phone-side
/
nuttx-bb
9
0
Fork 0
Code Releases Activity

NFS update 

git-svn-id: https://nuttx.svn.sourceforge.net/svnroot/nuttx/trunk@4790 7fd9a85b-ad96-42d3-883c-3090e2eb8679
This commit is contained in:
patacongo 2012-05-30 23:21:37 +00:00
parent 69a4141090
commit 60c45d4a32
5 changed files with 119 additions and 71 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
nuttx/ChangeLog
View File

@ -2850,4 +2850,8 @@
compatible with the F4 and needs the same work-around for the missing BTF
signal that was needed for the F4.
* drivers/lcd/ssd1289.*: Fix some of the initial register settings.
* configs/stm32f4discovery/src/up_ssd1289.c: FSMC address bit 16 is used to
distinguish command and data. But FSMC address bits 0-24 correspond to ARM
address bits 1-25, se we need to set bit 17 in order generate output on
FSMC_A16.

4
nuttx/drivers/lcd/ssd1289.c
View File

@ -819,7 +819,7 @@ static int ssd1289_setcontrast(FAR struct lcd_dev_s *dev, unsigned int contrast)
* Name: ssd1289_hwinitialize
*
* Description:
* Set LCD panel contrast (0-CONFIG_LCD_MAXCONTRAST).
* Initialize the LCD hardware.
*
**************************************************************************************/
@ -830,7 +830,7 @@ static inline int ssd1289_hwinitialize(FAR struct ssd1289_dev_s *priv)
uint16_t id;
#endif
/* Select the LCD and home the cursor position */
/* Select the LCD */
lcd->select(lcd);

2
nuttx/fs/nfs/nfs_node.h
View File

@ -123,7 +123,7 @@ struct nfsnode
bool n_open; /* true: The file is (still) open */
uint64_t n_size; /* Current size of file */
struct nfs_fattr n_fattr; /* nfs file attribute cache */
nfstype nfsv3_type; /* File type */
uint32_t nfsv3_type; /* File type */
time_t n_attrstamp; /* Attr. cache timestamp */
struct timespec n_mtime; /* Prev modify time. */
time_t n_ctime; /* Prev create time. */

145
nuttx/fs/nfs/nfs_vfsops.c
View File

@ -176,12 +176,11 @@ const struct mountpt_operations nfs_operations =
* check to see if the type is ok and that deletion is not in progress.
****************************************************************************/
static int
nfs_open(FAR struct file *filep, FAR const char *relpath,
int oflags, mode_t mode)
static int nfs_open(FAR struct file *filep, FAR const char *relpath,
int oflags, mode_t mode)
{
struct inode *in;
struct nfs_fattr vap;
//struct nfs_fattr vap;
struct nfsv3_sattr sp;
struct nfsmount *nmp;
struct nfsnode *np;
@ -220,17 +219,17 @@ nfs_open(FAR struct file *filep, FAR const char *relpath,
again:
nfsstats.rpccnt[NFSPROC_CREATE]++;
memset(&sp, 0, sizeof(struct nfsv3_sattr));
memset(&vap, 0, sizeof(struct nfs_fattr));
vap = nmp->nm_head->n_fattr;
//memset(&vap, 0, sizeof(struct nfs_fattr));
//vap = nmp->nm_head->n_fattr;
sp.sa_modetrue = true;
sp.sa_mode = txdr_unsigned(mode);
sp.sa_uidfalse = nfs_xdrneg1;
sp.sa_gidfalse = nfs_xdrneg1;
sp.sa_sizefalse = nfs_xdrneg1;
sp.sa_atimetype = txdr_unsigned(NFSV3SATTRTIME_TOCLIENT);
sp.sa_mtimetype = txdr_unsigned(NFSV3SATTRTIME_TOCLIENT);
txdr_nfsv3time2(&vap.fa3_atime, &sp.sa_atime);
txdr_nfsv3time2(&vap.fa3_mtime, &sp.sa_mtime);
sp.sa_uidfalse = 0;
sp.sa_gidfalse = 0;
sp.sa_sizefalse = 0;
sp.sa_atimetype = txdr_unsigned(NFSV3SATTRTIME_DONTCHANGE);
sp.sa_mtimetype = txdr_unsigned(NFSV3SATTRTIME_DONTCHANGE);
//txdr_nfsv3time2(&vap.fa3_atime, &sp.sa_atime);
//txdr_nfsv3time2(&vap.fa3_mtime, &sp.sa_mtime);
memset(&create, 0, sizeof(struct CREATE3args));
create.how = sp;
@ -259,7 +258,7 @@ again:
*/
np->n_open = true;
np->nfsv3_type = NFREG;
np->nfsv3_type = fxdr_unsigned(uint32_t, resok.attributes.fa_type);
np->n_fhp = resok.fshandle.handle;
np->n_size = fxdr_hyper(&resok.attributes.fa3_size);
np->n_fattr = resok.attributes;
@ -523,20 +522,19 @@ nfs_write(FAR struct file *filep, const char *buffer, size_t buflen)
nfsstats.rpccnt[NFSPROC_WRITE]++;
memset(&write, 0, sizeof(struct WRITE3args));
write.file = np->nfsv3_type;
write.offset = offset;
write.count = buflen;
write.stable = committed;
write.file = txdr_unsigned(np->nfsv3_type);
write.offset = txdr_unsigned(offset);
write.count = txdr_unsigned(buflen);
write.stable = txdr_unsigned(committed);
memcpy((void *)write.data, userbuffer, buflen);
error = nfs_request(nmp, NFSPROC_WRITE, (FAR const void *)&write, (FAR void *)&resok);
error = nfs_request(nmp, NFSPROC_WRITE, (FAR const void *)&write,
(FAR void *)&resok);
if (error)
{
goto errout_with_semaphore;
}
//bcopy (datareply, &resok, sizeof(struct WRITE3resok));
//resok = (struct WRITE3resok *) datareply;
writesize = resok.count;
if (writesize == 0)
{
@ -591,7 +589,8 @@ static int nfs_opendir(struct inode *mountpt, const char *relpath,
struct fs_dirent_s *dir)
{
struct nfsmount *nmp;
//struct romfs_dirinfo_s dirinfo;
struct nfsnode *np;
//struct romfs_dirinfo_s dirinfo;
int ret;
fvdbg("relpath: '%s'\n", relpath);
@ -603,6 +602,7 @@ static int nfs_opendir(struct inode *mountpt, const char *relpath,
/* Recover our private data from the inode instance */
nmp = mountpt->i_private;
np = nmp->nm_head;
/* Make sure that the mount is still healthy */
@ -616,9 +616,24 @@ static int nfs_opendir(struct inode *mountpt, const char *relpath,
/* The entry is a directory */
dir->u.nfs.nd_direoffset = false;
dir->u.nfs.cookie[0] = 0;
dir->u.nfs.cookie[1] = 0;
if (np->nfsv3_type != NFREG && np->nfsv3_type != NFDIR)
{
ndbg("open eacces type=%d\n", np->nfsv3_type);
nfs_semgive(nmp);
return EACCES;
}
if (np->n_flag & NMODIFIED)
{
if (np->nfsv3_type == NFDIR)
{
np->n_direofoffset = 0;
dir->u.nfs.nd_direoffset = false;
dir->u.nfs.cookie[0] = 0;
dir->u.nfs.cookie[1] = 0;
}
}
nfs_semgive(nmp);
return OK;
@ -637,9 +652,8 @@ int nfs_readdirrpc(struct nfsmount *nmp, struct nfsnode *np,
bool end_of_directory, struct fs_dirent_s *dir)
{
int error = 0;
void *datareply;
struct READDIR3args readir;
struct READDIR3resok *resok = NULL;
struct READDIR3resok resok;
/* Loop around doing readdir rpc's of size nm_readdirsize
* truncated to a multiple of NFS_READDIRBLKSIZ.
@ -669,32 +683,30 @@ int nfs_readdirrpc(struct nfsmount *nmp, struct nfsnode *np,
readir.cookieverf.nfsuquad[1] = np->n_cookieverf.nfsuquad[1];
}
error = nfs_request(nmp, NFSPROC_READDIR, &readir, &datareply);
error = nfs_request(nmp, NFSPROC_READDIR, (FAR const void *)&readir,
(FAR void *)&resok);
if (error)
{
goto nfsmout;
}
//bcopy (datareply, &resok, sizeof(struct READDIR3resok));
resok = (struct READDIR3resok *) datareply;
np->n_fattr = resok->dir_attributes;
np->n_cookieverf.nfsuquad[0] = resok->cookieverf.nfsuquad[0];
np->n_cookieverf.nfsuquad[1] = resok->cookieverf.nfsuquad[1];
dir->fd_dir.d_type = resok->reply.entries->fileid;
memcpy(&dir->fd_dir.d_name[NAME_MAX], &resok->reply.entries->name, NAME_MAX);
//dir->fd_dir.d_name = resok->reply.entries->name;//
dir->u.nfs.cookie[0] = resok->reply.entries->cookie.nfsuquad[0];
dir->u.nfs.cookie[1] = resok->reply.entries->cookie.nfsuquad[1];
np->n_fattr = resok.dir_attributes;
np->n_cookieverf.nfsuquad[0] = resok.cookieverf.nfsuquad[0];
np->n_cookieverf.nfsuquad[1] = resok.cookieverf.nfsuquad[1];
dir->fd_dir.d_type = resok.reply.entries->fileid;
memcpy(&dir->fd_dir.d_name[NAME_MAX], &resok.reply.entries->name, NAME_MAX);
//dir->fd_dir.d_name = resok->reply.entries->name;//
dir->u.nfs.cookie[0] = resok.reply.entries->cookie.nfsuquad[0];
dir->u.nfs.cookie[1] = resok.reply.entries->cookie.nfsuquad[1];
if (resok->reply.eof == true)
if (resok.reply.eof == true)
{
end_of_directory = true;
}
//more_dirs = fxdr_unsigned(int, *dp);
//more_dirs = fxdr_unsigned(int, *dp);
/* loop thru the dir entries*/
/* loop thru the dir entries */
/*
while (more_dirs && bigenough)
{
@ -719,7 +731,7 @@ int nfs_readdirrpc(struct nfsmount *nmp, struct nfsnode *np,
/* We are now either at the end of the directory */
if (resok->reply.entries == NULL)
if (resok.reply.entries == NULL)
{
np->n_direofoffset = fxdr_hyper(&dir->u.nfs.cookie[0]);
@ -1083,6 +1095,20 @@ int mountnfs(struct nfs_args *argp, void **handle)
nmp->nm_nam = argp->addr;
nfs_decode_args(nmp, argp);
/* Set up the sockets and per-host congestion */
nmp->nm_sotype = argp->sotype;
nmp->nm_soproto = argp->proto;
/* For Connection based sockets (TCP,...) defer the connect until
* the first request, in case the server is not responding.
*/
if (nmp->nm_sotype == SOCK_DGRAM && (error = nfs_connect(nmp)))
{
goto bad;
}
/* Create an instance of the file private data to describe the opened
* file.
*/
@ -1098,20 +1124,6 @@ int mountnfs(struct nfs_args *argp, void **handle)
np->n_open = true;
nmp->nm_head = np;
/* Set up the sockets and per-host congestion */
nmp->nm_sotype = argp->sotype;
nmp->nm_soproto = argp->proto;
/* For Connection based sockets (TCP,...) defer the connect until
* the first request, in case the server is not responding.
*/
if (nmp->nm_sotype == SOCK_DGRAM && (error = nfs_connect(nmp)))
{
goto bad;
}
/* Mounted! */
nmp->nm_mounted = true;
@ -1335,7 +1347,8 @@ static int nfs_remove(struct inode *mountpt, const char *relpath)
memset(&remove, 0, sizeof(struct REMOVE3args));
remove.object.dir.length = txdr_unsigned(np->n_fhsize);
remove.object.dir.handle = np->n_fhp;
//remove.object.name = relpath;
remove.object.length = txdr_unsigned(64);
strncpy(remove.object.name, relpath, 64);
error = nfs_request(nmp, NFSPROC_REMOVE, (FAR const void *)&remove,
(FAR void*)&resok);
@ -1413,8 +1426,8 @@ static int nfs_mkdir(struct inode *mountpt, const char *relpath, mode_t mode)
sp.sa_uidfalse = 0;
sp.sa_gidfalse = 0;
sp.sa_sizefalse = 0;
sp.sa_atimetype = txdr_unsigned(NFSV3SATTRTIME_TOCLIENT);
sp.sa_mtimetype = txdr_unsigned(NFSV3SATTRTIME_TOCLIENT);
sp.sa_atimetype = txdr_unsigned(NFSV3SATTRTIME_DONTCHANGE);
sp.sa_mtimetype = txdr_unsigned(NFSV3SATTRTIME_DONTCHANGE);
memset(&sp.sa_atime, 0, sizeof(nfstime3));
memset(&sp.sa_mtime, 0, sizeof(nfstime3));
@ -1428,10 +1441,14 @@ static int nfs_mkdir(struct inode *mountpt, const char *relpath, mode_t mode)
goto errout_with_semaphore;
}
np->nfsv3_type = NFDIR;
np->n_fhp = resok.fshandle.handle;
np->n_fattr = resok.obj_attributes;
np->n_flag |= NMODIFIED;
np->n_open = true;
np->nfsv3_type = fxdr_unsigned(uint32_t, resok.obj_attributes.fa_type);
np->n_fhp = resok.fshandle.handle;
np->n_size = fxdr_hyper(&resok.obj_attributes.fa3_size);
np->n_fattr = resok.obj_attributes;
fxdr_nfsv3time(&resok.obj_attributes.fa3_mtime, &np->n_mtime)
np->n_ctime = fxdr_hyper(&resok.obj_attributes.fa3_ctime);
np->n_flag |= NMODIFIED;
NFS_INVALIDATE_ATTRCACHE(np);

35
nuttx/fs/nfs/rpc_clnt.c
View File

@ -1267,7 +1267,7 @@ int rpcclnt_connect(struct rpcclnt *rpc)
if (error)
{
ndbg("psock_connect to ppmap port returns %d", error);
ndbg("psock_connect to PMAP port returns %d", error);
goto bad;
}
@ -1313,7 +1313,8 @@ int rpcclnt_connect(struct rpcclnt *rpc)
goto bad;
}
if ((fxdr_unsigned(uint32_t, mdata.mount.status)) != 0)
error = fxdr_unsigned(uint32_t, mdata.mount.status);
if (error != 0)
{
ndbg("error mounting with the server %d\n", error);
goto bad;
@ -1321,9 +1322,35 @@ int rpcclnt_connect(struct rpcclnt *rpc)
rpc->rc_fh = mdata.mount.fhandle;
/* NFS port in the socket*/
/* Do the RPC to get a dynamic bounding with the server using PMAP.
* NFS port in the socket.
*/
memset(&sdata, 0, sizeof(sdata));
memset(&rdata, 0, sizeof(rdata));
sa->sin_port = htons(PMAPPORT);
error = psock_connect(rpc->rc_so, saddr, sizeof(*saddr));
if (error)
{
ndbg("psock_connect PMAP port returns %d\n", error);
goto bad;
}
sdata.prog = txdr_unsigned(NFS_PROG);
sdata.vers = txdr_unsigned(NFS_VER3);
sdata.proc = txdr_unsigned(IPPROTO_UDP);
sdata.port = 0;
error = rpcclnt_request(rpc, PMAPPROC_GETPORT, PMAPPROG, PMAPVERS,
(void *)&rdata, (FAR const void *)&sdata);
if (error != 0)
{
goto bad;
}
sa->sin_port = htons(fxdr_unsigned(uint32_t, rdata.pmap.port));
sa->sin_port = htons(NFS_PORT);
error = psock_connect(rpc->rc_so, saddr, sizeof(*saddr));
if (error)
{