)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":1000687,"name":"Tomas Vanek","display_name":"Tomas Vanek","email":"vanekt@fbl.cz","username":"vanekt"},"change_message_id":"09b4c824d25fa43f8639a9b6c8d7431e5df4dfe8","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"2d0c2d28_a4e5ec12","updated":"2023-05-11 08:11:40.000000000","message":"This patch looks good, however the error propagation does not work as expected.\n\nI tested with Zynq 7000, it has a virtex2 compatible pld inside.\nFirst the status w/o error:\n \u003e virtex2 read_stat 0\n virtex2 status register: 0x56107ffc\n\nI stopped polling and disconnected TCLK.\n \u003e virtex2 read_stat 0\n virtex2 status register: 0xffffffff\n \nThis is probably ok, provided that the status register has no parity or any other means of error detection.\n\nI disconnected USB from the adapter (ftdi based).\n \u003e virtex2 read_stat 0\n libusb_handle_events() failed with LIBUSB_ERROR_NO_DEVICE\n unable to purge ftdi rx buffers: LIBUSB_ERROR_NO_DEVICE\n error while flushing MPSSE queue: -4\n virtex2 status register: 0xe02d2ac5\n\nHuh? An uninitialized memory is displayed as the result?\nI don\u0027t have time to dig deeper. If Daniel or anybody else wants to look at that it would be nice...","commit_id":"95d073dcee732898d4805a01cbe2c7012e5c0949"},{"author":{"_account_id":1001810,"name":"Daniel Anselmi","email":"danselmi@gmx.ch","username":"danselmi"},"change_message_id":"11d43d5df62262a8b7003d8f470555ca195e41ee","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"34ae2ee6_9d391bc0","in_reply_to":"1b5b3c3a_d39cfafa","updated":"2023-05-14 19:24:41.000000000","message":"Patch looks good!\n\nI can not reproduce Tomas\u0027 findings. Not with this patch anymore.\nError: libusb_handle_events() failed with LIBUSB_ERROR_NO_DEVICE\nError: unable to purge ftdi rx buffers: LIBUSB_ERROR_NO_DEVICE\nError: error while flushing MPSSE queue: -4\ncannot read virtex2 status register\n\n(Without this a patch I got the same results as Tomas. It was not a leak just an uninitialized variable on the stack.)","commit_id":"95d073dcee732898d4805a01cbe2c7012e5c0949"},{"author":{"_account_id":1000021,"name":"Antonio Borneo","email":"borneo.antonio@gmail.com","username":"borneoa"},"change_message_id":"69ca383b3b2e07db2aa50e77c8d62c622675e52c","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"1b5b3c3a_d39cfafa","in_reply_to":"2d0c2d28_a4e5ec12","updated":"2023-05-11 12:31:58.000000000","message":"\u003e This is probably ok, provided that the status register has no parity or any other means of error detection.\n\nI think so, the read stat function just send and receives from JTAG, no other way to detect if the value is correct.\n\n\u003e I disconnected USB from the adapter (ftdi based).\n\u003e  \u003e virtex2 read_stat 0\n\u003e  libusb_handle_events() failed with LIBUSB_ERROR_NO_DEVICE\n\u003e  unable to purge ftdi rx buffers: LIBUSB_ERROR_NO_DEVICE\n\u003e  error while flushing MPSSE queue: -4\n\u003e  virtex2 status register: 0xe02d2ac5\n\u003e \n\u003e Huh? An uninitialized memory is displayed as the result?\n\u003e I don\u0027t have time to dig deeper. If Daniel or anybody else wants to look at that it would be nice...\n\nI have checked the virtex2 code and I cannot find anything that could leak.\nMaybe something in JTAG generic code, or in the MPSSE code?","commit_id":"95d073dcee732898d4805a01cbe2c7012e5c0949"},{"author":{"_account_id":1000687,"name":"Tomas Vanek","display_name":"Tomas Vanek","email":"vanekt@fbl.cz","username":"vanekt"},"change_message_id":"a9279a934b5ffe038115eda52c345f9c6a105b28","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"b4844de0_339dc1bb","in_reply_to":"34ae2ee6_9d391bc0","updated":"2023-05-15 11:02:20.000000000","message":"\u003e I can not reproduce Tomas\u0027 findings. Not with this patch anymore.\n\nSorry for the noise!\nI unintentionally cherry-picked this patch on an older master where commit dd9137dc0e0c (\"pld/virtex2: add missing error checks\") was missing.\n\nRe-tested and looks good.","commit_id":"95d073dcee732898d4805a01cbe2c7012e5c0949"}]}