Re: [PATCH] of: Custom printk format specifier for device node

[Pantelis Antoniou <pantelis.antoniou@konsulko.com>]

Hi Rob,

> On Jan 20, 2015, at 16:47 , Rob Herring <robherring2@gmail.com> wrote:
> 
> On Tue, Jan 20, 2015 at 8:34 AM, Pantelis Antoniou
> <pantelis.antoniou@konsulko.com> wrote:
>> 90% of the usage of device node's full_name is printing it out
>> in a kernel message. Preparing for the eventual delayed allocation
>> introduce a custom printk format specifier that is both more
>> compact and more pleasant to the eye.
> 
> What a great idea. ;)
> 

It was a stroke of genius I tell you… :)

>> For instance typical use is:
>>        pr_info("Frobbing node %s\n", node->full_name);
>> 
>> Which can be written now as:
>>        pr_info("Frobbing node %pO\n", node);
>> 
>> A verbose format specifier (1-2) can be used to print extra
>> information about the node like its phandle and node flags.
>> 
>> Signed-off-by: Pantelis Antoniou <pantelis.antoniou@konsulko.com>
>> ---
>> Documentation/printk-formats.txt | 18 +++++++++++
>> lib/vsprintf.c                   | 67 ++++++++++++++++++++++++++++++++++++++++
>> 2 files changed, 85 insertions(+)
>> 
>> diff --git a/Documentation/printk-formats.txt b/Documentation/printk-formats.txt
>> index 5a615c1..6ad199f 100644
>> --- a/Documentation/printk-formats.txt
>> +++ b/Documentation/printk-formats.txt
>> @@ -231,6 +231,24 @@ struct va_format:
>>        Do not use this feature without some mechanism to verify the
>>        correctness of the format string and va_list arguments.
>> 
>> +Device tree nodes:
>> +
>> +       %pO{,1,2}
> 
> 'O' is not very obvious, but I imagine we are somewhat limted in our
> choice here?
> 

All the good women are married, all the handsome men are gay, all the obvious
format specifiers are taken.

>> +
>> +       For printing device tree nodes. The (optional) number is the verbosity
>> +       level.
>> +
>> +       Examples:
>> +
>> +       %pO     /foo/bar@0              - Node full name
>> +       %pO0    /foo/bar@0              - Same as above
>> +       %pO1    /foo/bar@0[10]          - Node full name + phandle
>> +       %pO2    /foo/bar@0[10:DdPB]     - Node full name + phandle + node flags
>> +                                        D - dynamic
>> +                                        d - detached
>> +                                        P - Populated
>> +                                        B - Populated bus
> 
> We should think about what else we want to print for a node. Perhaps
> 'On' for name, 'Oc' for compatible, etc.
> 

The verbosity argument is something simple that works, but I take requests...

>> +
>> u64 SHOULD be printed with %llu/%llx:
>> 
>>        printk("%llu", u64_var);
>> diff --git a/lib/vsprintf.c b/lib/vsprintf.c
>> index ec337f6..72896cc 100644
>> --- a/lib/vsprintf.c
>> +++ b/lib/vsprintf.c
>> @@ -29,6 +29,7 @@
>> #include <linux/dcache.h>
>> #include <linux/cred.h>
>> #include <net/addrconf.h>
>> +#include <linux/of.h>
>> 
>> #include <asm/page.h>          /* for PAGE_SIZE */
>> #include <asm/sections.h>      /* for dereference_function_descriptor() */
>> @@ -1240,6 +1241,68 @@ char *address_val(char *buf, char *end, const void *addr,
>>        return number(buf, end, num, spec);
>> }
>> 
>> +#if IS_ENABLED(CONFIG_OF)
> 
> You can move this into the function:
> 
> if (!IS_ENABLED(CONFIG_OF)
>  return NULL;
> 
> Looks like you'll need the patch for empty of_node_check_flag as well.
> 

Yes, I tried to avoid that.

>> +static noinline_for_stack
>> +char *device_node_string(char *buf, char *end, struct device_node *dn,
>> +                        struct printf_spec spec, const char *fmt)
>> +{
>> +       char tbuf[sizeof("[xxxxxxxxxx:xxxx]") + 1];
>> +       const char *full_name;
>> +       int verbosity, len, flen, i;
>> +
>> +       if ((unsigned long)dn < PAGE_SIZE)
>> +               return string(buf, end, "(null)", spec);
>> +
>> +       full_name = of_node_full_name(dn);
>> +       verbosity = 0;
>> +       if (fmt[1] >= '0' && fmt[1] <= '2')
>> +               verbosity = fmt[1] - '0';
>> +
>> +       /* fast and simple case */
>> +       switch (verbosity) {
>> +       default:
>> +       case 0:
>> +               tbuf[0] = '\0';
>> +               break;
>> +       case 1:
>> +               snprintf(tbuf, sizeof(tbuf), "[%u]",
>> +                       (u32)dn->phandle);
>> +               break;
>> +       case 2:
>> +               snprintf(tbuf, sizeof(tbuf), "[%u:%c%c%c%c]",
>> +                       (u32)dn->phandle,
>> +                       of_node_check_flag(dn, OF_DYNAMIC) ? 'D' : '-',
>> +                       of_node_check_flag(dn, OF_DETACHED) ? 'd' : '-',
>> +                       of_node_check_flag(dn, OF_POPULATED) ? 'P' : '-',
>> +                       of_node_check_flag(dn, OF_POPULATED_BUS) ? 'B' : '-');
>> +               break;
>> +       }
>> +       flen = strlen(full_name);
>> +       len = flen + strlen(tbuf);
>> +       if (spec.precision > 0 && len > spec.precision)
>> +               len = spec.precision;
>> +
>> +       if (!(spec.flags & LEFT)) {
>> +               while (len < spec.field_width--) {
>> +                       if (buf < end)
>> +                               *buf = ' ';
>> +                       ++buf;
>> +               }
>> +       }
>> +       for (i = 0; i < len; ++i) {
>> +               if (buf < end)
>> +                       *buf = i < flen ? full_name[i] : tbuf[i - flen];
>> +               ++buf;
>> +       }
>> +       while (len < spec.field_width--) {
>> +               if (buf < end)
>> +                       *buf = ' ';
>> +               ++buf;
>> +       }
>> +       return buf;
>> +}
>> +#endif
>> +
>> int kptr_restrict __read_mostly;
>> 
>> /*
>> @@ -1459,6 +1522,10 @@ char *pointer(const char *fmt, char *buf, char *end, void *ptr,
>>                return dentry_name(buf, end,
>>                                   ((const struct file *)ptr)->f_path.dentry,
>>                                   spec, fmt);
>> +#if IS_ENABLED(CONFIG_OF)
> 
> With the above, then this ifdef should not be needed.
> 

Yes

>> +       case 'O':
>> +               return device_node_string(buf, end, ptr, spec, fmt);
>> +#endif
>>        }
>>        spec.flags |= SMALL;
>>        if (spec.field_width == -1) {
>> --
>> 1.7.12
>> 

OK, I’ll prepare the empty flag accessors patch and I’ll repost with the requested changes.

Regards

— Pantelis