ContributionsMost RecentMost LikesSolutionsNot receiving interrupts from FPGA to HPS Hi, I have upgraded kernel from linux-socfpga 3.10.31 to 4.14.73 and I am having some troubles with getting the interrupts from the FPGA. I can see, when I do "cat /proc/interrupts" my kernel modules are not getting the interrupts from GIC as they used to from the older kernel but is now receiving from gpio-dwapb. In case of 3.10.31, IRQ number is set to expeced value and refer to the following please. #cat /proc/interrupts … 72 ... GIC-0 84 Level my_uio_driver In case of 4.14.73 I got … 72 ... gpio-dwapb 25 Edge my_uio_drive r I would be grateful if someone have already look something similar and could tell me where the mistake is. Thanks a lot in advance, Simone Re: Could not handle interrupt on E0-Nano-SoC Altera 5CSEMA4U23 Hello @Altera_Forum , I'm having the same issue. Did you find a solution? Thanks in advance, Simone Re: Interrupt handling problem using linux with DE1_SoC cyclone V Hello, I'm having the same issue with 4.14 kernel. I'm using an IP block on FPGA registered on interrupt n°73 and a Userspace driver as interface. If I type "cat /proc/interrupts" it returns .. ... ... ..... .... ..... 73: xxx 0 gpio-dwapb 3 Edge "my_uio_driver " .. ... ... ..... .... ..... with xxx different from 0, but blocked. I also try to make an IP which rise an interrupt (on 73) periodically but the number of interrupt is still blocked to initial value different from 0. Have you find a solution? Thanks in advance, Simone Re: Sending interrupts from FPGA to HPS Hello @dbhag2 , I'm having the same issue with 4.14 kernel. Have you find a solution? Thanks in advance, Simone Re: why cyclone v soc HPS running linux can only get 200 BogoMIPS? Hi @minmax , I still see "200.00 BogoMIPS" but I can see that the overall performances was heavily improved. I updated, in .dts file, the clocks/pll section as reported in 4.14.x of linux-socfpga git repo. Regards, Simone Re: why cyclone v soc HPS running linux can only get 200 BogoMIPS? HI AWu6, thanks for the replay. The slow boot, and generally the low performance of the processor was caused by a wrong definition of clocks/PLL/oscillators section in the devices tree. I update that device tree section with one found in linux-socfpga, (4.14.x) and everything goes well. Slow boot and generally low performance disappeared. That devices tree works well with 3.10.x kernel but with 4.14.x no. With the old device tree seems that the 4.14.x kernel started using a clock frequency lower than 800MHz. No one log or indication was present at startup in boot logs. Have a good day, Simone Re: why cyclone v soc HPS running linux can only get 200 BogoMIPS? I'm facing with the same issue.... Boot logs show me: " Switching to timer-based delay loop, resolution 10ns Calibrating delay loop (skipped), value calculated using timer frequency.. 200.00 BogoMIPS (lpj=1000000) " Moreover I'm getting a slower boot time respect when I used a 3.10.36-ltsi linux-socfpga kernel. @AWu6 did you get a slower boot too? Did you find a fix for that issue? @EBERLAZARE_I_Intel, I think is a Devices Tree related issue, may you share an example of dts working on 4.x linux-socfpga kernel? Thanks a lot in advance, Simone