2012/8/21 Stefan Metzmacher <metze@xxxxxxxxx>: > Hi Pavel, > >> This is a patchset that includes the rest of SMB2 changes: brlocks, leases and multicredit io. >> >> The changes are applied on top of Jeff's recent patchset and can be found here: >> http://git.altlinux.org/people/piastry/public/?p=cifs-2.6.git;a=shortlog;h=refs/heads/smb2-dev-cork >> >> Pavel Shilovsky (10): >> CIFS: Improve byte-range locks handling >> CIFS: Handle SMB2 lock flags >> CIFS: Move brlock code to ops struct >> CIFS: Add brlock support for SMB2 >> CIFS: Request SMB2.1 leases >> CIFS: Add SMB2.1 lease break support >> CIFS: Fix cache coherency for read oplock case >> CIFS: Make use of multicredit write for SMB2 >> CIFS: Make use of multicredit reads for SMB2 > > Leases and multi-credit support are negotiated features in the capabilities > bits of the server. You should not rely on the dialect before you use them, > as there're servers which support SMB 2.1, but do not support leases or > multi-credit. Ok, thanks - will fix this. > > I think smb2_get_credit_size() should return 1 if the value in the header > is 0. It brings no problem because CreditCharge value is set to 1 in header assemble (and can be rewritten further in async read and write codepath) - so, get_credit_size gets at least 1. > > Regarding > + /* random value */ > + req->ClientGUID[0] = req->ClientGUID[3] = req->ClientGUID[7] = 7; > > I think you should use a random value here as the server uses this field > to detect if requests come from the same client. This should be a global > in the cifs.ok, which can change on reboot/module reload. > Yes, this is a good point. I just set it this way temporarily (for testing purpose) and forgot to replace with random bytes. Thanks. -- Best regards, Pavel Shilovsky. -- To unsubscribe from this list: send the line "unsubscribe linux-cifs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
