* Re: [ath6kl:pending 14/24] drivers/mtd/devices/bcm47xxsflash.c:299:2: error: implicit declaration of function 'ioremap_cache'
[not found] <201607190916.c8xDOxpA%fengguang.wu@intel.com>
@ 2016-07-19 5:56 ` Rafał Miłecki
0 siblings, 0 replies; only message in thread
From: Rafał Miłecki @ 2016-07-19 5:56 UTC (permalink / raw)
To: kbuild test robot, Brian Norris
Cc: kbuild-all, ath10k, Kalle Valo, linux-mtd, linux-wireless
Resending reply with proper To/Cc.
On 07/19/2016 03:14 AM, kbuild test robot wrote:
> tree: https://git.kernel.org/pub/scm/linux/kernel/git/kvalo/ath.git pending
> head: 7a1b79bd39dda1f12b9f6708583250f40354c981
> commit: 57d8f7dd2132df3ac21044e93a8ecdc9744b4459 [14/24] bcma: allow enabling serial flash support on non-MIPS SoCs
> config: cris-allmodconfig (attached as .config)
> compiler: cris-linux-gcc (GCC) 4.6.3
> reproduce:
> wget https://git.kernel.org/cgit/linux/kernel/git/wfg/lkp-tests.git/plain/sbin/make.cross -O ~/bin/make.cross
> chmod +x ~/bin/make.cross
> git checkout 57d8f7dd2132df3ac21044e93a8ecdc9744b4459
> # save the attached .config to linux build tree
> make.cross ARCH=cris
>
> All error/warnings (new ones prefixed by >>):
>
> drivers/mtd/devices/bcm47xxsflash.c: In function 'bcm47xxsflash_bcma_probe':
>>> drivers/mtd/devices/bcm47xxsflash.c:299:2: error: implicit declaration of function 'ioremap_cache' [-Werror=implicit-function-declaration]
>>> drivers/mtd/devices/bcm47xxsflash.c:299:15: warning: assignment makes pointer from integer without a cast [enabled by default]
> cc1: some warnings being treated as errors
Oh, great :( With commit:
5651d6aaf489 ("mtd: bcm47xxsflash: use ioremap_cache() instead of KSEG0ADDR()")
we believed to make this driver compilable on all architectures. It seems that
ioremap_cache is available only on arm, arm64, ia64, mips, sh, x86 and xtensa.
I noticed that kernel/memremap.c implements ioremap_cache on its own for archs
that don't provide it. There is a comment in this file:
/* temporary while we convert existing ioremap_cache users to memremap */
Should we use memremap? I don't understand this function well, it seems I could
call it with with MEMREMAP_WB flag (and trigger ioremap_cache call), but is it a
correct thing to do?
Or should we simply make some symbol (CONFIG_MTD_BCM47XXSFLASH?) depend on
ARM || MIPS?
> vim +/ioremap_cache +299 drivers/mtd/devices/bcm47xxsflash.c
>
> 5fe42d5bf Rafał Miłecki 2012-09-17 283
> 5651d6aaf Brian Norris 2016-02-26 284 b47s = devm_kzalloc(dev, sizeof(*b47s), GFP_KERNEL);
> d2b1bd142 Libo Chen 2013-05-30 285 if (!b47s)
> d2b1bd142 Libo Chen 2013-05-30 286 return -ENOMEM;
> a2f74a7da Rafał Miłecki 2013-01-06 287 sflash->priv = b47s;
> a2f74a7da Rafał Miłecki 2013-01-06 288
> 5651d6aaf Brian Norris 2016-02-26 289 res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> 5651d6aaf Brian Norris 2016-02-26 290 if (!res) {
> 5651d6aaf Brian Norris 2016-02-26 291 dev_err(dev, "invalid resource\n");
> 5651d6aaf Brian Norris 2016-02-26 292 return -EINVAL;
> 5651d6aaf Brian Norris 2016-02-26 293 }
> 5651d6aaf Brian Norris 2016-02-26 294 if (!devm_request_mem_region(dev, res->start, resource_size(res),
> 5651d6aaf Brian Norris 2016-02-26 295 res->name)) {
> 5651d6aaf Brian Norris 2016-02-26 296 dev_err(dev, "can't request region for resource %pR\n", res);
> 5651d6aaf Brian Norris 2016-02-26 297 return -EBUSY;
> 5651d6aaf Brian Norris 2016-02-26 298 }
> 5651d6aaf Brian Norris 2016-02-26 @299 b47s->window = ioremap_cache(res->start, resource_size(res));
> 5651d6aaf Brian Norris 2016-02-26 300 if (!b47s->window) {
> 5651d6aaf Brian Norris 2016-02-26 301 dev_err(dev, "ioremap failed for resource %pR\n", res);
> 5651d6aaf Brian Norris 2016-02-26 302 return -ENOMEM;
> 5651d6aaf Brian Norris 2016-02-26 303 }
> 5651d6aaf Brian Norris 2016-02-26 304
> 41c81536e Rafał Miłecki 2013-03-06 305 b47s->bcma_cc = container_of(sflash, struct bcma_drv_cc, sflash);
> 265dfbd9a Rafał Miłecki 2013-03-24 306 b47s->cc_read = bcm47xxsflash_bcma_cc_read;
> 265dfbd9a Rafał Miłecki 2013-03-24 307 b47s->cc_write = bcm47xxsflash_bcma_cc_write;
>
> :::::: The code at line 299 was first introduced by commit
> :::::: 5651d6aaf489d1db48c253cf884b40214e91c2c5 mtd: bcm47xxsflash: use ioremap_cache() instead of KSEG0ADDR()
>
> :::::: TO: Brian Norris <computersforpeace@gmail.com>
> :::::: CC: Brian Norris <computersforpeace@gmail.com>
>
> ---
> 0-DAY kernel test infrastructure Open Source Technology Center
> https://lists.01.org/pipermail/kbuild-all Intel Corporation
>
^ permalink raw reply [flat|nested] only message in thread