erebion/coreboot
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

drivers/intel/dtbt: Skip mailbox commands on downstream bridges

Browse source 
Downstream bridges share the same PCI device ID as the upstream
bridge but have no firmware mailbox. Sending commands to them
causes timeouts. Add a forward declaration of dtbt_device_ops
to detect and skip bridges whose parent is also a dTBT device.

Tested on thinkpad t480: The 5s timeout is now gone.

Change-Id: I96febb0e52e0ffbe52a237212b8f708a7b05c6d7
Signed-off-by: Arthur Heymans <arthur@aheymans.xyz>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/91689
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Matt DeVillier <matt.devillier@gmail.com>
Reviewed-by: Angel Pons <th3fanbus@gmail.com>
This commit is contained in:
Arthur Heymans 2026-03-16 20:53:33 +01:00 • committed by Matt DeVillier
commit 674000732d
1 changed files with 11 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

11
src/drivers/intel/dtbt/dtbt.c
View file

@ -11,6 +11,7 @@
#include "chip.h"
#include "dtbt.h"
static struct device_operations dtbt_device_ops;
/* Returns true on success, false on timeout or TBT2PCIE Error field non-zero */
static bool dtbt_cmd(struct device *dev, u8 command, u32 data, u32 timeout)
@ -183,6 +184,16 @@ static void dtbt_enable(struct device *dev)
if (!is_dev_enabled(dev) || dev->path.pci.devfn != 0)
return;
/*
* Only the upstream bridge has the firmware mailbox (PCIE2TBT/TBT2PCIE
* registers). Downstream bridges behind it share the same PCI device ID
* but have no mailbox, so sending commands to them will timeout. Skip any
* bridge whose parent is also a dTBT bridge (i.e. a downstream bridge).
*/
if (dev->upstream && dev->upstream->dev &&
dev->upstream->dev->ops == &dtbt_device_ops)
return;
printk(BIOS_INFO, "dTBT controller found at %s\n", dev_path(dev));
/* Set security level (Table 37/428); failure aborts enable */