[PATCH v4 1/3] resource: Use list_head to link sibling resource

Baoquan He bhe at redhat.com
Tue May 8 22:09:13 EDT 2018


On 05/07/18 at 07:42pm, kbuild test robot wrote:
> Hi Baoquan,
> 
> I love your patch! Yet something to improve:
> 
> [auto build test ERROR on linus/master]
> [also build test ERROR on v4.17-rc4 next-20180504]
> [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/Baoquan-He/resource-Use-list_head-to-link-sibling-resource/20180507-144345
> config: powerpc-defconfig (attached as .config)
> compiler: powerpc64-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=powerpc 
> 
> All errors (new ones prefixed by >>):
> 
>    arch/powerpc/kernel/pci-common.c: In function 'pci_process_bridge_OF_ranges':
> >> arch/powerpc/kernel/pci-common.c:764:44: error: incompatible types when assigning to type 'struct list_head' from type 'void *'
>        res->parent = res->child = res->sibling = NULL;

Pasted code can fix above error.

diff --git a/arch/powerpc/kernel/pci-common.c b/arch/powerpc/kernel/pci-common.c
index fe9733ffffaa..a7e68f6f9f24 100644
--- a/arch/powerpc/kernel/pci-common.c
+++ b/arch/powerpc/kernel/pci-common.c
@@ -761,7 +761,9 @@ void pci_process_bridge_OF_ranges(struct pci_controller *hose,
 			res->flags = range.flags;
 			res->start = range.cpu_addr;
 			res->end = range.cpu_addr + range.size - 1;
-			res->parent = res->child = res->sibling = NULL;
+			res->parent = NULL;
+			INIT_LIST_HEAD(&res->child);
+			INIT_LIST_HEAD(&res->sibling);
 		}
 	}
 }

>                                                ^
>    arch/powerpc/kernel/pci-common.c: In function 'reparent_resources':
> >> arch/powerpc/kernel/pci-common.c:1100:10: error: assignment from incompatible pointer type [-Werror=incompatible-pointer-types]
>      for (pp = &parent->child; (p = *pp) != NULL; pp = &p->sibling) {
>              ^

This reparent_resources() function is duplicated with the one in
arch/microblaze/pci/pci-common.c which has been fixed in v4. I planned
to move it to kernel/resource.c in a separate patch since it's shared by
different ARCH, then fix it in this patch.

>    arch/powerpc/kernel/pci-common.c:1100:50: error: assignment from incompatible pointer type [-Werror=incompatible-pointer-types]
>      for (pp = &parent->child; (p = *pp) != NULL; pp = &p->sibling) {
>                                                      ^
> >> arch/powerpc/kernel/pci-common.c:1113:13: error: incompatible types when assigning to type 'struct list_head' from type 'struct resource *'
>      res->child = *firstpp;
>                 ^
>    arch/powerpc/kernel/pci-common.c:1114:15: error: incompatible types when assigning to type 'struct list_head' from type 'struct resource *'
>      res->sibling = *pp;
>                   ^
> >> arch/powerpc/kernel/pci-common.c:1117:9: error: incompatible types when assigning to type 'struct resource *' from type 'struct list_head'
>      for (p = res->child; p != NULL; p = p->sibling) {
>             ^
>    arch/powerpc/kernel/pci-common.c:1117:36: error: incompatible types when assigning to type 'struct resource *' from type 'struct list_head'
>      for (p = res->child; p != NULL; p = p->sibling) {
>                                        ^
>    cc1: all warnings being treated as errors
> 
> vim +764 arch/powerpc/kernel/pci-common.c
> 
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  642  
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  643  /**
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  644   * pci_process_bridge_OF_ranges - Parse PCI bridge resources from device tree
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  645   * @hose: newly allocated pci_controller to be setup
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  646   * @dev: device node of the host bridge
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  647   * @primary: set if primary bus (32 bits only, soon to be deprecated)
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  648   *
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  649   * This function will parse the "ranges" property of a PCI host bridge device
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  650   * node and setup the resource mapping of a pci controller based on its
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  651   * content.
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  652   *
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  653   * Life would be boring if it wasn't for a few issues that we have to deal
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  654   * with here:
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  655   *
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  656   *   - We can only cope with one IO space range and up to 3 Memory space
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  657   *     ranges. However, some machines (thanks Apple !) tend to split their
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  658   *     space into lots of small contiguous ranges. So we have to coalesce.
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  659   *
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  660   *   - Some busses have IO space not starting at 0, which causes trouble with
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  661   *     the way we do our IO resource renumbering. The code somewhat deals with
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  662   *     it for 64 bits but I would expect problems on 32 bits.
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  663   *
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  664   *   - Some 32 bits platforms such as 4xx can have physical space larger than
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  665   *     32 bits so we need to use 64 bits values for the parsing
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  666   */
> cad5cef6 Greg Kroah-Hartman     2012-12-21  667  void pci_process_bridge_OF_ranges(struct pci_controller *hose,
> cad5cef6 Greg Kroah-Hartman     2012-12-21  668  				  struct device_node *dev, int primary)
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  669  {
> 858957ab Kevin Hao              2013-05-16  670  	int memno = 0;
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  671  	struct resource *res;
> 654837e8 Andrew Murray          2014-02-25  672  	struct of_pci_range range;
> 654837e8 Andrew Murray          2014-02-25  673  	struct of_pci_range_parser parser;
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  674  
> b7c670d6 Rob Herring            2017-08-21  675  	printk(KERN_INFO "PCI host bridge %pOF %s ranges:\n",
> b7c670d6 Rob Herring            2017-08-21  676  	       dev, primary ? "(primary)" : "");
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  677  
> 654837e8 Andrew Murray          2014-02-25  678  	/* Check for ranges property */
> 654837e8 Andrew Murray          2014-02-25  679  	if (of_pci_range_parser_init(&parser, dev))
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  680  		return;
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  681  
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  682  	/* Parse it */
> 654837e8 Andrew Murray          2014-02-25  683  	for_each_of_pci_range(&parser, &range) {
> e9f82cb7 Benjamin Herrenschmidt 2008-10-14  684  		/* If we failed translation or got a zero-sized region
> e9f82cb7 Benjamin Herrenschmidt 2008-10-14  685  		 * (some FW try to feed us with non sensical zero sized regions
> e9f82cb7 Benjamin Herrenschmidt 2008-10-14  686  		 * such as power3 which look like some kind of attempt at exposing
> e9f82cb7 Benjamin Herrenschmidt 2008-10-14  687  		 * the VGA memory hole)
> e9f82cb7 Benjamin Herrenschmidt 2008-10-14  688  		 */
> 654837e8 Andrew Murray          2014-02-25  689  		if (range.cpu_addr == OF_BAD_ADDR || range.size == 0)
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  690  			continue;
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  691  
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  692  		/* Act based on address space type */
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  693  		res = NULL;
> 654837e8 Andrew Murray          2014-02-25  694  		switch (range.flags & IORESOURCE_TYPE_BITS) {
> 654837e8 Andrew Murray          2014-02-25  695  		case IORESOURCE_IO:
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  696  			printk(KERN_INFO
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  697  			       "  IO 0x%016llx..0x%016llx -> 0x%016llx\n",
> 654837e8 Andrew Murray          2014-02-25  698  			       range.cpu_addr, range.cpu_addr + range.size - 1,
> 654837e8 Andrew Murray          2014-02-25  699  			       range.pci_addr);
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  700  
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  701  			/* We support only one IO range */
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  702  			if (hose->pci_io_size) {
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  703  				printk(KERN_INFO
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  704  				       " \\--> Skipped (too many) !\n");
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  705  				continue;
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  706  			}
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  707  #ifdef CONFIG_PPC32
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  708  			/* On 32 bits, limit I/O space to 16MB */
> 654837e8 Andrew Murray          2014-02-25  709  			if (range.size > 0x01000000)
> 654837e8 Andrew Murray          2014-02-25  710  				range.size = 0x01000000;
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  711  
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  712  			/* 32 bits needs to map IOs here */
> 654837e8 Andrew Murray          2014-02-25  713  			hose->io_base_virt = ioremap(range.cpu_addr,
> 654837e8 Andrew Murray          2014-02-25  714  						range.size);
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  715  
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  716  			/* Expect trouble if pci_addr is not 0 */
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  717  			if (primary)
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  718  				isa_io_base =
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  719  					(unsigned long)hose->io_base_virt;
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  720  #endif /* CONFIG_PPC32 */
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  721  			/* pci_io_size and io_base_phys always represent IO
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  722  			 * space starting at 0 so we factor in pci_addr
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  723  			 */
> 654837e8 Andrew Murray          2014-02-25  724  			hose->pci_io_size = range.pci_addr + range.size;
> 654837e8 Andrew Murray          2014-02-25  725  			hose->io_base_phys = range.cpu_addr - range.pci_addr;
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  726  
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  727  			/* Build resource */
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  728  			res = &hose->io_resource;
> 654837e8 Andrew Murray          2014-02-25  729  			range.cpu_addr = range.pci_addr;
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  730  			break;
> 654837e8 Andrew Murray          2014-02-25  731  		case IORESOURCE_MEM:
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  732  			printk(KERN_INFO
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  733  			       " MEM 0x%016llx..0x%016llx -> 0x%016llx %s\n",
> 654837e8 Andrew Murray          2014-02-25  734  			       range.cpu_addr, range.cpu_addr + range.size - 1,
> 654837e8 Andrew Murray          2014-02-25  735  			       range.pci_addr,
> 654837e8 Andrew Murray          2014-02-25  736  			       (range.pci_space & 0x40000000) ?
> 654837e8 Andrew Murray          2014-02-25  737  			       "Prefetch" : "");
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  738  
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  739  			/* We support only 3 memory ranges */
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  740  			if (memno >= 3) {
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  741  				printk(KERN_INFO
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  742  				       " \\--> Skipped (too many) !\n");
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  743  				continue;
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  744  			}
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  745  			/* Handles ISA memory hole space here */
> 654837e8 Andrew Murray          2014-02-25  746  			if (range.pci_addr == 0) {
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  747  				if (primary || isa_mem_base == 0)
> 654837e8 Andrew Murray          2014-02-25  748  					isa_mem_base = range.cpu_addr;
> 654837e8 Andrew Murray          2014-02-25  749  				hose->isa_mem_phys = range.cpu_addr;
> 654837e8 Andrew Murray          2014-02-25  750  				hose->isa_mem_size = range.size;
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  751  			}
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  752  
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  753  			/* Build resource */
> 654837e8 Andrew Murray          2014-02-25  754  			hose->mem_offset[memno] = range.cpu_addr -
> 654837e8 Andrew Murray          2014-02-25  755  							range.pci_addr;
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  756  			res = &hose->mem_resources[memno++];
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  757  			break;
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  758  		}
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  759  		if (res != NULL) {
> aeba3731 Michael Ellerman       2014-10-16  760  			res->name = dev->full_name;
> aeba3731 Michael Ellerman       2014-10-16  761  			res->flags = range.flags;
> aeba3731 Michael Ellerman       2014-10-16  762  			res->start = range.cpu_addr;
> aeba3731 Michael Ellerman       2014-10-16  763  			res->end = range.cpu_addr + range.size - 1;
> aeba3731 Michael Ellerman       2014-10-16 @764  			res->parent = res->child = res->sibling = NULL;
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  765  		}
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  766  	}
> 13dccb9e Benjamin Herrenschmidt 2007-12-11  767  }
> fa462f2d Benjamin Herrenschmidt 2007-12-20  768  
> 
> :::::: The code at line 764 was first introduced by commit
> :::::: aeba3731b150188685225b510886f1370d8814de powerpc/pci: Fix IO space breakage after of_pci_range_to_resource() change
> 
> :::::: TO: Michael Ellerman <mpe at ellerman.id.au>
> :::::: CC: Michael Ellerman <mpe at ellerman.id.au>
> 
> ---
> 0-DAY kernel test infrastructure                Open Source Technology Center
> https://lists.01.org/pipermail/kbuild-all                   Intel Corporation





More information about the kexec mailing list