On 6/23/23 2:49 AM, Damien Le Moal wrote: >> Here are 8 patches against the 'for-next' branch of Damien's 'libata.git' repo. >> >> The libata code still often uses the 'unsigned long' type for the millisecond >> timeouts, while the kernel functions like msecs_to_jiffies() or msleep() only >> take 'unsigned int' parameters for those. I've started fixing the timeout types >> from ata_exec_internal[_sg]() that tripped the Svace static analyzer and posted >> couple patches, promising to post a large continuation series somewhat later... >> in my worst nightmare I couldn't imagine that this would take a whole year! :-( >> >> Sergey Shtylyov (8): >> ata: libata: fix parameter type of ata_deadline() >> ata: libata-core: fix parameter types of ata_wait_register() >> ata: libata-eh: fix reset timeout type >> ata: fix debounce timings type >> ata: libata-scsi: fix timeout type in ata_scsi_park_store() >> ata: libahci: fix parameter type of ahci_exec_polled_cmd() >> ata: ahci_xgene: fix parameter types of xgene_ahci_poll_reg_val() >> ata: sata_sil24: fix parameter type of sil24_exec_polled_cmd() > > We could merge patches 1-5 together as core updates, followed by driver patches. I'd like to avoid that. Please! :-) > Anyway, looks all OK to me but we are a little too close to the merge window. So > I will apply this once we have 6.5-rc1. Sigh, even more delaying... /me blames his stupidly long illness on December tho... [...] MBR, Sergey
