Greg Kroah-Hartman, i cili \u00ebsht\u00eb p\u00ebrgjegj\u00ebs p\u00ebr mb\u00ebshtetje t\u00eb deg\u00ebs s\u00eb stabilizuar t\u00eb kernelit Linux, ka vendosur t\u00eb ndaloj\u00eb pranimin e \u00e7do ndryshimi n\u00eb kernelin Linux q\u00eb vjen nga Universiteti i Minnesota, si dhe t\u00eb kthej\u00eb t\u00eb gjitha patch t\u00eb pranuara m\u00eb par\u00eb dhe ti rishikoj\u00eb ato. Arsyetimi p\u00ebr bllokimin ishte aktivizmi i grupit k\u00ebrkimor q\u00eb studionte mund\u00ebsit\u00eb e vendosjes s\u00eb dob\u00ebsive t\u00eb fshehura n\u00eb kodin e projekteve t\u00eb hapura. Ky grup kishte d\u00ebrguar patch-e q\u00eb p\u00ebrfshinin nj\u00eb gam\u00eb t\u00eb ndryshme gabimesh, kishte v\u00ebzhguar reagimin e komunitetit dhe kishte studiuar m\u00ebnyrat p\u00ebr t\u00eb mashtruar procesin e rishikimit t\u00eb ndryshimeve. Sipas Gregut, kryerja e eksperimenteve t\u00eb tilla p\u00ebr p\u00ebrfshirjen e ndryshimeve t\u00eb d\u00ebmshme \u00ebsht\u00eb e papranueshme dhe e pabesueshme. <\/p>\n
Arsye e bllokimit ishte d\u00ebrgimi nga an\u00ebtar\u00ebt e k\u00ebtij grupi i nj\u00eb patch-i, i cili shtonte nj\u00eb kontroll t\u00eb treguesit p\u00ebr t\u00eb p\u00ebrjashtuar mund\u00ebsin\u00eb e thirrjes s\u00eb dyfisht\u00eb t\u00eb funksionit \u00abfree\u00bb. Duke marr\u00eb parasysh kontekstin e p\u00ebrdorimit t\u00eb treguesit, kontrolli ishte i panevojsh\u00ebm. Q\u00ebllimi i d\u00ebrgimit t\u00eb patch-it ishte t\u00eb studiohej n\u00ebse ndryshimi i gabuar do t\u00eb kalonte rishikimin nga zhvilluesit e b\u00ebrtham\u00ebs. P\u00ebrve\u00e7 patch-it t\u00eb p\u00ebrmendur, dol\u00ebn edhe p\u00ebrpjekje t\u00eb tjera nga zhvilluesit e Universitetit t\u00eb Minesot\u00ebs p\u00ebr t\u00eb b\u00ebr\u00eb ndryshime t\u00eb dyshimta n\u00eb b\u00ebrtham\u00eb, p\u00ebrfshir\u00eb ato q\u00eb lidhen me shtimin e dob\u00ebsive t\u00eb fshehta. <\/p>\n
Pjes\u00ebtari q\u00eb d\u00ebrgoi patch-t\u00eb p\u00ebrpiqej t\u00eb mbrohej me pretekstin se po testonte nj\u00eb analizues t\u00eb ri statik dhe ndryshimi ishte p\u00ebrgatitur mbi baz\u00ebn e rezultateve t\u00eb kontrollit n\u00eb t\u00eb. Por Greg theksoi se korrigjimet e propozuara nuk ishin karakteristike p\u00ebr gabimet q\u00eb zbulohen nga analizuesit statik\u00eb dhe t\u00eb gjitha patch-et e d\u00ebrguara nuk e rregullonin asgj\u00eb n\u00eb t\u00ebr\u00ebsi. Duke marr\u00eb parasysh q\u00eb grupi i k\u00ebrkuesve n\u00eb fjal\u00eb kishte provuar t\u00eb promovonte m\u00eb par\u00eb korrigjime me dob\u00ebsi t\u00eb fshehura, \u00ebsht\u00eb evidente se ata vazhduan eksperimentet e tyre mbi komunitetin e zhvilluesve t\u00eb kernelit. <\/p>\n
\u00cbsht\u00eb interesante se, n\u00eb t\u00eb kaluar\u00ebn, drejtori i grupit q\u00eb po kryente eksperimentet kishte marr\u00eb pjes\u00eb n\u00eb nj\u00eb rregullim legjitim t\u00eb dob\u00ebsive, p\u00ebr shembull, kishte zbuluar rrjedhje informacioni n\u00eb tonin USB (CVE-2016-4482) dhe n\u00eb n\u00ebnstruktur\u00ebn rrjet t\u00eb (CVE-2016-4485). N\u00eb studimin mbi p\u00ebrparimin e fsheht\u00eb t\u00eb dob\u00ebsive, grupi nga Universiteti i Minesot\u00ebs p\u00ebrmend shembullin e dob\u00ebsis\u00eb CVE-2019-12819, e shkaktuar nga nj\u00eb rregullim q\u00eb ishte miratuar n\u00eb b\u00ebrtham\u00eb n\u00eb vitin 2014. Rregullimi shtonte nj\u00eb thirrje p\u00ebr put_device n\u00eb blokun e p\u00ebrpunimit t\u00eb gabimeve n\u00eb mdio_bus, por pas pes\u00eb vjet\u00ebsh doli se nj\u00eb manipulim i till\u00eb shkaktonte qasje n\u00eb nj\u00eb bllok memorie pas lirimit t\u00eb tij (\u00abuse-after-free\u00bb). <\/p>\n
Megjithat\u00eb, autor\u00ebt e studimit pretendojn\u00eb se n\u00eb pun\u00ebn e tyre kan\u00eb p\u00ebrgjith\u00ebsuar t\u00eb dh\u00ebnat mbi 138 patch-e q\u00eb shkaktojn\u00eb gabime dhe q\u00eb nuk lidhen me pjes\u00ebmarr\u00ebsit e studimit. P\u00ebrpjekjet p\u00ebr t\u00eb d\u00ebrguar patch-e t\u00eb tyre me gabime ishin t\u00eb kufizuara n\u00eb shk\u00ebmbimin e email-eve, dhe ndryshime t\u00eb tilla nuk arrit\u00ebn n\u00eb Git (n\u00ebse pas d\u00ebrgimit t\u00eb patch-it p\u00ebrmes email-it, mir\u00ebmbajt\u00ebsi e konsideronte patch-in normal, ata k\u00ebrkonin q\u00eb ndryshimi t\u00eb mos p\u00ebrfshihej, sepse kishte nj\u00eb gabim, pas s\u00eb cil\u00ebs d\u00ebrgonin nj\u00eb patch t\u00eb sakt\u00eb). <\/p>\n
Shtes\u00eb 1: Duke par\u00eb aktivitetin e autorit t\u00eb korrigjimit t\u00eb kritikuar, ai ka d\u00ebrguar prej koh\u00ebsh patch-e n\u00eb n\u00ebndeg\u00eb t\u00eb ndryshme t\u00eb b\u00ebrtham\u00ebs. P\u00ebr shembull, n\u00eb drejtorit\u00eb radeon dhe nouveau, s\u00eb fundmi u pranua nj\u00eb ndryshim me thirrjen pm_runtime_put_autosuspend(dev->dev) n\u00eb blokun e gabimit, nuk p\u00ebrjashtohet se kjo mund t\u00eb \u00e7oj\u00eb n\u00eb p\u00ebrdorimin e nj\u00eb bufferi pasi \u00ebsht\u00eb liruar memoria e lidhur me t\u00eb. <\/p>\n