Messages in this thread | | | Date | Wed, 22 Nov 2023 19:39:44 +0100 | From | Christian Marangi <> | Subject | Re: [net-next PATCH] net: phy: aquantia: drop wrong endianness conversion for addr and CRC |
| |
On Wed, Nov 22, 2023 at 10:23:47AM -0800, Jakub Kicinski wrote: > On Wed, 22 Nov 2023 18:53:39 +0100 Christian Marangi wrote: > > So they DO get converted to the HOST endian on reading the firmware from > > an nvmem cell or a filesystem? > > They don't get converted when "reading from nvmem / fs". > They get converted when you do: > > word = get_unaligned((const u32 *)(data + pos)); > > get_unaligned() is basically: > > #if BIGENDIAN > #define get_unaligned get_unaligned_be32 > #else > #define get_unaligned get_unaligned_le32 > #endif > > so you'll get different behavior here depending on the CPU.
Ugh... If that is true this is bad...
When get_unaligned was suggested, I checked if the thing was doing any kind of conversion and from [1] I tought it was just getting the pointer.
I can't find the entry where the thing is done. Is this some kind of include magic with asm specific API?
[1] https://elixir.bootlin.com/linux/latest/source/include/asm-generic/unaligned.h#L22
-- Ansuel
| |