On Mon, 03 Oct 2005 12:02:00 MDT, Jordan Crouse said: > +static u32 geode_data_read(void) { > + u32 val; > + > + val = *((u32 *) (geode_rng_base + GEODE_RNG_DATA_REG)); > + return val; > +} > + > +static unsigned int geode_data_present(void) { > + u32 val; > + > + val = *((u32 *) (geode_rng_base + GEODE_RNG_STATUS_REG)); > + return val; > +} Yowza. At least the intel_* routines do this sort of thing to semi-check that the sucker exists: assert (rng_mem != NULL); writeb (hw_status, rng_mem + INTEL_RNG_HW_STATUS); What does your code do if geode_init() manages to fail somehow?
Attachment:
pgpHvDv0OwjXp.pgp
Description: PGP signature
- References:
- [PATCH 6/7] AMD Geode GX/LX support
- From: "Jordan Crouse" <[email protected]>
- [PATCH 6/7] AMD Geode GX/LX support
- Prev by Date: Re: nforce3 lockup problem
- Next by Date: Re: what's next for the linux kernel?
- Previous by thread: [PATCH 6/7] AMD Geode GX/LX support
- Next by thread: Re: [PATCH 6/7] AMD Geode GX/LX support
- Index(es):