Post-facto backport request: USB: serial: f81232: fix incomplete serial port generation

Post-facto backport request: USB: serial: f81232: fix incomplete serial port generation

[PS10 PETER HONG 洪繼澤]

Hi Stable Team,

I would like to request a backport for the following commit to the
currently supported stable trees 6.1.y, 6.6.y, 6.12.y and 6.18.y

   cd644b805da8 ("USB: serial: f81232: fix incomplete serial port 
generation")

Reason:
   This patch fixes a stability issue where Fintek F81532A/534A/535/536
   devices fail to initialize all serial ports during fast load/unload 
cycles.
   The fix involves a dummy read to clear the device's stale internal state.

The patch should apply cleanly to most recent stable branches.

Best regards,
Peter
-- 

*洪繼澤 **Peter Hong*

精拓科技股份有限公司

Feature Integration Technology

Address: 302新竹縣竹北市台元二街10號7樓

TEL: 03-5600168 #813

FAX: 03-5600166

E-Mail﹕peter_hong@fintek.com.tw

Re: Post-facto backport request: USB: serial: f81232: fix incomplete serial port generation

[Johan Hovold]

On Tue, Mar 17, 2026 at 10:21:17AM +0800, PS10 PETER HONG 洪繼澤 wrote:
> Hi Stable Team,
> 
> I would like to request a backport for the following commit to the
> currently supported stable trees 6.1.y, 6.6.y, 6.12.y and 6.18.y
> 
>    cd644b805da8 ("USB: serial: f81232: fix incomplete serial port 
> generation")
> 
> Reason:
>    This patch fixes a stability issue where Fintek F81532A/534A/535/536
>    devices fail to initialize all serial ports during fast load/unload 
> cycles.
>    The fix involves a dummy read to clear the device's stale internal state.
> 
> The patch should apply cleanly to most recent stable branches.

My understanding was that this was only something you'd hit if you
unload and reload the driver which isn't something a user would normally
do. That's why I didn't add a stable tag.

The diff is a bit on the bigger side but it's all straight-forward
enough so I'm fine with backporting if you think it will be useful also
for older kernels.

Johan

Re: Post-facto backport request: USB: serial: f81232: fix incomplete serial port generation

[PS10 PETER HONG 洪繼澤]

Hi Johan,

Johan Hovold 於 2026/3/17 下午 03:22 寫道:
> On Tue, Mar 17, 2026 at 10:21:17AM +0800, PS10 PETER HONG 洪繼澤 wrote:
>> Hi Stable Team,
>>
>> I would like to request a backport for the following commit to the
>> currently supported stable trees 6.1.y, 6.6.y, 6.12.y and 6.18.y
>>
>>     cd644b805da8 ("USB: serial: f81232: fix incomplete serial port
>> generation")
>>
>> Reason:
>>     This patch fixes a stability issue where Fintek F81532A/534A/535/536
>>     devices fail to initialize all serial ports during fast load/unload
>> cycles.
>>     The fix involves a dummy read to clear the device's stale internal state.
>>
>> The patch should apply cleanly to most recent stable branches.
> My understanding was that this was only something you'd hit if you
> unload and reload the driver which isn't something a user would normally
> do. That's why I didn't add a stable tag.
>
> The diff is a bit on the bigger side but it's all straight-forward
> enough so I'm fine with backporting if you think it will be useful also
> for older kernels.

Apologies for the resend. My previous email wasn't set to plain-text mode.

Typical users are unlikely to encounter this, but one of our customers is
currently facing the issue. They have requested that we submit this patch to
the stable tree so they can utilize the mainstream driver without needing a
separate vendor driver.

In theory, applying this patch will improve the product's overall stability.
We would like to see this integrated into the mainstream kernel tree.

Best regards,
Peter

-- 

*洪繼澤 **Peter Hong*

精拓科技股份有限公司

Feature Integration Technology

Address: 302新竹縣竹北市台元二街10號7樓

TEL: 03-5600168 #813

FAX: 03-5600166

E-Mail﹕peter_hong@fintek.com.tw

Re: Post-facto backport request: USB: serial: f81232: fix incomplete serial port generation

[Johan Hovold]

Hi Peter,

On Wed, Mar 18, 2026 at 08:46:28AM +0800, PS10 PETER HONG 洪繼澤 wrote:
> Johan Hovold 於 2026/3/17 下午 03:22 寫道:
> > On Tue, Mar 17, 2026 at 10:21:17AM +0800, PS10 PETER HONG 洪繼澤 wrote:

> >> I would like to request a backport for the following commit to the
> >> currently supported stable trees 6.1.y, 6.6.y, 6.12.y and 6.18.y
> >>
> >>     cd644b805da8 ("USB: serial: f81232: fix incomplete serial port
> >> generation")
> >>
> >> Reason:
> >>     This patch fixes a stability issue where Fintek F81532A/534A/535/536
> >>     devices fail to initialize all serial ports during fast load/unload
> >> cycles.
> >>     The fix involves a dummy read to clear the device's stale internal state.

> Typical users are unlikely to encounter this, but one of our customers is
> currently facing the issue. They have requested that we submit this patch to
> the stable tree so they can utilize the mainstream driver without needing a
> separate vendor driver.
> 
> In theory, applying this patch will improve the product's overall stability.
> We would like to see this integrated into the mainstream kernel tree.

Sounds good.

Johan