regd: sleeping in atomic

[Jiri Slaby <jirislaby@gmail.com>]

Hi,

Stanse found an atomic error in reg_copy_regd:

static int reg_copy_regd(const struct ieee80211_regdomain **dst_regd,
                          const struct ieee80211_regdomain *src_regd)
{
         struct ieee80211_regdomain *regd;
         int size_of_regd = 0;
         unsigned int i;

         size_of_regd = sizeof(struct ieee80211_regdomain) +
           ((src_regd->n_reg_rules + 1) * sizeof(struct 
ieee80211_reg_rule));

         regd = kzalloc(size_of_regd, GFP_KERNEL);       <---- here

Called from:

static void reg_regdb_search(struct work_struct *work)
{
         spin_lock(&reg_regdb_search_lock);
         while (!list_empty(&reg_regdb_search_list)) {
...
                 for (i=0; i<reg_regdb_size; i++) {
                         curdom = reg_regdb[i];

                         if (!memcmp(request->alpha2, curdom->alpha2, 2)) {
                                 r = reg_copy_regd(&regdom, curdom);
...
         spin_unlock(&reg_regdb_search_lock);
}

Whole error temporarily available at:
http://decibel.fi.muni.cz/~xslaby/stanse/error.cgi?db=34-rc&id=578

It is introduced by 3b377ea9d4efc94dc52fe41b4dfdb463635ab298.

Do you plan to extend it somehow or may the spinlock be converted to 
mutex? If not how much may size_of_regd be -- can we safely switch to 
GFP_ATOMIC?

-- 
js