Skip to contentSkip to conversation

build warnings: multi_v7_defconfig + LPAE, printk with variable types

DISCUSSION Kevin Hilman 3 participants 8 messages Aug 21, 2013 – Aug 22, 2013

Subsystems (5)
lore
Share & tools
lore
#0 lore raw

I've started auto-building the multi_v7_defconfig with LPAE enabled, and noticed a handful of warnings[1]

Several are due to using prink format on a type that changes with a config option (e.g several below use %x to print a resource_size_t, which is obviously different between 32- and 64-bit builds.

So this got me to wondering if there is "one true way" for using printk on types that might change with config option (preferably without ifdef.)

The one that comes to mind would be to just always cast to the larger of the types and use the correspondig format. For the ones below, it would mean casting to (u64) and using '%llx'.

Is there a better way?

Kevin

[1]
Warnings:
arch/arm/mach-omap2/gpmc.c:1495:4: warning: format '%x' expects argument of type 'unsigned int', but argument 4 has type 'resource_size_t' [-Wformat] net/ipv6/addrconf.c:2793:22: warning: 'link_dev' may be used uninitialized in this function [-Wuninitialized] drivers/dma/ste_dma40.c:3228:2: warning: format '%x' expects argument of type 'unsigned int', but argument 4 has type 'resource_size_t' [-Wformat] drivers/dma/ste_dma40.c:3582:3: warning: format '%x' expects argument of type 'unsigned int', but argument 4 has type 'resource_size_t' [-Wformat] drivers/dma/ste_dma40.c:3582:3: warning: format '%x' expects argument of type 'unsigned int', but argument 5 has type 'resource_size_t' [-Wformat] drivers/dma/ste_dma40.c:3593:5: warning: format '%x' expects argument of type 'unsigned int', but argument 5 has type 'resource_size_t' [-Wformat] drivers/spi/spi-pl022.c:2178:9: warning: format '%x' expects argument of type 'unsigned int', but argument 2 has type 'resource_size_t' [-Wformat] drivers/tty/serial/imx.c:1542:6: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]

#1 lore raw
Quoted 11 lines · Kevin Hilman

I've started auto-building the multi_v7_defconfig with LPAE enabled, and noticed a handful of warnings[1]

Several are due to using prink format on a type that changes with a config option (e.g several below use %x to print a resource_size_t, which is obviously different between 32- and 64-bit builds.

So this got me to wondering if there is "one true way" for using printk on types that might change with config option (preferably without ifdef.)

Quoted 5 lines

The one that comes to mind would be to just always cast to the larger of the types and use the correspondig format. For the ones below, it would mean casting to (u64) and using '%llx'.

Is there a better way?

Yes, according to Documentation/printk-formats.txt:

"Physical addresses:

%pa 0x01234567 or 0x0123456789abcdef

For printing a phys_addr_t type (and its derivatives, such as resource_size_t) which can vary based on build options, regardless of the width of the CPU data path. Passed by reference."

I can send a patch fixing these warnings if you want.

Regards,

Fabio Estevam

#2 lore raw
Quoted 20 lines · Fabio Estevam
Quoted 11 lines · Kevin Hilman

I've started auto-building the multi_v7_defconfig with LPAE enabled, and noticed a handful of warnings[1]

Several are due to using prink format on a type that changes with a config option (e.g several below use %x to print a resource_size_t, which is obviously different between 32- and 64-bit builds.

So this got me to wondering if there is "one true way" for using printk on types that might change with config option (preferably without ifdef.)

Quoted 5 lines

The one that comes to mind would be to just always cast to the larger of the types and use the correspondig format. For the ones below, it would mean casting to (u64) and using '%llx'.

Is there a better way?

Yes, according to Documentation/printk-formats.txt:

Quoted 7 lines

"Physical addresses:

%pa 0x01234567 or 0x0123456789abcdef

For printing a phys_addr_t type (and its derivatives, such as resource_size_t) which can vary based on build options, regardless of the width of the CPU data path. Passed by reference."

doh, RTFM. That one is new since the last time I read it.

Quoted 3 lines

I can send a patch fixing these warnings if you want.

Yes, please. Thanks!

Please split it up into in to one patch for arch/arm/mach-omap2, one for drivers/dma/* and one for drivers/spi/* so they can be merged through the proper maintainers.

Thanks,

Kevin

#4 lore raw
In reply to Kevin Hilman
Quoted 3 lines · Kevin Hilman

[1]
Warnings:
arch/arm/mach-omap2/gpmc.c:1495:4: warning: format '%x' expects argument of type 'unsigned int', but argument 4 has type 'resource_size_t' [-Wformat]

For this warning, what about this?

arch/arm/mach-omap2/gpmc.c +0 -0
@@ -1491,8 +1491,8 @@ static int gpmc_probe_generic_child(struct

platform_device *pdev,

         */
        ret = gpmc_cs_remap(cs, res.start);
        if (ret < 0) {
-               dev_err(&pdev->dev, "cannot remap GPMC CS %d to 0x%x\n",
-                       cs, res.start);
+               dev_err(&pdev->dev, "cannot remap GPMC CS %d to 0x%pa\n",
+                       cs, &res.start);
                goto err;
        }
#3 lore raw
In reply to Kevin Hilman
Quoted 5 lines · Kevin Hilman

Yes, please. Thanks!

Please split it up into in to one patch for arch/arm/mach-omap2, one for drivers/dma/* and one for drivers/spi/* so they can be merged through the proper maintainers.

Ok, will do it later today.

Regards,

Fabio Estevam

#5 lore raw
In reply to Fabio Estevam
Quoted 12 lines · Fabio Estevam
Quoted 5 lines · Kevin Hilman

[1]
Warnings:

        arch/arm/mach-omap2/gpmc.c:1495:4: warning: format '%x'

expects argument of type 'unsigned int', but argument 4 has type resource_size_t' [-Wformat]

For this warning, what about this?

arch/arm/mach-omap2/gpmc.c static int gpmc_probe_generic_child(struct · line 1496

Quoted 8 lines
arch/arm/mach-omap2/gpmc.c +2 -2
@@ -1491,8 +1491,8 @@ static int gpmc_probe_generic_child(struct
 platform_device *pdev,
         */
        ret = gpmc_cs_remap(cs, res.start);
        if (ret < 0) {
-               dev_err(&pdev->dev, "cannot remap GPMC CS %d to 0x%x\n",
-                       cs, res.start);
+               dev_err(&pdev->dev, "cannot remap GPMC CS %d to 0x%pa\n",
+                       cs, &res.start);

hmm, why did you add the '&' ? That changes the behavior, not just the format.

Kevin

#6 lore raw
Quoted 11 lines · Kevin Hilman
Quoted 9 lines · Fabio Estevam
Quoted 5 lines · Kevin Hilman

[1]
Warnings:

        arch/arm/mach-omap2/gpmc.c:1495:4: warning: format '%x'

expects argument of type 'unsigned int', but argument 4 has type resource_size_t' [-Wformat]

For this warning, what about this?

Quoted 14 lines

arch/arm/mach-omap2/gpmc.c static int gpmc_probe_generic_child(struct · line 1496

Quoted 11 lines
arch/arm/mach-omap2/gpmc.c +2 -2
@@ -1491,8 +1491,8 @@ static int gpmc_probe_generic_child(struct
 platform_device *pdev,
         */
        ret = gpmc_cs_remap(cs, res.start);
        if (ret < 0) {
-               dev_err(&pdev->dev, "cannot remap GPMC CS %d to 0x%x\n",
-                       cs, res.start);
+               dev_err(&pdev->dev, "cannot remap GPMC CS %d to 0x%pa\n",
+                       cs, &res.start);

hmm, why did you add the '&' ? That changes the behavior, not just the format.

Remember that these strings are checked by the standard printf-format checker. So, in order to remain compatible, these extra extensions are made to the %p format - hence %p with an a suffix. Normally, the 'a' would get printed with standard printf, but the kernel's version will detect the following letter and use it to decide how to print the pointer argument.

See Documentation/printk-formats.txt, and note that for %pa is says "passed by reference".

If you don't like passing by reference here, the alternative solution is this, from the same document:

If <type> is dependent on a config option for its size (e.g., sector_t, blkcnt_t) or is architecture-dependent for its size (e.g., tcflag_t), use a format specifier of its largest possible type and explicitly cast to it. Example:

printk("test: sector number/total blocks: %llu/%llu\n",
(unsigned long long)sector, (unsigned long long)blockcount);

#7 lore raw
In reply to Kevin Hilman
Quoted 11 lines · Kevin Hilman
Quoted 9 lines · Fabio Estevam
Quoted 5 lines · Kevin Hilman

[1]
Warnings:

        arch/arm/mach-omap2/gpmc.c:1495:4: warning: format '%x'

expects argument of type 'unsigned int', but argument 4 has type resource_size_t' [-Wformat]

For this warning, what about this?

Quoted 14 lines

arch/arm/mach-omap2/gpmc.c static int gpmc_probe_generic_child(struct · line 1496

Quoted 11 lines
arch/arm/mach-omap2/gpmc.c +2 -2
@@ -1491,8 +1491,8 @@ static int gpmc_probe_generic_child(struct
 platform_device *pdev,
         */
        ret = gpmc_cs_remap(cs, res.start);
        if (ret < 0) {
-               dev_err(&pdev->dev, "cannot remap GPMC CS %d to 0x%x\n",
-                       cs, res.start);
+               dev_err(&pdev->dev, "cannot remap GPMC CS %d to 0x%pa\n",
+                       cs, &res.start);

hmm, why did you add the '&' ? That changes the behavior, not just the format.

Yes, this is required. Please refer to the 'Passed by reference.' in the end of the paragrah of Documention/printk-formats.txt.

The other good thing of using '%pa' is that I can remove the '0x' as well. Just noticed this after doing some local tests. Will submit the patches shortly.