Re: minor patch for 2.1.87 NFS client
Christopher Faylor (cgf@bbc.com)
Mon, 23 Feb 1998 20:14:11 GMT
In article <Pine.LNX.3.95.980223102736.10348A-100000@penguin.transmeta.com>,
Linus Torvalds <torvalds@transmeta.com> wrote:
>On Mon, 23 Feb 1998, Bill Hawes wrote:
>>
>>The attached minor patch provides a workaround for problems reported
>>with certain HP NFS servers. It appears that the servers are treating
>>the readdir buffer size as a longword rather than byte count, and then
>>overrun the available buffer. The patch works by detecting the first
>>overrun and then passes the size in longwords on subsequent readdir
>>messages. I've added a flags field to the RPC client structure that we
>>can use for future server-specific bug workarounds as well.
>>
>>The patch also includes conditionally-enabled code to pad RPC write
>>messages to a longword length. This provides a workaround for a few
>>older servers that return a "garbage args" error if the total RPC
>>message length isn't a longword multiple.
>
>I'd like to have others test this out for a bit: I actually tried the
>first "pad to longword" patch you sent out, and it resulted in some
>strange NFS behaviour for me (notably it made my mail-connections lock
>up: probably because my mail-client is one of the few programs that
>does a lot of unaligned writing due to writing out single messages with
>a fsync() in between..)
I've seen similar behavior with the "pad to longword" patch. In my case
it was a log file that triggered the problem. I got an "RPC Busy" and
was unable to perform any NFS accesses on any nfs disk until I rebooted.
--
http://www.bbc.com/ cgf@bbc.com "Strange how unreal
VMS=>UNIX Solutions Boston Business Computing the real can be."
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu