On 23:07-20240124, Dhruva Gole wrote:
On Jan 24, 2024 at 10:39:10 -0600, Nishanth Menon wrote:
On 18:38-20240124, Dhruva Gole wrote:
On Jan 24, 2024 at 16:42:12 +0530, Kamlesh Gurudasani wrote:
Dhruva Gole d-gole@ti.com writes:
The secure_hdr needs to be 0 init-ed however this was never being put into the secure_buf, leading to possibility of the first 4 bytes of secure_buf being possibly garbage.
Fix this by initialising the secure_hdr itself to the secure_buf location, thus when we make it 0, it automatically ensures the first 4 bytes are 0.
Fixes: 32cd25128bd849 ("firmware: Add basic support for TI System Control Interface (TI SCI)") Signed-off-by: Dhruva Gole d-gole@ti.com
Boot tested for sanity on AM62x SK https://gist.github.com/DhruvaG2000/724ceba3a0db03f4b0bff47de1160074
Cc: Nishanth Menon nm@ti.com Cc: Tom Rini trini@konsulko.com
drivers/firmware/ti_sci.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/firmware/ti_sci.c b/drivers/firmware/ti_sci.c index f5f659c11274..83ee8401a731 100644 --- a/drivers/firmware/ti_sci.c +++ b/drivers/firmware/ti_sci.c @@ -236,13 +236,13 @@ static int ti_sci_do_xfer(struct ti_sci_info *info, { struct k3_sec_proxy_msg *msg = &xfer->tx_message; u8 secure_buf[info->desc->max_msg_size];
- struct ti_sci_secure_msg_hdr secure_hdr;
struct ti_sci_secure_msg_hdr *secure_hdr = (struct ti_sci_secure_msg_hdr *)secure_buf; int ret;
if (info->is_secure) { /* ToDo: get checksum of the entire message */
secure_hdr.checksum = 0;secure_hdr.reserved = 0;
secure_hdr->checksum = 0;secure_hdr->reserved = 0;secure_hdr is pointer now, the value may be same but (struct ti_sci_secure_msg_hdr) would make more sense
Good catch Kamlesh! I have sent a new revision addressing this.
Makes no sense why we have secure API support in U-Boot. what is using this?
In my understanding of generic K3 boot flow, things like proc_boot and even applying or removing of firewalls will need a secure channel to talk to TIFS right? From my understanding secure host can only talk to TIFS and make such requests hence secure API.
U-boot runs in NS world, you cant even talk to the Trustzone Sec proxy without triggering a firewall violation.
https://software-dl.ti.com/tisci/esd/latest/2_tisci_msgs/general/TISCI_heade... "The Secure Messaging Header is only required when sending messages over secure transport. Messages sent over non-secure transport must not contain the secure messaging header."
btw, that checksum field should be renamed integ_check, but anyways..
So, I do not see a reason to even have that if condition in the first place and what real bug was getting fixed in this patch.