Hi,
On 1/26/24 17:07, Hans de Goede wrote:
> After commit 936e4d49ecbc ("Input: atkbd - skip ATKBD_CMD_GETID in
> translated mode") the keyboard on Dell XPS 13 9350 / 9360 / 9370 models
> has stopped working after a suspend/resume.
>
> The problem appears to be that atkbd_probe() fails when called
> from atkbd_reconnect() on resume, which on systems where
> ATKBD_CMD_GETID is skipped can only happen by ATKBD_CMD_SETLEDS
> failing. ATKBD_CMD_SETLEDS failing because ATKBD_CMD_GETID was
> skipped is weird, but apparently that is what is happening.
Thinking more about it, what is likely happening here is that
ATKBD_CMD_SETLEDS is being send from atkbd_probe() where as
before atkbd_probe() would call ATKBD_CMD_GETID() and if that
succeeded (which it likely did) atkbd_probe() would continue with
calling atkbd_deactivate() and then exit, never calling
ATKBD_CMD_SETLEDS (at least not from atkbd_probe()).
So the problem seems to be that the embedded controller
does not like receiving ending ATKBD_CMD_SETLEDS as
the first command after resume and that being the first
command after resume is new behavior introduced by
936e4d49ecbc ("Input: atkbd - skip ATKBD_CMD_GETID in translated mode")
After applying both patches from this set (which is what
Paul tested), the ATKBD_CMD_GETID will still be skipped
but instead of replacing it with a ATKBD_CMD_SETLEDS
atkbd_probe() now continues with calling atkbd_deactivate()
and then exits as before the recent changes.
So after applying both patches here the behavior change
compared to before 936e4d49ecbc is limited to just
skipping ATKBD_CMD_GETID rather then effectively
replacing it with ATKBD_CMD_SETLEDS.
Regards,
Hans
>
> Fix this by also skipping ATKBD_CMD_SETLEDS when skipping
> ATKBD_CMD_GETID.
>
> Fixes: 936e4d49ecbc ("Input: atkbd - skip ATKBD_CMD_GETID in translated mode")
> Reported-by: Paul Menzel <pmenzel@xxxxxxxxxxxxx>
> Closes: https://lore.kernel.org/linux-input/0aa4a61f-c939-46fe-a572-08022e8931c7@xxxxxxxxxxxxx/
> Closes: https://bbs.archlinux.org/viewtopic.php?pid=2146300
> Closes: https://bugzilla.kernel.org/show_bug.cgi?id=218424
> Closes: https://bugzilla.redhat.com/show_bug.cgi?id=2260517
> Tested-by: Paul Menzel <pmenzel@xxxxxxxxxxxxx>
> Cc: stable@xxxxxxxxxxxxxxx
> Signed-off-by: Hans de Goede <hdegoede@xxxxxxxxxx>
> ---
> drivers/input/keyboard/atkbd.c | 13 ++++++++-----
> 1 file changed, 8 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/input/keyboard/atkbd.c b/drivers/input/keyboard/atkbd.c
> index 13ef6284223d..c229bd6b3f7f 100644
> --- a/drivers/input/keyboard/atkbd.c
> +++ b/drivers/input/keyboard/atkbd.c
> @@ -811,7 +811,6 @@ static int atkbd_probe(struct atkbd *atkbd)
> {
> struct ps2dev *ps2dev = &atkbd->ps2dev;
> unsigned char param[2];
> - bool skip_getid;
>
> /*
> * Some systems, where the bit-twiddling when testing the io-lines of the
> @@ -825,6 +824,11 @@ static int atkbd_probe(struct atkbd *atkbd)
> "keyboard reset failed on %s\n",
> ps2dev->serio->phys);
>
> + if (atkbd_skip_getid(atkbd)) {
> + atkbd->id = 0xab83;
> + return 0;
> + }
> +
> /*
> * Then we check the keyboard ID. We should get 0xab83 under normal conditions.
> * Some keyboards report different values, but the first byte is always 0xab or
> @@ -833,18 +837,17 @@ static int atkbd_probe(struct atkbd *atkbd)
> */
>
> param[0] = param[1] = 0xa5; /* initialize with invalid values */
> - skip_getid = atkbd_skip_getid(atkbd);
> - if (skip_getid || ps2_command(ps2dev, param, ATKBD_CMD_GETID)) {
> + if (ps2_command(ps2dev, param, ATKBD_CMD_GETID)) {
>
> /*
> - * If the get ID command was skipped or failed, we check if we can at least set
> + * If the get ID command failed, we check if we can at least set
> * the LEDs on the keyboard. This should work on every keyboard out there.
> * It also turns the LEDs off, which we want anyway.
> */
> param[0] = 0;
> if (ps2_command(ps2dev, param, ATKBD_CMD_SETLEDS))
> return -1;
> - atkbd->id = skip_getid ? 0xab83 : 0xabba;
> + atkbd->id = 0xabba;
> return 0;
> }
>
