* Randy Dunlap <rdunlap@xxxxxxxxxxxxx> [231210 18:54]: > > > On 12/10/23 14:56, Andrew Morton wrote: > > On Sun, 10 Dec 2023 21:13:15 +0000 Matthew Wilcox <willy@xxxxxxxxxxxxx> wrote: > > > >> On Sat, Dec 09, 2023 at 10:38:39PM -0800, Randy Dunlap wrote: > >>> - * mas_topiary_node() - Dispose of a singe node > >>> + * mas_topiary_node() - Dispose of a single node > >> > >> That's all well and good, but how do we dispose of singed nodes now? > > > > Lol, this is my all-time favorite typo! > > > > > > hp2:/usr/src/mm> git log | grep Singed-off-by | wc -l > > 59 > > > > (sounds painful) > > > > hp2:/usr/src/linux-6.7-rc4> grep -ri singe . > > > > Current linux-next shows: > > $ git grep -iw singe > arch/x86/events/intel/pt.c: * In case of singe-entry ToPA, always put the self-referencing END > arch/x86/kvm/vmx/vmx.c: * evmcs in singe VM shares same assist page. > drivers/input/misc/yealink.c: /* Combine up to <len> consecutive LCD bytes in a singe request > drivers/net/ethernet/brocade/bna/bfa_defs_status.h: * NOTE: The error msgs are auto generated from the comments. Only singe line > kernel/trace/trace_events_filter.c: * struct prog_entry - a singe entry in the filter program > lib/maple_tree.c: * mas_topiary_node() - Dispose of a singe node > lib/maple_tree.c: /* New root needs a singe node */ > sound/core/seq/seq_queue.c:/* enqueue a event to singe queue */ > sound/pci/rme9652/hdspm.c: /* Changing between Singe, Double and Quad speed is not > > but there is also "unsinged", "assinged", & "singel" if the -w is omitted. > > And somehow the two Chinese translations of submitting-patches.rst contain > Singed-off-by. > Nodes become singed, especially around the edges, when writes are too fast. We have to replace them with fresh nodes, otherwise tasks can't find the VMAs in that area. I will have to research how these other people were able to implement 'unsinged' - maybe there's a way to reuse these singed areas more effectively.
