Differences
This shows you the differences between two versions of the page.
| Both sides previous revision Previous revision Next revision | Previous revision | ||
|
soc:2010:cooldavid:journal:start [2010/07/29 07:06] cooldavid |
soc:2010:cooldavid:journal:start [2010/08/07 04:15] (current) cooldavid |
||
|---|---|---|---|
| Line 23: | Line 23: | ||
| * [[soc:2010:cooldavid:journal:week11|Week 11]]: | * [[soc:2010:cooldavid:journal:week11|Week 11]]: | ||
| * Testing TCP performance with different window size, network environment. | * Testing TCP performance with different window size, network environment. | ||
| - | * Discuss the testing results and find a reasonable size. | + | * Port wscale branch against current gPXE master |
| * [[soc:2010:cooldavid:journal:week12|Week 12]]: | * [[soc:2010:cooldavid:journal:week12|Week 12]]: | ||
| - | * Possible more TCP cleanup, tuning. | + | * Port wscale branch against current gPXE master |
| + | * Continue testing TCP performance with different window size, network environment. | ||
| + | * Trace the issue that gPXE download speed slow on KVM. | ||
| + | * Maybe after GSoC period | ||
| + | * Discuss the testing results and find a reasonable size. | ||
| * gPXE scheduling and program flow documentation. | * gPXE scheduling and program flow documentation. | ||
| * Possible more devel-docs. | * Possible more devel-docs. | ||
| Line 32: | Line 36: | ||
| * Gillies: 88E8055 PCI-E Gigabit Ethernet Controller [11ab:4363], after the "ifopen net0" I have the prompt back and it crashes (Keyboad is blocked, no message). If I am fast enough, I can type 2 or 3 letters on the new prompt before it's blocked. | * Gillies: 88E8055 PCI-E Gigabit Ethernet Controller [11ab:4363], after the "ifopen net0" I have the prompt back and it crashes (Keyboad is blocked, no message). If I am fast enough, I can type 2 or 3 letters on the new prompt before it's blocked. | ||
| * meteger: there are a few PCI IDs in the current tg3 that work fine when you do a tftp or HTTP transfer but fail miserably for iscsi, for some reason. but that driver was ported to etherboot in 2003, and it took 7 years for someone to request that specific PCI ID. | * meteger: there are a few PCI IDs in the current tg3 that work fine when you do a tftp or HTTP transfer but fail miserably for iscsi, for some reason. but that driver was ported to etherboot in 2003, and it took 7 years for someone to request that specific PCI ID. | ||
| - | * mcb30: better performance in kvm with 8k than with 64k. (TCP RX Window) | ||
