Hi Gagan,Thank you very much for the quick response and for looking into this issue on your side.We will wait for your update. Please let us know if you need any further information, logs, or testing assistance from our side. Best regards Mr. Gore Tel:+86-18038128078 <tel:+86-18038128078 > =================================== 高佛庥 南京易科腾信息技术有限公司 供应链 地址：南京市江宁区联域路3号天域互联科技中心C1号楼7楼 电话：180-3812-8078 邮箱：[email protected] https://e-quantum.com.cn/ <https://e-quantum.com.cn/ > =================================== ====================================== 本邮件含有保密信息，未经允许，此邮件不得转发他人； 若邮件误发至您邮箱，请及时删除，不得作为他用。 ====================================== ------------------------------------------------------------------ 发件人：Gagandeep Singh <[email protected]> 发送时间：2026年3月12日(周四) 18:59 收件人："高佛庥"<[email protected]>; Hemant Agrawal<[email protected]> 抄 送：dev<[email protected]> 主 题：RE: URGENT: PFDR still exhausted after applying FQRNI patch - need immediate guidance Hi Mr. Gore, Let me check it on our side, I will update you soon. Regards, Gagan From:高佛庥 <[email protected]> Sent: Thursday, March 12, 2026 12:03 PM To: Gagandeep Singh <[email protected]>; Hemant Agrawal <[email protected]> Cc: dev <[email protected]> Subject: URGENT: PFDR still exhausted after applying FQRNI patch - need immediate guidance Some people who received this message don't often get email from [email protected] <mailto:[email protected] >. Learn why this is important <https://aka.ms/LearnAboutSenderIdentification > Dear Gagandeep, Hemant, We are facing a critical production outage on NXP LS1043A platforms running OVS-DPDK（ver:22.11.7）. Status:
* ✅ Your PFDR leak patch (bus/dpaa: fix PFDRs leaks due to FQRNIs) is applied * ❌ After ~90 days, PFDR still exhausts (from 524k to 487k entries) * ❌ Error: Qman ErrInt: PFDR Enqueues Blocked Interrupt * ❌ Device requires reboot to recover Your patch note states: "Please note that this drain_mr_fqrni() only look for FQRNI type messages to consume. If there are other type of messages like FQRN, FQRL, FQPN, ERN etc. also coming on MR then those messages need to be handled separately." We believe this is exactly our situation - other message types are accumulating and consuming PFDRs. Urgent request: * Are there known patches to handle FQRN, ERN, etc. messages? * Should we drain MR in additional locations (interrupt handlers, periodic timers)? * Can you provide a complete fix or point us to the right code modifications? We are willing to test any patch immediately. Our production systems are at risk. Please advise urgently. Best regards, Mr. Gore Tel:+86-18038128078 <Tel:+86-18038128078 > 高佛庥 南京易科腾信息技术有限公司（Nanjing EQTINFO Information Technology Co., Ltd） 电话：180-3812-8078 邮箱：[email protected] <mailto:[email protected] >
