This patch enables driver support of MTUs greater than 1500. Though the adapter documentation says that maximum MTU is 16k, the largest packet I could send was 7440. So, I am setting that as the max
Jon Mason <jdmason@xxxxxxxxxx> : Ok. Powow: - rtl8169_close() is not protected as if it was issued during dev->close. It races with the irq handler. - I see no reason why the state of the device shou
Jon Mason wrote: This patch enables driver support of MTUs greater than 1500. Though the adapter documentation says that maximum MTU is 16k, the largest packet I could send was 7440. So, I am setting
I disagree. It should fail, complain loudly, and let the user try and fix it (or return it to the previous state). The code provided below will act like it is successful in changing the MTU, even if
Jon Mason <jdmason@xxxxxxxxxx> : Mmmm... Actually both patches try to paper over the real issue: the driver should reserve the ressources it needs for the change of mtu first. Ok, I'll take the patch
Agreed. I did a quick look in the other drivers, and didn't find any drivers that handle this possibility. Most of them ignore the fact. This sounds like an open hole in all of the drivers.
This patch enables driver support of MTUs greater than 1500. Though the adapter documentation says that maximum MTU is 16k, the largest packet I could send was 7440. So, I am setting that as the max
Jon Mason <jdmason@xxxxxxxxxx> : [...] Ok. Powow: - rtl8169_close() is not protected as if it was issued during dev->close. It races with the irq handler. - I see no reason why the state of the devic
I disagree. It should fail, complain loudly, and let the user try and fix it (or return it to the previous state). The code provided below will act like it is successful in changing the MTU, even if
Jon Mason <jdmason@xxxxxxxxxx> : Mmmm... Actually both patches try to paper over the real issue: the driver should reserve the ressources it needs for the change of mtu first. Ok, I'll take the patch
Agreed. I did a quick look in the other drivers, and didn't find any drivers that handle this possibility. Most of them ignore the fact. This sounds like an open hole in all of the drivers.