Linux Kernel 2.6.12 and later has a "cp2101" driver and its source (by Craig Shelley) in the usb section. It may help you to learn on the CP210x driver.

The Linux driver source from SiLabs (MCCI) is what never be called a "source code", actually it's a C parser output. Though I'm not a Linux freak, I feel it's a shameful manner, if SiLabs thinks they can cheat GPL by this way.
"Linux support for SiLabs CP210x devices, take 2" on "linux kernel monkey log"

Don't ask me for the protocol, I never do such a troublesome thing.
If I were you, I don't use CP210x in your situation.

As an enhancement to our Linux support, we have undertaken a development effort to update, enhance, and package the existing CP2101 driver that is currently in the public domain at:

http://www.kernel.org/

These enhancements include adding support for CP2102 and CP2103 devices in addition to porting the source to work with Linux kernel version 2.4. The current version supports Linux 2.6.

Once the work is completed, the source will be updated and made available in the public domain. The estimated release schedule is before the end of Q1 2006.

Regards,

If this means SiLabs continues to release the Linux driver by themselves, I don't think it's a good way. I suppose following story is better.
a) SiLabs opens the USB protocol of CP210x on demand under the prohibition such as SiLabs always attaches to their email.
b) A volunteer makes a Linux driver using SiLabs document at his own risk as usual as Linux way.
c) SiLabs refers the Linux driver by courtesy of the volunteer.

As a programmer, I highly appreciate Linux and the open-source movement. I've learnt much from their source code, I enjoy benefit from them in daily life. At the same time, however, I don't think they are all of the world.
Linux and the open-source movement is based on the spirit of "at my own risk". On the other hand, enterprise activity is based on "at the companies' risk". We pay our money for the companies' accomplishment to reduce our risk. If a company try to cross this boundary without correct recognition, it often brings a miserable result for both world such as SiLabs(MCCI) made on their Linux source code. So, I suppose the above way is happier solution for this problem.

If SiLabs hesitates to open the USB protocol of CP210x on demand, I should say SiLabs fears are utterly groundless.
If a mean engineer would try to reverse-engineer it, it doesn't take a day using an USB protocol analyzer with small cost. Nowadays, USB protocol analyzer costs as cheap as $400.
On the other hand, if SiLabs opens the protocol on demand under the prohibition, SiLabs can still reserve the proprietary right on the protocol.

So, I recommend you to reconsider the way SiLabs supports Linux.

Tsuneo

1) Encrypted Linux source code
In the SiLabs(MCCI) source code for CP210x Linux driver, following five files are encrypted as follows.
mcci_usb_ids.c, mcci_usb.c, usbif.h, vspctlda.h, vspwire.h

//------ mcci_usb_ids.c ------
#define ¥
l111110100 ¥
  ¥
(010304)
#define ¥
l111110101 ¥
  ¥
(0xea60)
#define ¥
l111110110 ¥
  ¥
(1)
int
mcci_vids
[
]
=
{
l111110100
,
l111110100
,
0
}
;
int
mcci_pids
[
]
=
{
l111110101
,
l111110110
,
0
}
;
//------ end of mcci_usb_ids.c ------

#define l111110100   (010304)
#define l111110101   (0xea60)
#define l111110110   (1)
int mcci_vids[]={l111110100,l111110100,0};
int mcci_pids[]={l111110101,l111110110,0};

Please consider what will happen if many companies would follow the way SiLabs(MCCI) encrypted the source code. It will threaten the open-source community. Therefore, it was accused as an outlaw.

If SiLabs still continues to release this kind of encrypted code in future version, it will never be accepted from Linux community. Linux 2.6 adopted the driver written by Craig Shelley instead of SiLabs one. Greg K-H, the maintainer of "USB SERIAL DRIVER" part of the Linux kernel, wrote as follows. "Linux support for SiLabs CP210x devices, take 2" Actually he has authority to decide which driver is included in Linux release.

2) "On demand"
With the term "on demand", I meant following situation.
a) The protocol is not published as a document such as an appnote on the web.
b) However, the protocol is opened when an user requests to SiLabs by mail.
c) SiLabs limits the usage of the information opened by email.
- SiLabs prohibits for the user to open the information itself.
- SiLabs allows for the user to make a software such as a Linux device driver using the information.
- SiLabs allows for the user to open their software.

I suppose the actual case such as the FLASH programming protocol over the JTAG on 'F12x. I received a kind reply when I asked SiLabs about this protocol.
a) The 'F12x programming protocol is not opened publicly on SiLabs documents, though 'F00x-02x protocol is published as an appnote.
b) However, it is opened to the user when (s)he asks it to SiLabs by email.
c) The usage of the information is limited by SiLabs statement attached to the reply.

If SiLabs provides the correct information on the CP210x protocol to the Linux driver authors, they will make their driver more reliable. Linux 2.6 has "CP2101" driver, independently coded by Craig Shelley. He wrote in the release note as follows.

------ "2.6.12-rc2 long-format changelog" -----
http://lwn.net/Articles/130476/

        [PATCH] USB: add driver for CP2101/CP2102 RS232 adaptors
        
        This patch against 2.6.10 provides the driver for the CP2101 and CP2102
        USB to RS232 adaptor chips which are found in some serial converter
        cables and usb mobile phone cables.
        
        It provides cp2101.c, and patches Kconfig and Makefile in
        drivers/usb/serial Since there is no technical data sheet freely
        available for these chips, much of the code in this driver has been
        based on protocol analysis, assumptions and guesswork.
        
        The cp210x family of devices have OEM programmable device IDs so updates
        will be required as and when new devices are found.
        
        The driver currently only supports configuration of baud rate, data
        bits, stop bits and parity mode. While this is enough to enable the use
        of a wide range of devices based on this chip, the driver is still
        incomplete in many areas such as flow control.  Unfortunately without
        technical datasheets, implementing these features is made very difficult
        indeed.
        
        Signed-off-by: Craig Shelley craig@microtron.org.uk
        Signed-off-by: Greg Kroah-Hartman 
------ "More USB bugfixes for 2.6.12-rc5" -----
http://lkml.org/lkml/2005/6/3/37
Craig Shelley:
  o USB: CP2101 Add support for flow control
------------------------