similar to: [PATCH] bug in nfsmount/main.c parse_int()

Commit-ID: c4b811a1e4647224ddc717fac59900d16d0e9d4d Gitweb: http://git.kernel.org/?p=libs/klibc/klibc.git;a=commit;h=c4b811a1e4647224ddc717fac59900d16d0e9d4d Author: Baptiste Jonglez <baptiste.jonglez at imag.fr> AuthorDate: Thu, 14 Sep 2017 09:22:21 -0700 Committer: Ben Hutchings <ben at decadent.org.uk> CommitDate: Wed, 2 Jan 2019 03:08:04 +0000 [klibc] nfsmount: support

In case anyone's interested, I followed up in the linux-nfs mailing list: https://marc.info/?l=linux-nfs&m=156887818618861&w=2 Thanks, Alkis On 9/15/19 10:51 AM, Alkis Georgopoulos wrote: > I think I got it. > > Both nfsmount and `mount -t nfs` now default to rsize/wsize = 1 MB. > By lowering this to 32K, all issues are gone, even with the default > timeo=7. And

The standard mount option nowadays to specify NFS version is "nfsvers", as documented in nfs(5) on modern Linux systems. Up to now, nfsmount only supported the old "v2" or "v3" boolean options. Extend option parsing to support both "nfsvers=X" and "vers=X", with X being equal to either 2 or 3 (nfsmount does not support NFSv4 at present). If both

nfsmount doesnt work anymore for me. This patch fixes at least udp nfs mounts. tcp does still hang. client and server is 2.6. --- /dev/shm/bindresvport.c 2003-12-08 02:33:24.000000000 +0100 +++ klibc/inet/bindresvport.c 2004-01-06 21:05:02.283779072 +0100 @@ -23,7 +23,7 @@ int bindresvport(int sd, struct sockaddr if (sin == NULL) { sin = &me;

I think I got it. Both nfsmount and `mount -t nfs` now default to rsize/wsize = 1 MB. By lowering this to 32K, all issues are gone, even with the default timeo=7. And nfsroot=xxx client responsiveness is a whole lot better. I think when nfsmount was initially written, the default rsize/wsize were much lower, which matched the timeo=7. Now they cause the lags/timeouts that I reported. So

I managed to get to the bottom of this, and filed a bug report for NFS: https://bugzilla.kernel.org/show_bug.cgi?id=204939 Klibc nfsmount still a bug: it needs to NOT hardcode timeo=7. Either the NFS defaults should be used, which result in: timeo=600,rsize=1048576,wsize=1048576, or at least the kernel documented defaults, https://www.kernel.org/doc/Documentation/filesystems/nfs/nfsroot.txt

kinit/nfsmount.c:mount_nfs_root() should use the bootpath specified by bootp/dhcp. If the "nfsroot" option is specified then it overrides the boot server bootpath and a message indicating the override is printed. --- klibc-0.194/kinit/nfsroot.c.orig 2005-01-05 04:13:47.043897880 -0700 +++ klibc-0.194/kinit/nfsroot.c 2005-01-05 04:13:09.316633296 -0700 @@ -66,34 +66,21 @@ const int

Hi all, I have implemented portmap spoofing in klibc nfsmount (released as klibc-0.144) This is basically a vestigial portmap daemon which gets launched before the mount() call and then just records any transactions it gets to a file and sends back an affirmative reply. There are two ways to use it (this belongs in a README file, but it's too late at night right now): a) Set a fixed

Hi, The following set of patches is intended mostly to improve ipconfig and nfsmount compatibility with glibc. The context: I'm working on yaird, an alternative to mkinitrd, and currently implementing nfs root file systems. This is based on ipconfig and nfsmount from klibc, with a rewritten version of kinit. Users should be able to build the package with vanilla glibc, since klibc is not

Hey all, bug report time. I started tracking this down, but am at a bit of a loss for time, so I figured I'd report over this list to see if anyone else has the time to fix this or at least find the error. When calling nfsmount on it's own, say "nfsmount 1.2.3.4:/home/nfs /root" it works just fine, however, when calling nfsmount_main from within kinit (via nfsroot=), things get

Commit-ID: 886783e7a10fb7a638bc6034e4cdcb6296cea6a1 Gitweb: http://git.kernel.org/?p=libs/klibc/klibc.git;a=commit;h=886783e7a10fb7a638bc6034e4cdcb6296cea6a1 Author: Ben Hutchings <ben at decadent.org.uk> AuthorDate: Mon, 7 Oct 2019 17:18:41 +0100 Committer: Ben Hutchings <ben at decadent.org.uk> CommitDate: Mon, 7 Oct 2019 17:27:49 +0100 [klibc] nfsmount: Use kernel

Peter, we found that nfs over udp will corrupt data under very extrem load, there is no way to fix it due to the way how UDP works. TCP will not have these problems. I also wonder why the package size is only 1k. Everyone who wishes a slow connection can pass the desired options via the kernel cmdline. Everyone else prefers probably the fast mount. The defaults should look more like that: ---

I can't explain why 700 msecs aren't enough to avoid timeouts in 100 Mbps networks, but my tests verify it, so I'm writing to the list to request that you increase the default timeo to at least 30, or to 600 which is the default for `mount -t nfs`. How to reproduce: 1) Cabling: server <=> 100 Mbps switch <=> client Alternatively, one can use a 1000 Mbps switch and

There is a conflict in glibc between <linux/in.h> and <netinet/in.h>: one defines protocols with an enum, the other does the same, but adds a self-define for every enumerated value. #define IPPROTO_IP IPPROTO_IP This means the two cannot be mixed, and since loads of include files refer to <netinet/in.h>, effectively <linux/in.h> cannot be used with glibc. Since

Hello all, I have observed a problem with a diskless PXE client I am attempting to configure. PXE/NFS/DHCP/TFTPd server is running CentOS5.1 and the Diskless workstation's root and kernel was extracted from a CentOS5.1 (custom kernel due to setting to enable Root File System support). Problem: When the diskless client boots and logs in I notice that my root user is being squashed, even if I

Hi Bryan and other klibc people. I have taken a stamp on the Makefiles for klibc, this is what I came up with. 1) No longer recompile on every invocation 2) Correct checking on dependencies 3) Simpler makefile syntax (almost all over the place) I compile-time tested it only. Two open issues: a) Do we realy use .a files for initramfs. I renemed that to the executable. b) I renamed

linux-2.6 2ad780978b7c0c3e7877949f098cbd06e7c73839 cleaned up linux/nfsmount.h and thus a klibc build would produce this: > usr/kinit/nfsmount/mount.c:179: error: 'MNTPROC_MNT' undeclared (first use in this function) > usr/kinit/nfsmount/mount.c:179: error: (Each undeclared identifier is reported only once > usr/kinit/nfsmount/mount.c:179: error: for each function it appears

use AF_INET directly. Signed-off-by: maximilian attems <max at stro.at> --- usr/kinit/nfsmount/dummypmap.c | 2 +- usr/kinit/nfsmount/mount.c | 4 ++-- usr/kinit/nfsmount/sunrpc.c | 4 ++-- 3 files changed, 5 insertions(+), 5 deletions(-) diff --git a/usr/kinit/nfsmount/dummypmap.c b/usr/kinit/nfsmount/dummypmap.c index 43abe37..481e23b 100644 ---

The in-kernel nfsroot code is obsoleted by kinit. Remove it; it causes conflicts. Signed-off-by: H. Peter Anvin <hpa at zytor.com> --- commit 161e1dc16ec1129b30b634a2a8dcbbd1937800c5 tree c30da837d746fe65d8a13ccf6f27bd381948edb4 parent 018604e070e143657abcf0cb256a1e2dda205d97 author H. Peter Anvin <hpa at zytor.com> Sat, 20 May 2006 16:24:05 -0700 committer H. Peter Anvin <hpa at

hi, i trying to install a fai installation server in a DomU for DomUs, but i have some problems with nfs. The DomUs says, that they can''t reach the NFS Server =========== Begin: Retrying nfs mount ... IP-Config: eth0 hardware address 00:16:3e:7a:25:7c mtu 1500 DHCP IP-Config: eth0 guessed broadcast address 172.16.0.255 IP-Config: eth0 complete (from 172.16.0.151): address: