)]}'
{"src/erasurecode.c":[{"author":{"_account_id":15343,"name":"Tim Burke","email":"tburke@nvidia.com","username":"tburke"},"change_message_id":"450b1e5773aeba1a92ea2923be50b22176584c87","unresolved":false,"context_lines":[{"line_number":276,"context_line":""},{"line_number":277,"context_line":"    if (args-\u003ek \u003c 0 || args-\u003em \u003c 0)"},{"line_number":278,"context_line":"        return -EINVALIDPARAMS;"},{"line_number":279,"context_line":"    if ((args-\u003ek + args-\u003em) \u003e EC_MAX_FRAGMENTS) {"},{"line_number":280,"context_line":"        log_error(\"Total number of fragments (k + m) must be less than %d\\n\","},{"line_number":281,"context_line":"                  EC_MAX_FRAGMENTS);"},{"line_number":282,"context_line":"        return -EINVALIDPARAMS;"}],"source_content_type":"text/x-csrc","patch_set":1,"id":"5fc1f717_91d64a53","line":279,"updated":"2019-04-05 23:35:01.000000000","message":"EC_MAX_FRAGMENTS is still 32? So we haven\u0027t *actually* changed the cap yet...\n\nI wonder if this change *should* do that, and include some tests that exercise the new limit...\n\nLooking at https://github.com/01org/isa-l/blob/master/examples/ec/ec_simple_example.c it seems like ISA-L should support up to 255, FWIW -- thought it looks there may be more work than just bumping up EC_MAX_FRAGMENTS if we want to get close to *that* limit: https://github.com/openstack/liberasurecode/blob/master/src/backends/isa-l/isa_l_common.c#L59 :-/\n\nI\u0027ll at least see about making a patch to bump EC_MAX_FRAGMENTS up to 64, anyway...\n\nBut I guess if quadiron is looking for *hundreds* of frags, we\u0027re going to have to push this check down into each backend\u0027s init() anyway.","commit_id":"3251b297094b14cb22ed8b766946e9aabe0882e4"},{"author":{"_account_id":15343,"name":"Tim Burke","email":"tburke@nvidia.com","username":"tburke"},"change_message_id":"450b1e5773aeba1a92ea2923be50b22176584c87","unresolved":false,"context_lines":[{"line_number":555,"context_line":"    char **data_segments \u003d NULL;"},{"line_number":556,"context_line":"    char **parity_segments \u003d NULL;"},{"line_number":557,"context_line":"    int *missing_idxs \u003d NULL;"},{"line_number":558,"context_line":"    int *realloc_bv \u003d NULL;"},{"line_number":559,"context_line":""},{"line_number":560,"context_line":"    ec_backend_t instance \u003d liberasurecode_backend_instance_get_by_desc(desc);"},{"line_number":561,"context_line":"    if (NULL \u003d\u003d instance) {"}],"source_content_type":"text/x-csrc","patch_set":1,"id":"5fc1f717_3c187874","line":558,"range":{"start_line":558,"start_character":4,"end_line":558,"end_character":7},"updated":"2019-04-05 23:35:01.000000000","message":"Seems kinda silly to use a full int for something that should only ever store 0 or 1, but... eh, it\u0027s small...","commit_id":"3251b297094b14cb22ed8b766946e9aabe0882e4"},{"author":{"_account_id":29916,"name":"vianney rancurel","email":"vianney.rancurel@gmail.com","username":"vrancurel"},"change_message_id":"36d1b197c71d0abd8e58e520013c49d067971996","unresolved":false,"context_lines":[{"line_number":555,"context_line":"    char **data_segments \u003d NULL;"},{"line_number":556,"context_line":"    char **parity_segments \u003d NULL;"},{"line_number":557,"context_line":"    int *missing_idxs \u003d NULL;"},{"line_number":558,"context_line":"    int *realloc_bv \u003d NULL;"},{"line_number":559,"context_line":""},{"line_number":560,"context_line":"    ec_backend_t instance \u003d liberasurecode_backend_instance_get_by_desc(desc);"},{"line_number":561,"context_line":"    if (NULL \u003d\u003d instance) {"}],"source_content_type":"text/x-csrc","patch_set":1,"id":"5fc1f717_3ca71853","line":558,"range":{"start_line":558,"start_character":4,"end_line":558,"end_character":7},"in_reply_to":"5fc1f717_3c187874","updated":"2019-04-06 01:01:51.000000000","message":"yes :)","commit_id":"3251b297094b14cb22ed8b766946e9aabe0882e4"},{"author":{"_account_id":15343,"name":"Tim Burke","email":"tburke@nvidia.com","username":"tburke"},"change_message_id":"450b1e5773aeba1a92ea2923be50b22176584c87","unresolved":false,"context_lines":[{"line_number":584,"context_line":"    k \u003d instance-\u003eargs.uargs.k;"},{"line_number":585,"context_line":"    m \u003d instance-\u003eargs.uargs.m;"},{"line_number":586,"context_line":""},{"line_number":587,"context_line":"    realloc_bv \u003d alloca((k + m) * sizeof (int));"},{"line_number":588,"context_line":"    memset(realloc_bv, 0, (k + m) * sizeof (int));"},{"line_number":589,"context_line":"    "},{"line_number":590,"context_line":"    if (num_fragments \u003c k) {"}],"source_content_type":"text/x-csrc","patch_set":1,"id":"5fc1f717_b1f10604","line":587,"range":{"start_line":587,"start_character":17,"end_line":587,"end_character":23},"updated":"2019-04-05 23:35:01.000000000","message":"We\u0027re still going to have *some* reasonable limit on k and m, right? \u0027Cause if this blows the stack, we\u0027re in trouble. I\u0027m still a little nervous about our use of alloca() in shss... but I guess if this is reasonably small (1k? 10k?) it ought to be safe? IDK","commit_id":"3251b297094b14cb22ed8b766946e9aabe0882e4"},{"author":{"_account_id":29916,"name":"vianney rancurel","email":"vianney.rancurel@gmail.com","username":"vrancurel"},"change_message_id":"36d1b197c71d0abd8e58e520013c49d067971996","unresolved":false,"context_lines":[{"line_number":584,"context_line":"    k \u003d instance-\u003eargs.uargs.k;"},{"line_number":585,"context_line":"    m \u003d instance-\u003eargs.uargs.m;"},{"line_number":586,"context_line":""},{"line_number":587,"context_line":"    realloc_bv \u003d alloca((k + m) * sizeof (int));"},{"line_number":588,"context_line":"    memset(realloc_bv, 0, (k + m) * sizeof (int));"},{"line_number":589,"context_line":"    "},{"line_number":590,"context_line":"    if (num_fragments \u003c k) {"}],"source_content_type":"text/x-csrc","patch_set":1,"id":"5fc1f717_5ca4644b","line":587,"range":{"start_line":587,"start_character":17,"end_line":587,"end_character":23},"in_reply_to":"5fc1f717_b1f10604","updated":"2019-04-06 01:01:51.000000000","message":"100K fragments shall still be safe","commit_id":"3251b297094b14cb22ed8b766946e9aabe0882e4"}]}
