[SLOF] [PATCH v2 09/10] libnet: Simplify the net-load arguments passing

[Thomas Huth]

On 10.10.2016 04:46, Alexey Kardashevskiy wrote:
> On 20/09/16 05:08, Thomas Huth wrote:
>> There is no need anymore to pass most of the arguments as
>> strings, we can use integers and pointers now instead.
>>
>> Signed-off-by: Thomas Huth <thuth at redhat.com>
>> ---
>>  lib/libnet/libnet.code       | 23 +++++++++--------------
>>  lib/libnet/netapps.h         |  3 ++-
>>  lib/libnet/netload.c         | 21 ++++++++++++---------
>>  slof/fs/packages/obp-tftp.fs | 15 ++++-----------
>>  4 files changed, 27 insertions(+), 35 deletions(-)
...
>> diff --git a/slof/fs/packages/obp-tftp.fs b/slof/fs/packages/obp-tftp.fs
>> index 30070a6..2e6bcfc 100644
>> --- a/slof/fs/packages/obp-tftp.fs
>> +++ b/slof/fs/packages/obp-tftp.fs
>> @@ -33,20 +33,13 @@ INSTANCE VARIABLE ciregs-buffer
>>      my-parent ihandle>phandle node>path encode-string
>>      s" bootpath" set-chosen
>>  
>> -    \ Generate arg string for snk like
>> -    \ "netboot load-addr length filename"
>> -    (u.) s" netboot " 2swap $cat s"  60000000 " $cat
>> +    60000000                        ( addr maxlen )
>>  
>>      \ Allocate 1720 bytes to store the BOOTP-REPLY packet
>> -    6B8 alloc-mem dup >r (u.) $cat
>> -    huge-tftp-load @ IF s"  1 " ELSE s"  0 " THEN $cat
>> -    \ Add desired TFTP-Blocksize as additional argument
>> -    s" 1432 " $cat
>> +    6B8 alloc-mem dup >r            ( addr maxlen replybuf )
>> +    huge-tftp-load @  d# 1428       ( addr maxlen replybuf hugetftp blocksize )
> 
> 
> This 1432 -> 1428 change does not seem to belong to this patch, why is this
> change?

1428 is the right maximum size that should be used for TFTP transfers
(otherwise there might be problems if networking is using VLANs
underneath, IIRC).
The old 1432 value is silently fixed by the code in tftp.c:

    /* Preferred blocksize - used as option for the read request */
    if (_blocksize < 8)
            _blocksize = 8;
    else if (_blocksize > MAX_BLOCKSIZE)
            _blocksize = MAX_BLOCKSIZE;

with MAX_BLOCKSIZE defined to 1428.

I guess I should have mentioned that in the commit message... or if you
prefer, I can also do it in a separate patch instead?

 Thomas