Re: [dpdk-dev] [PATCH v2 2/2] eal/linux: Add support for handling built-in kernel modules

[David Marchand <david.marchand@6wind.com>]

Hello Kamil,

On Mon, Dec 7, 2015 at 5:57 PM, Kamil Rytarowski <
Kamil.Rytarowski@caviumnetworks.com> wrote:

> Currently rte_eal_check_module() detects Linux kernel modules via reading
> /proc/modules. Built-in ones aren't listed there and therefore they are not
> being found by the script.
>
> Add support for checking built-in modules with parsing the sysfs files
>
> Signed-off-by: Kamil Rytarowski <Kamil.Rytarowski@caviumnetworks.com>
> ---
>  lib/librte_eal/linuxapp/eal/eal.c | 21 ++++++++++++++++++++-
>  1 file changed, 20 insertions(+), 1 deletion(-)
>
> diff --git a/lib/librte_eal/linuxapp/eal/eal.c
> b/lib/librte_eal/linuxapp/eal/eal.c
> index 635ec36..6cab906 100644
> --- a/lib/librte_eal/linuxapp/eal/eal.c
> +++ b/lib/librte_eal/linuxapp/eal/eal.c
> @@ -52,6 +52,8 @@
>  #if defined(RTE_ARCH_X86_64) || defined(RTE_ARCH_I686)
>  #include <sys/io.h>
>  #endif
> +#include <sys/types.h>
> +#include <sys/stat.h>
>
>  #include <rte_common.h>
>  #include <rte_debug.h>
> @@ -902,7 +904,10 @@ int
>  rte_eal_check_module(const char *module_name)
>  {
>         char mod_name[30]; /* Any module names can be longer than 30
> bytes? */
> +       char sysfs_mod_name[PATH_MAX];
> +       struct stat st;
>         int ret = 0;
> +       int rv;
>         int n;
>
>         if (NULL == module_name)
> @@ -918,9 +923,23 @@ rte_eal_check_module(const char *module_name)
>                 n = fscanf(fd, "%29s %*[^\n]", mod_name);
>                 if ((n == 1) && !strcmp(mod_name, module_name)) {
>                         ret = 1;
> -                       break;
> +                       goto finish;
>                 }
>         }
> +       RTE_LOG(DEBUG, EAL, "Module %s not found in /proc/modules",
> +               module_name);
> +
> +       /* A module might be builtin, try sysfs */
> +       snprintf(sysfs_mod_name, PATH_MAX, "/sys/module/%s", module_name);
> +       if ((rv = stat(sysfs_mod_name, &st)) == 0) {
> +               ret = 1;
> +               goto finish;
> +       }
> +
> +       RTE_LOG(DEBUG, EAL, "Open %s failed! error %i (%s)\n",
> +               sysfs_mod_name, errno, strerror(errno));
> +
> +finish:
>         fclose(fd);
>
>         return ret;
>

Well, in the end, won't all modules end up in /sys/module ?
So, I would say we can get rid of /proc/modules parsing.

The only thing that bothers me is this comment in the kernel documentation :


/sys/module/MODULENAME
                The name of the module that is in the kernel.
This
                module name will always show up if the module is loaded as
a
                dynamic module.  If it is built directly into the kernel,
it
                will only show up if it has a version or at least
one

parameter.


                Note: The conditions of creation in the built-in case are
not
                by design and may be removed in the
future.

But, at the moment, I suppose we are fine.


-- 
David Marchand