gcc-4.3 can't decide whether the constant value in kempld_prescaler[PRESCALER_21] is built-time constant or not, and gets confused by the logic in do_div():
drivers/watchdog/kempld_wdt.o: In function `kempld_wdt_set_stage_timeout': kempld_wdt.c:(.text.kempld_wdt_set_stage_timeout+0x130): undefined reference to `__aeabi_uldivmod'
This adds a call to ACCESS_ONCE() to force it to not consider it to be constant, and leaves the more efficient normal case in place for modern compilers, using an #ifdef to annotate why we do this hack.
Signed-off-by: Arnd Bergmann arnd@arndb.de --- drivers/watchdog/kempld_wdt.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-)
diff --git a/drivers/watchdog/kempld_wdt.c b/drivers/watchdog/kempld_wdt.c index 8e302d0e346c..3efa295ac627 100644 --- a/drivers/watchdog/kempld_wdt.c +++ b/drivers/watchdog/kempld_wdt.c @@ -140,12 +140,19 @@ static int kempld_wdt_set_stage_timeout(struct kempld_wdt_data *wdt_data, unsigned int timeout) { struct kempld_device_data *pld = wdt_data->pld; - u32 prescaler = kempld_prescaler[PRESCALER_21]; + u32 prescaler; u64 stage_timeout64; u32 stage_timeout; u32 remainder; u8 stage_cfg;
+#if GCC_VERSION < 40400 + /* work around a bug compiling do_div() */ + prescaler = READ_ONCE(kempld_prescaler[PRESCALER_21]); +#else + prescaler = kempld_prescaler[PRESCALER_21]; +#endif + if (!stage) return -EINVAL;