Patch would not apply so manually updated it and then merged to cifs-2.6.git for-next Please doublecheck that ok (see attached) On Wed, Dec 19, 2018 at 11:37 AM Pavel Shilovskiy <pshilov@xxxxxxxxxxxxx> wrote: > > Looks good. Thanks! > > Reviewed-by: Pavel Shilovsky <pshilov@xxxxxxxxxxxxx> > > -----Original Message----- > From: Georgy Bystrenin <gkot@xxxxxxxxxx> > Sent: Wednesday, December 19, 2018 9:15 AM > To: linux-cifs@xxxxxxxxxxxxxxx > Cc: Pavel Shilovskiy <pshilov@xxxxxxxxxxxxx>; Michael Shigorin <mike@xxxxxxxxxxxx>; Evgeny Sinelnikov <sin@xxxxxxxxxxxx> > Subject: [PATCH] cifs: Fixed OFD locks do not conflict with eachother > > While resolving a bug with locks on samba shares found a strange behavior. > When a file locked by one node and we trying to lock it from another node it fail with errno 5 (EIO) but in that case errno must be set to (EACCES | EAGAIN). > This isn't happening when we try to lock file second time on same node. > In this case it returns EACCES as expected. > Also this issue not reproduces when we use SMB1 protocol (vers=1.0 in mount options). > > Further investigation showed that the mapping from status_to_posix_error is different for SMB1 and SMB2+ implementations. > For SMB1 mapping is [NT_STATUS_LOCK_NOT_GRANTED to ERRlock] > (https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Ftorvalds%2Flinux%2Fblob%2Fmaster%2Ffs%2Fcifs%2Fnetmisc.c%23L66&data=02%7C01%7Cpshilov%40microsoft.com%7C32719cbffb6e4fa3a18208d665d588a8%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C1%7C636808365142084033&sdata=ymTs1OSE2HBAjH%2BcYb0rmsWNxAwp%2Beg23vOE%2B6HaQkU%3D&reserved=0) > but for SMB2+ mapping is [STATUS_LOCK_NOT_GRANTED to -EIO] > (https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Ftorvalds%2Flinux%2Fblob%2Fmaster%2Ffs%2Fcifs%2Fsmb2maperror.c%23L383&data=02%7C01%7Cpshilov%40microsoft.com%7C32719cbffb6e4fa3a18208d665d588a8%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C1%7C636808365142084033&sdata=hK5X5JyOJ0KGduL7Ty4cru51TOLrSaqNMY11Yeywy78%3D&reserved=0) > > Quick changes in SMB2+ mapping from EIO to EACCES has fixed issue. > > BUG: https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Fbugzilla.kernel.org%2Fshow_bug.cgi%3Fid%3D201971&data=02%7C01%7Cpshilov%40microsoft.com%7C32719cbffb6e4fa3a18208d665d588a8%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C1%7C636808365142084033&sdata=D4or6WC3MuSi31KpUfIpE%2BjDC6xSEBjJCMiMNDvC3cM%3D&reserved=0 > > Signed-off-by: Georgy A Bystrenin <gkot@xxxxxxxxxxxx> > --- > fs/cifs/smb2maperror.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/fs/cifs/smb2maperror.c b/fs/cifs/smb2maperror.c index 62c88dfed57b..d7e839cb773f 100644 > --- a/fs/cifs/smb2maperror.c > +++ b/fs/cifs/smb2maperror.c > @@ -378,8 +378,8 @@ static const struct status_to_posix_error smb2_error_map_table[] = { > {STATUS_NONEXISTENT_EA_ENTRY, -EIO, "STATUS_NONEXISTENT_EA_ENTRY"}, > {STATUS_NO_EAS_ON_FILE, -ENODATA, "STATUS_NO_EAS_ON_FILE"}, > {STATUS_EA_CORRUPT_ERROR, -EIO, "STATUS_EA_CORRUPT_ERROR"}, > - {STATUS_FILE_LOCK_CONFLICT, -EIO, "STATUS_FILE_LOCK_CONFLICT"}, > - {STATUS_LOCK_NOT_GRANTED, -EIO, "STATUS_LOCK_NOT_GRANTED"}, > + {STATUS_FILE_LOCK_CONFLICT, -EACCES, "STATUS_FILE_LOCK_CONFLICT"}, > + {STATUS_LOCK_NOT_GRANTED, -EACCES, "STATUS_LOCK_NOT_GRANTED"}, > {STATUS_DELETE_PENDING, -ENOENT, "STATUS_DELETE_PENDING"}, > {STATUS_CTL_FILE_NOT_SUPPORTED, -ENOSYS, > "STATUS_CTL_FILE_NOT_SUPPORTED"}, > -- > 2.17.1 -- Thanks, Steve
From 3e1ae04c4d3713449696d91f7d29347aed1df79b Mon Sep 17 00:00:00 2001 From: Georgy A Bystrenin <gkot@xxxxxxxxxxxx> Date: Fri, 21 Dec 2018 00:11:42 -0600 Subject: [PATCH] cifs: Fixed OFD locks do not conflict with each other While resolving a bug with locks on samba shares found a strange behavior. When a file locked by one node and we trying to lock it from another node it fail with errno 5 (EIO) but in that case errno must be set to (EACCES | EAGAIN). This isn't happening when we try to lock file second time on same node. In this case it returns EACCES as expected. Also this issue not reproduces when we use SMB1 protocol (vers=1.0 in mount options). Further investigation showed that the mapping from status_to_posix_error is different for SMB1 and SMB2+ implementations. For SMB1 mapping is [NT_STATUS_LOCK_NOT_GRANTED to ERRlock] (See fs/cifs/netmisc.c line 66) but for SMB2+ mapping is [STATUS_LOCK_NOT_GRANTED to -EIO] (see fs/cifs/smb2maperror.c line 383) Quick changes in SMB2+ mapping from EIO to EACCES has fixed issue. BUG: https://bugzilla.kernel.org/show_bug.cgi?id=201971 Signed-off-by: Georgy A Bystrenin <gkot@xxxxxxxxxxxx> Reviewed-by: Pavel Shilovsky <pshilov@xxxxxxxxxxxxx> --- fs/cifs/smb2maperror.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/fs/cifs/smb2maperror.c b/fs/cifs/smb2maperror.c index d47b7f5dfa6c..924269cec135 100644 --- a/fs/cifs/smb2maperror.c +++ b/fs/cifs/smb2maperror.c @@ -379,8 +379,8 @@ static const struct status_to_posix_error smb2_error_map_table[] = { {STATUS_NONEXISTENT_EA_ENTRY, -EIO, "STATUS_NONEXISTENT_EA_ENTRY"}, {STATUS_NO_EAS_ON_FILE, -ENODATA, "STATUS_NO_EAS_ON_FILE"}, {STATUS_EA_CORRUPT_ERROR, -EIO, "STATUS_EA_CORRUPT_ERROR"}, - {STATUS_FILE_LOCK_CONFLICT, -EIO, "STATUS_FILE_LOCK_CONFLICT"}, - {STATUS_LOCK_NOT_GRANTED, -EIO, "STATUS_LOCK_NOT_GRANTED"}, + {STATUS_FILE_LOCK_CONFLICT, -EACCES, "STATUS_FILE_LOCK_CONFLICT"}, + {STATUS_LOCK_NOT_GRANTED, -EACCES, "STATUS_LOCK_NOT_GRANTED"}, {STATUS_DELETE_PENDING, -ENOENT, "STATUS_DELETE_PENDING"}, {STATUS_CTL_FILE_NOT_SUPPORTED, -ENOSYS, "STATUS_CTL_FILE_NOT_SUPPORTED"}, -- 2.17.1
