[Letux-kernel] X1600 / LX16 support - here: fw_getenv() issue
H. Nikolaus Schaller
hns at goldelico.com
Tue Feb 6 22:38:08 CET 2024
Hi all,
> Am 06.02.2024 um 20:20 schrieb H. Nikolaus Schaller <hns at goldelico.com>:
>
> Hi all,
>
>> Am 06.02.2024 um 19:24 schrieb H. Nikolaus Schaller <hns at goldelico.com>:
>>
> All this does not fit to copying fw_arg2 to _fw_envp in fw_init_cmdline()
> and starting to scan at index 0 in fw_getenv().
>
> Here is a hint on what U-Boot bootm is doing:
>
> https://docs.u-boot.org/en/stable/usage/cmd/bootm.html?highlight=bootm
>
> If I read correctly there are two different formats. And, the second parameter
> is NOT a list of some environment variables but the start address of a RAMDISK?
>
> But what is then the story behind fw_getenv() at all?
>
> And U-Boot starts the kernel by
>
> bootcmd=fatload mmc 0 0x80a00000 /uImage; bootm 0x80a00000
>
> I.e. there is no second parameter???
I just realized that this all fw_getenv() is MIPS specific! No other architecture
or driver uses this.
arch/mips/fw/lib/cmdline.c:char *fw_getenv(char *envname)
So I am not sure if this needs some patch in U-Boot to do what is expected.
And I do not exactly know what is expected :)
Some MIPS processors seem to use fw_getenv() to look for "initrd_start", "fdt_start",
"yamontty", "cpuclock", "memsize", "modetty0".
And because I grepped this on the 5.10 kernel, I found the answer why it works
on 5.10 but not with 6.8:
In 5.10.y fw_getenv() isn't called at all... Hence it is no problem that it would fail.
There is simply no setup_rng_seed() calling fw_getenv("rngseed"):
https://elixir.bootlin.com/linux/v6.8-rc3/source/arch/mips/kernel/setup.c#L760
https://elixir.bootlin.com/linux/v5.10.209/source/arch/mips/kernel/setup.c#L765
It was easy to ask git blame where this was introduced...
commit 056a68cea01edfa78b3474af1bfa39cc6bcc7bee
Author: Jason A. Donenfeld <Jason at zx2c4.com>
Date: Fri Sep 30 16:01:38 2022 +0200
mips: allow firmware to pass RNG seed to kernel
Nearly all other firmware environments have some way of passing a RNG
seed to initialize the RNG: DTB's rng-seed, EFI's RNG protocol, m68k's
bootinfo block, x86's setup_data, and so forth. This adds something
similar for MIPS, which will allow various firmware environments,
bootloaders, and hypervisors to pass an RNG seed to initialize the
kernel's RNG.
Signed-off-by: Jason A. Donenfeld <Jason at zx2c4.com>
Signed-off-by: Thomas Bogendoerfer <tsbogend at alpha.franken.de>
diff --git a/arch/mips/kernel/setup.c b/arch/mips/kernel/setup.c
index 2ca156a5b2319..39c79f67c7a33 100644
--- a/arch/mips/kernel/setup.c
+++ b/arch/mips/kernel/setup.c
@@ -42,6 +42,7 @@
#include <asm/setup.h>
#include <asm/smp-ops.h>
#include <asm/prom.h>
+#include <asm/fw/fw.h>
#ifdef CONFIG_MIPS_ELF_APPENDED_DTB
char __section(".appended_dtb") __appended_dtb[0x100000];
@@ -756,6 +757,24 @@ static void __init prefill_possible_map(void)
static inline void prefill_possible_map(void) {}
#endif
+static void __init setup_rng_seed(void)
+{
+ char *rng_seed_hex = fw_getenv("rngseed");
+ u8 rng_seed[512];
+ size_t len;
+
+ if (!rng_seed_hex)
+ return;
+
+ len = min(sizeof(rng_seed), strlen(rng_seed_hex) / 2);
+ if (hex2bin(rng_seed, rng_seed_hex, len))
+ return;
+
+ add_bootloader_randomness(rng_seed, len);
+ memzero_explicit(rng_seed, len);
+ memzero_explicit(rng_seed_hex, len * 2);
+}
+
void __init setup_arch(char **cmdline_p)
{
cpu_probe();
@@ -786,6 +805,8 @@ void __init setup_arch(char **cmdline_p)
paging_init();
memblock_dump_all();
+
+ setup_rng_seed();
}
unsigned long kernelsp[NR_CPUS];
Now the question is how to fix it:
a) in kernel (may be difficult and not upstreamable - or it should get a CONFIG option)
b) in U-Boot (may also be difficult)
So we have to research what the MIPS specific "firmware environment" is
where a pointer is passed in a register to fw_arg2.
BR,
Nikolaus
More information about the Letux-kernel
mailing list