxime Ripard 2014-02-05 319 else
2c98d976c Sergey Suloev 2018-04-03 320 sun6i_spi_set(sspi, SUN6I_TFR_CTL_REG, SUN6I_TFR_CTL_DHB);
3558fe900 Maxime Ripard 2014-02-05 321
09d186f3c Sergey Suloev 2018-04-03 322
3558fe900 Maxime Ripard 2014-02-05 323 /* Ensure that we have a parent clock fast enough */
3558fe900 Maxime Ripard 2014-02-05 324 mclk_rate = clk_get_rate(sspi->mclk);
47284e3e0 Marcus Weseloh 2015-11-08 325 if (mclk_rate < (2 * tfr->speed_hz)) {
47284e3e0 Marcus Weseloh 2015-11-08 326 clk_set_rate(sspi->mclk, 2 * tfr->speed_hz);
3558fe900 Maxime Ripard 2014-02-05 327 mclk_rate = clk_get_rate(sspi->mclk);
3558fe900 Maxime Ripard 2014-02-05 328 }
3558fe900 Maxime Ripard 2014-02-05 329
3558fe900 Maxime Ripard 2014-02-05 330 /*
3558fe900 Maxime Ripard 2014-02-05 331 * Setup clock divider.
3558fe900 Maxime Ripard 2014-02-05 332 *
3558fe900 Maxime Ripard 2014-02-05 333 * We have two choices there. Either we can use the clock
3558fe900 Maxime Ripard 2014-02-05 334 * divide rate 1, which is calculated thanks to this formula:
3558fe900 Maxime Ripard 2014-02-05 335 * SPI_CLK = MOD_CLK / (2 ^ cdr)
3558fe900 Maxime Ripard 2014-02-05 336 * Or we can use CDR2, which is calculated with the formula:
3558fe900 Maxime Ripard 2014-02-05 337 * SPI_CLK = MOD_CLK / (2 * (cdr + 1))
3558fe900 Maxime Ripard 2014-02-05 338 * Wether we use the former or the latter is set through the
3558fe900 Maxime Ripard 2014-02-05 339 * DRS bit.
3558fe900 Maxime Ripard 2014-02-05 340 *
3558fe900 Maxime Ripard 2014-02-05 341 * First try CDR2, and if we can't reach the expected
3558fe900 Maxime Ripard 2014-02-05 342 * frequency, fall back to CDR1.
3558fe900 Maxime Ripard 2014-02-05 343 */
47284e3e0 Marcus Weseloh 2015-11-08 344 div = mclk_rate / (2 * tfr->speed_hz);
3558fe900 Maxime Ripard 2014-02-05 345 if (div <= (SUN6I_CLK_CTL_CDR2_MASK + 1)) {
3558fe900 Maxime Ripard 2014-02-05 346 if (div > 0)
3558fe900 Maxime Ripard 2014-02-05 347 div--;
3558fe900 Maxime Ripard 2014-02-05 348
3558fe900 Maxime Ripard 2014-02-05 349 reg = SUN6I_CLK_CTL_CDR2(div) | SUN6I_CLK_CTL_DRS;
3558fe900 Maxime Ripard 2014-02-05 350 } else {
47284e3e0 Marcus Weseloh 2015-11-08 351 div = ilog2(mclk_rate) - ilog2(tfr->speed_hz);
3558fe900 Maxime Ripard 2014-02-05 352 reg = SUN6I_CLK_CTL_CDR1(div);
3558fe900 Maxime Ripard 2014-02-05 353 }
3558fe900 Maxime Ripard 2014-02-05 354
3558fe900 Maxime Ripard 2014-02-05 355 sun6i_spi_write(sspi, SUN6I_CLK_CTL_REG, reg);
3558fe900 Maxime Ripard 2014-02-05 356
3558fe900 Maxime Ripard 2014-02-05 357 /* Setup the transfer now... */
3558fe900 Maxime Ripard 2014-02-05 358 if (sspi->tx_buf)
3558fe900 Maxime Ripard 2014-02-05 359 tx_len = tfr->len;
3558fe900 Maxime Ripard 2014-02-05 360
3558fe900 Maxime Ripard 2014-02-05 361 /* Setup the counters */
3558fe900 Maxime Ripard 2014-02-05 362 sun6i_spi_write(sspi, SUN6I_BURST_CNT_REG, SUN6I_BURST_CNT(tfr->len));
3558fe900 Maxime Ripard 2014-02-05 363 sun6i_spi_write(sspi, SUN6I_XMIT_CNT_REG, SUN6I_XMIT_CNT(tx_len));
3558fe900 Maxime Ripard 2014-02-05 364 sun6i_spi_write(sspi, SUN6I_BURST_CTL_CNT_REG,
3558fe900 Maxime Ripard 2014-02-05 365 SUN6I_BURST_CTL_CNT_STC(tx_len));
3558fe900 Maxime Ripard 2014-02-05 366
3558fe900 Maxime Ripard 2014-02-05 367 /* Fill the TX FIFO */
10565dfd3 Milo Kim 2016-10-28 368 sun6i_spi_fill_fifo(sspi, sspi->fifo_depth);
3558fe900 Maxime Ripard 2014-02-05 369
3558fe900 Maxime Ripard 2014-02-05 370 /* Enable the interrupts */
913f536c6 Icenowy Zheng 2017-03-06 @371 sun6i_spi_enable_interrupt(sspi, SUN6I_INT_CTL_TC |
913f536c6 Icenowy Zheng 2017-03-06 372 SUN6I_INT_CTL_RF_RDY);
3558fe900 Maxime Ripard 2014-02-05 373
3558fe900 Maxime Ripard 2014-02-05 374 /* Start the transfer */
2c98d976c Sergey Suloev 2018-04-03 375 sun6i_spi_set(sspi, SUN6I_TFR_CTL_REG, SUN6I_TFR_CTL_XCH);
3558fe900 Maxime Ripard 2014-02-05 376
43836daab Sergey Suloev 2018-04-03 377 /* Wait for completion */
43836daab Sergey Suloev 2018-04-03 378 ret = sun6i_spi_wait_for_transfer(spi, tfr);
3558fe900 Maxime Ripard 2014-02-05 379
3558fe900 Maxime Ripard 2014-02-05 380 sun6i_spi_write(sspi, SUN6I_INT_CTL_REG, 0);
3558fe900 Maxime Ripard 2014-02-05 381
3558fe900 Maxime Ripard 2014-02-05 382 return ret;
3558fe900 Maxime Ripard 2014-02-05 383 }
3558fe900 Maxime Ripard 2014-02-05 384
:::::: The code at line 371 was first introduced by commit
:::::: 913f536c6c18a2e19e32f06971101c1d0ae3739c spi: sun6i: Allow transfers larger than FIFO size
:::::: TO: Icenowy Zheng <icenowy@...c.xyz>
:::::: CC: Mark Brown <broonie@...nel.org>
---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all Intel Corporation
Download attachment ".config.gz" of type "application/gzip" (52801 bytes)
Powered by blists - more mailing lists