Content-Type: multipart/mixed; boundary="===============3278623134940252736==" MIME-Version: 1.0 From: Denis Kenzior Subject: Re: [PATCH] udev: Adding PCIe as a subsystem in udev Date: Thu, 26 Sep 2019 20:09:57 -0500 Message-ID: In-Reply-To: <9BCC1739A7EADA4790E1FFF4123FB3268BB6557C@BGSMSX101.gar.corp.intel.com> List-Id: To: ofono@ofono.org --===============3278623134940252736== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Hi Antara, On 9/26/19 4:46 AM, Borwankar, Antara wrote: > Hi Denis, > = > Regarding your below query > = > - Can you tell me why this new addition uses vid/pid with a 0x prefix whi= le the rest of the table does not? It seems weird & inconsistent. > Actually in case of pci device vendor and device id read using libudev AP= I is returning as "0x8086" and "0x7560" strings respectively. So I have add= ed the same values in vendor_list table. That is strange. Okay, so maybe we should just add a different look-up = table to handle this. Perhaps pci_driver_list[] and use vids/pids with = 0x prefixes there. > In case of usb device the vendor and device strings returned by same libu= dev API does not have "0x" prefix. I don=E2=80=99t have any idea why it is = so. > = > If you insist I can remove this prefix 0x from vendor_list table and chan= ge the implementation of check_pci_device() function to handle comparison o= f vendor and model strings with vid, pid form vendor_list. Yep, that might be another alternative. I'd be fine with either. = Having some entries with 0x and some without will confuse people and = lead to bugs. Regards, -Denis --===============3278623134940252736==--