linux-kernel - Re: [PATCH] usb: typec: tipd: drop double register read in tps6598x

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <2025110840-varnish-exhale-c362@gregkh>
Date: Sat, 8 Nov 2025 09:43:14 +0900
From: Greg KH <gregkh@...uxfoundation.org>
To: Peter Korsgaard <peter@...sgaard.com>
Cc: javier.carrasco@...fvision.net, heikki.krogerus@...ux.intel.com,
	neal@...pa.dev, linux-usb@...r.kernel.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH] usb: typec: tipd: drop double register read in
 tps6598x_interrupt

On Fri, Nov 07, 2025 at 04:03:36PM +0100, Peter Korsgaard wrote:
> >>>>> "Greg" == Greg KH <gregkh@...uxfoundation.org> writes:
> 
>  > On Thu, Nov 06, 2025 at 05:48:49PM +0100, Peter Korsgaard wrote:
>  >> Commit 409c1cfb5a80 ("usb: typec: tipd: fix event checking for tps6598x")
>  >> added (by accident?) a double read of the TPS_REG_INT_EVENT1 register.  Drop
>  >> that.
> 
>  > Are you sure?  Sometimes 2 reads are required.  How was this tested?
> 
> Hard to be 100% sure, but the code did not have a double read before the
> above commit and sticking a printk in the driver like this:
> 
> diff --git a/drivers/usb/typec/tipd/core.c b/drivers/usb/typec/tipd/core.c
> index 01db27cbf1d1..6687d192dbd4 100644
> --- a/drivers/usb/typec/tipd/core.c
> +++ b/drivers/usb/typec/tipd/core.c
> @@ -536,8 +536,9 @@ static irqreturn_t tps6598x_interrupt(int irq, void *data)
>                 intev_len = TPS_65987_8_INTEVENT_LEN;
> 
>         ret = tps6598x_block_read(tps, TPS_REG_INT_EVENT1, event1, intev_len);
> -
> +       printk(KERN_ERR "1st: %llx %llx\n", event1[0], event1[1]);
>         ret = tps6598x_block_read(tps, TPS_REG_INT_EVENT1, event1, intev_len);
> +       printk(KERN_ERR "2nd: %llx %llx\n", event1[0], event1[1]);
>         if (ret) {
>                 dev_err(tps->dev, "%s: failed to read event1\n", __func__);
>                 goto err_unlock;
> 
> 
> and (un)plugging the USB cable I see:
> 
> [ 3267.257341] 1st: 3000008 0
> [ 3267.262097] 2nd: 3000008 0
> 
> [ 3267.345179] 1st: 1000000 0
> [ 3267.350512] 2nd: 1000000 0
> 
> [ 3267.388947] 1st: 1000000 0
> [ 3267.393707] 2nd: 1000000 0
> 
> [ 3267.912112] 1st: 1000000 0
> [ 3267.916872] 2nd: 1000000 0
> 
> [ 3268.049505] 1st: 1000000 0
> [ 3268.054773] 2nd: 1000000 0
> 
> [ 3269.105173] 1st: 1000000 0
> [ 3269.109970] 2nd: 1000000 0
> 
> [ 3280.049111] 1st: 3000008 0
> [ 3280.053865] 2nd: 3000008 0
> 
> So I am fairly sure it is not needed.

Sometimes hardware requires it, even if it is not noticed by the actual
read value, so I would like to get an ack from the original author on
this before accepting it.

thanks,

greg k-h