OpenVZ Forum


Home » Mailing lists » Devel » [RFC PATCH 5/5] use next syscall data to predefine the file descriptor value
[RFC PATCH 5/5] use next syscall data to predefine the file descriptor value [message #31741] Tue, 08 July 2008 11:24 Go to next message
Nadia Derbey is currently offline  Nadia Derbey
Messages: 114
Registered: January 2008
Senior Member
From: openvz.org
[PATCH 05/05]

This patch uses the value written into the next_syscall_data proc file
as a target file descriptor for the next file to be opened.

This makes it easy to restart a process with the same fds as the ones it was
using during the checkpoint phase, instead of 1. opening the file, 2. dup2'ing
the open file descriptor.

The following syscalls are impacted if next_syscall_data is set:
. open()
. openat()

Signed-off-by: Nadia Derbey <Nadia.Derbey@bull.net>

---
 fs/open.c |   62 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 61 insertions(+), 1 deletion(-)

Index: linux-2.6.26-rc8-mm1/fs/open.c
===================================================================
--- linux-2.6.26-rc8-mm1.orig/fs/open.c	2008-07-08 12:12:34.000000000 +0200
+++ linux-2.6.26-rc8-mm1/fs/open.c	2008-07-08 13:23:03.000000000 +0200
@@ -974,6 +974,59 @@ struct file *dentry_open(struct dentry *
 EXPORT_SYMBOL(dentry_open);
 
 /*
+ * Marks a given file descriptor entry as busy (should not be busy when this
+ * routine is called.
+ *
+ * files->next_fd is not updated: this lets the potentially created hole be
+ * filled up on next calls to get_unused_fd_flags.
+ *
+ * Returns the specified fd if successful, -errno else.
+ */
+static int get_predefined_fd_flags(int fd, int flags)
+{
+	struct files_struct *files = current->files;
+	int error;
+	struct fdtable *fdt;
+
+	error = -EINVAL;
+	if (fd < 0)
+		goto out;
+
+	error = -EMFILE;
+	if (fd >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur)
+		goto out;
+
+	spin_lock(&files->file_lock);
+	fdt = files_fdtable(files);
+
+	error = expand_files(files, fd);
+	if (error < 0)
+		goto out_unlock;
+
+	error = -EBUSY;
+	if (FD_ISSET(fd, fdt->open_fds))
+		goto out_unlock;
+
+	FD_SET(fd, fdt->open_fds);
+	if (flags & O_CLOEXEC)
+		FD_SET(fd, fdt->close_on_exec);
+	else
+		FD_CLR(fd, fdt->close_on_exec);
+
+	/* Sanity check */
+	if (fdt->fd[fd] != NULL) {
+		printk(KERN_WARNING "get_unused_fd: slot %d not NULL!\n", fd);
+		fdt->fd[fd] = NULL;
+	}
+
+	error = fd;
+out_unlock:
+	spin_unlock(&files->file_lock);
+out:
+	return error;
+}
+
+/*
  * Find an empty file descriptor entry, and mark it busy.
  */
 int get_unused_fd_flags(int flags)
@@ -1088,7 +1141,14 @@ long do_sys_open(int dfd, const char __u
 	int fd = PTR_ERR(tmp);
 
 	if (!IS_ERR(tmp)) {
-		fd = get_unused_fd_flags(flags);
+		if (unlikely(next_data_set(current))) {
+			int next_fd = get_next_data(current);
+
+			fd = get_predefined_fd_flags(next_fd, flags);
+			reset_next_syscall_data(current);
+		} else
+			fd = get_unused_fd_flags(flags);
+
 		if (fd >= 0) {
 			struct file *f = do_filp_open(dfd, tmp, flags, mode);
 			if (IS_ERR(f)) {

--
_______________________________________________
Containers mailing list
Containers@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers
Re: [RFC PATCH 5/5] use next syscall data to predefine the file descriptor value [message #31757 is a reply to message #31741] Tue, 08 July 2008 20:14 Go to previous messageGo to next message
serue is currently offline  serue
Messages: 750
Registered: February 2006
Senior Member
From: openvz.org
Quoting Nadia.Derbey@bull.net (Nadia.Derbey@bull.net):
> [PATCH 05/05]
> 
> This patch uses the value written into the next_syscall_data proc file
> as a target file descriptor for the next file to be opened.
> 
> This makes it easy to restart a process with the same fds as the ones it was
> using during the checkpoint phase, instead of 1. opening the file, 2. dup2'ing
> the open file descriptor.
> 
> The following syscalls are impacted if next_syscall_data is set:
> . open()
> . openat()

Oh, neat, I somehow missed the fact that you had this in your previous
posting  :)

> Signed-off-by: Nadia Derbey <Nadia.Derbey@bull.net>

It'd be nice if the get_predefined_fd_flags() could share a helper
with get_unused_fd_flags() (in particular because the "/* snaity check */
at the end is between a '#if 1' which sounds like it may one day be
removed), but I'm not sure offhand the best way to do that.  So for now

Acked-by: Serge Hallyn <serue@us.ibm.com>

Thanks, Nadia.

Kathy, I'd love to see a -lxc release with this patchset so we can test
it with cryo.

Suka, the open with specified id here might help your simplify your pipe
c/r patches for cryo?

-serge

> ---
>  fs/open.c |   62 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++-
>  1 file changed, 61 insertions(+), 1 deletion(-)
> 
> Index: linux-2.6.26-rc8-mm1/fs/open.c
> ===================================================================
> --- linux-2.6.26-rc8-mm1.orig/fs/open.c	2008-07-08 12:12:34.000000000 +0200
> +++ linux-2.6.26-rc8-mm1/fs/open.c	2008-07-08 13:23:03.000000000 +0200
> @@ -974,6 +974,59 @@ struct file *dentry_open(struct dentry *
>  EXPORT_SYMBOL(dentry_open);
> 
>  /*
> + * Marks a given file descriptor entry as busy (should not be busy when this
> + * routine is called.
> + *
> + * files->next_fd is not updated: this lets the potentially created hole be
> + * filled up on next calls to get_unused_fd_flags.
> + *
> + * Returns the specified fd if successful, -errno else.
> + */
> +static int get_predefined_fd_flags(int fd, int flags)
> +{
> +	struct files_struct *files = current->files;
> +	int error;
> +	struct fdtable *fdt;
> +
> +	error = -EINVAL;
> +	if (fd < 0)
> +		goto out;
> +
> +	error = -EMFILE;
> +	if (fd >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur)
> +		goto out;
> +
> +	spin_lock(&files->file_lock);
> +	fdt = files_fdtable(files);
> +
> +	error = expand_files(files, fd);
> +	if (error < 0)
> +		goto out_unlock;
> +
> +	error = -EBUSY;
> +	if (FD_ISSET(fd, fdt->open_fds))
> +		goto out_unlock;
> +
> +	FD_SET(fd, fdt->open_fds);
> +	if (flags & O_CLOEXEC)
> +		FD_SET(fd, fdt->close_on_exec);
> +	else
> +		FD_CLR(fd, fdt->close_on_exec);
> +
> +	/* Sanity check */
> +	if (fdt->fd[fd] != NULL) {
> +		printk(KERN_WARNING "get_unused_fd: slot %d not NULL!\n", fd);
> +		fdt->fd[fd] = NULL;
> +	}
> +
> +	error = fd;
> +out_unlock:
> +	spin_unlock(&files->file_lock);
> +out:
> +	return error;
> +}
> +
> +/*
>   * Find an empty file descriptor entry, and mark it busy.
>   */
>  int get_unused_fd_flags(int flags)
> @@ -1088,7 +1141,14 @@ long do_sys_open(int dfd, const char __u
>  	int fd = PTR_ERR(tmp);
> 
>  	if (!IS_ERR(tmp)) {
> -		fd = get_unused_fd_flags(flags);
> +		if (unlikely(next_data_set(current))) {
> +			int next_fd = get_next_data(current);
> +
> +			fd = get_predefined_fd_flags(next_fd, flags);
> +			reset_next_syscall_data(current);
> +		} else
> +			fd = get_unused_fd_flags(flags);
> +
>  		if (fd >= 0) {
>  			struct file *f = do_filp_open(dfd, tmp, flags, mode);
>  			if (IS_ERR(f)) {
> 
> --
_______________________________________________
Containers mailing list
Containers@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers
Re: [RFC PATCH 5/5] use next syscall data to predefine the file descriptor value [message #31769 is a reply to message #31757] Wed, 09 July 2008 04:59 Go to previous messageGo to next message
kathys is currently offline  kathys
Messages: 3
Registered: June 2008
Junior Member
From: openvz.org
Hi Nadia,

I am trying with great difficulty to incorporate these patches into the 
existing lxc-tree on 2.6.26-rc8-mm1-lxc1, they are conflicting with a 
number of other patches from checkpoint/. Serge has asked me to include 
them in the next lxc release so I need to know how to make them fit.

I will put out 2.6.26-rc8-mm1-lxc1 without your patches because its 
taking me too long, I will endeavor to include them in the 
2.6.26-rc8-mm1-lxc2, so if you could have a look at them against the 
next release of lxc which I hope to get out by tomorrow (Thursday) 
afternoon.

Thanks,

Kathy

Serge E. Hallyn wrote:
> Quoting Nadia.Derbey@bull.net (Nadia.Derbey@bull.net):
>   
>> [PATCH 05/05]
>>
>> This patch uses the value written into the next_syscall_data proc file
>> as a target file descriptor for the next file to be opened.
>>
>> This makes it easy to restart a process with the same fds as the ones it was
>> using during the checkpoint phase, instead of 1. opening the file, 2. dup2'ing
>> the open file descriptor.
>>
>> The following syscalls are impacted if next_syscall_data is set:
>> . open()
>> . openat()
>>     
>
> Oh, neat, I somehow missed the fact that you had this in your previous
> posting  :)
>
>   
>> Signed-off-by: Nadia Derbey <Nadia.Derbey@bull.net>
>>     
>
> It'd be nice if the get_predefined_fd_flags() could share a helper
> with get_unused_fd_flags() (in particular because the "/* snaity check */
> at the end is between a '#if 1' which sounds like it may one day be
> removed), but I'm not sure offhand the best way to do that.  So for now
>
> Acked-by: Serge Hallyn <serue@us.ibm.com>
>
> Thanks, Nadia.
>
> Kathy, I'd love to see a -lxc release with this patchset so we can test
> it with cryo.
>
> Suka, the open with specified id here might help your simplify your pipe
> c/r patches for cryo?
>
> -serge
>
>   
>> ---
>>  fs/open.c |   62 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++-
>>  1 file changed, 61 insertions(+), 1 deletion(-)
>>
>> Index: linux-2.6.26-rc8-mm1/fs/open.c
>> ===================================================================
>> --- linux-2.6.26-rc8-mm1.orig/fs/open.c	2008-07-08 12:12:34.000000000 +0200
>> +++ linux-2.6.26-rc8-mm1/fs/open.c	2008-07-08 13:23:03.000000000 +0200
>> @@ -974,6 +974,59 @@ struct file *dentry_open(struct dentry *
>>  EXPORT_SYMBOL(dentry_open);
>>
>>  /*
>> + * Marks a given file descriptor entry as busy (should not be busy when this
>> + * routine is called.
>> + *
>> + * files->next_fd is not updated: this lets the potentially created hole be
>> + * filled up on next calls to get_unused_fd_flags.
>> + *
>> + * Returns the specified fd if successful, -errno else.
>> + */
>> +static int get_predefined_fd_flags(int fd, int flags)
>> +{
>> +	struct files_struct *files = current->files;
>> +	int error;
>> +	struct fdtable *fdt;
>> +
>> +	error = -EINVAL;
>> +	if (fd < 0)
>> +		goto out;
>> +
>> +	error = -EMFILE;
>> +	if (fd >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur)
>> +		goto out;
>> +
>> +	spin_lock(&files->file_lock);
>> +	fdt = files_fdtable(files);
>> +
>> +	error = expand_files(files, fd);
>> +	if (error < 0)
>> +		goto out_unlock;
>> +
>> +	error = -EBUSY;
>> +	if (FD_ISSET(fd, fdt->open_fds))
>> +		goto out_unlock;
>> +
>> +	FD_SET(fd, fdt->open_fds);
>> +	if (flags & O_CLOEXEC)
>> +		FD_SET(fd, fdt->close_on_exec);
>> +	else
>> +		FD_CLR(fd, fdt->close_on_exec);
>> +
>> +	/* Sanity check */
>> +	if (fdt->fd[fd] != NULL) {
>> +		printk(KERN_WARNING "get_unused_fd: slot %d not NULL!\n", fd);
>> +		fdt->fd[fd] = NULL;
>> +	}
>> +
>> +	error = fd;
>> +out_unlock:
>> +	spin_unlock(&files->file_lock);
>> +out:
>> +	return error;
>> +}
>> +
>> +/*
>>   * Find an empty file descriptor entry, and mark it busy.
>>   */
>>  int get_unused_fd_flags(int flags)
>> @@ -1088,7 +1141,14 @@ long do_sys_open(int dfd, const char __u
>>  	int fd = PTR_ERR(tmp);
>>
>>  	if (!IS_ERR(tmp)) {
>> -		fd = get_unused_fd_flags(flags);
>> +		if (unlikely(next_data_set(current))) {
>> +			int next_fd = get_next_data(current);
>> +
>> +			fd = get_predefined_fd_flags(next_fd, flags);
>> +			reset_next_syscall_data(current);
>> +		} else
>> +			fd = get_unused_fd_flags(flags);
>> +
>>  		if (fd >= 0) {
>>  			struct file *f = do_filp_open(dfd, tmp, flags, mode);
>>  			if (IS_ERR(f)) {
>>
>> --
>>     
>
>   

_______________________________________________
Containers mailing list
Containers@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers
Re: [RFC PATCH 5/5] use next syscall data to predefine the file descriptor value [message #31789 is a reply to message #31741] Thu, 10 July 2008 00:32 Go to previous messageGo to next message
ebiederm is currently offline  ebiederm
Messages: 1354
Registered: February 2006
Senior Member
From: openvz.org
Nadia.Derbey@bull.net writes:

> [PATCH 05/05]
>
> This patch uses the value written into the next_syscall_data proc file
> as a target file descriptor for the next file to be opened.
>
> This makes it easy to restart a process with the same fds as the ones it was
> using during the checkpoint phase, instead of 1. opening the file, 2. dup2'ing
> the open file descriptor.

As it happens the behavior of open is deterministic.  So if you open
the files in the right order you should not need this.  dup2 is only needed
if there is a gap in the fds used.

Eric
_______________________________________________
Containers mailing list
Containers@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers
Re: [RFC PATCH 5/5] use next syscall data to predefine the file descriptor value [message #31798 is a reply to message #31769] Thu, 10 July 2008 06:12 Go to previous messageGo to next message
Nadia Derbey is currently offline  Nadia Derbey
Messages: 114
Registered: January 2008
Senior Member
From: openvz.org
kathys wrote:
> Hi Nadia,
> 
> I am trying with great difficulty to incorporate these patches into the 
> existing lxc-tree on 2.6.26-rc8-mm1-lxc1, they are conflicting with a 
> number of other patches from checkpoint/.

Kathy,

Is it the same problem as the one we have solved by private e-mail?

Regards,
Nadia

> Serge has asked me to include 
> them in the next lxc release so I need to know how to make them fit.
> 
> I will put out 2.6.26-rc8-mm1-lxc1 without your patches because its 
> taking me too long, I will endeavor to include them in the 
> 2.6.26-rc8-mm1-lxc2, so if you could have a look at them against the 
> next release of lxc which I hope to get out by tomorrow (Thursday) 
> afternoon.
> 
> Thanks,
> 
> Kathy
> 
> Serge E. Hallyn wrote:
> 
>> Quoting Nadia.Derbey@bull.net (Nadia.Derbey@bull.net):
>>  
>>
>>> [PATCH 05/05]
>>>
>>> This patch uses the value written into the next_syscall_data proc file
>>> as a target file descriptor for the next file to be opened.
>>>
>>> This makes it easy to restart a process with the same fds as the ones 
>>> it was
>>> using during the checkpoint phase, instead of 1. opening the file, 2. 
>>> dup2'ing
>>> the open file descriptor.
>>>
>>> The following syscalls are impacted if next_syscall_data is set:
>>> . open()
>>> . openat()
>>>     
>>
>>
>> Oh, neat, I somehow missed the fact that you had this in your previous
>> posting  :)
>>
>>  
>>
>>> Signed-off-by: Nadia Derbey <Nadia.Derbey@bull.net>
>>>     
>>
>>
>> It'd be nice if the get_predefined_fd_flags() could share a helper
>> with get_unused_fd_flags() (in particular because the "/* snaity check */
>> at the end is between a '#if 1' which sounds like it may one day be
>> removed), but I'm not sure offhand the best way to do that.  So for now
>>
>> Acked-by: Serge Hallyn <serue@us.ibm.com>
>>
>> Thanks, Nadia.
>>
>> Kathy, I'd love to see a -lxc release with this patchset so we can test
>> it with cryo.
>>
>> Suka, the open with specified id here might help your simplify your pipe
>> c/r patches for cryo?
>>
>> -serge
>>
>>  
>>
>>> ---
>>>  fs/open.c |   62 
>>> +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++-
>>>  1 file changed, 61 insertions(+), 1 deletion(-)
>>>
>>> Index: linux-2.6.26-rc8-mm1/fs/open.c
>>> ===================================================================
>>> --- linux-2.6.26-rc8-mm1.orig/fs/open.c    2008-07-08 
>>> 12:12:34.000000000 +0200
>>> +++ linux-2.6.26-rc8-mm1/fs/open.c    2008-07-08 13:23:03.000000000 
>>> +0200
>>> @@ -974,6 +974,59 @@ struct file *dentry_open(struct dentry *
>>>  EXPORT_SYMBOL(dentry_open);
>>>
>>>  /*
>>> + * Marks a given file descriptor entry as busy (should not be busy 
>>> when this
>>> + * routine is called.
>>> + *
>>> + * files->next_fd is not updated: this lets the potentially created 
>>> hole be
>>> + * filled up on next calls to get_unused_fd_flags.
>>> + *
>>> + * Returns the specified fd if successful, -errno else.
>>> + */
>>> +static int get_predefined_fd_flags(int fd, int flags)
>>> +{
>>> +    struct files_struct *files = current->files;
>>> +    int error;
>>> +    struct fdtable *fdt;
>>> +
>>> +    error = -EINVAL;
>>> +    if (fd < 0)
>>> +        goto out;
>>> +
>>> +    error = -EMFILE;
>>> +    if (fd >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur)
>>> +        goto out;
>>> +
>>> +    spin_lock(&files->file_lock);
>>> +    fdt = files_fdtable(files);
>>> +
>>> +    error = expand_files(files, fd);
>>> +    if (error < 0)
>>> +        goto out_unlock;
>>> +
>>> +    error = -EBUSY;
>>> +    if (FD_ISSET(fd, fdt->open_fds))
>>> +        goto out_unlock;
>>> +
>>> +    FD_SET(fd, fdt->open_fds);
>>> +    if (flags & O_CLOEXEC)
>>> +        FD_SET(fd, fdt->close_on_exec);
>>> +    else
>>> +        FD_CLR(fd, fdt->close_on_exec);
>>> +
>>> +    /* Sanity check */
>>> +    if (fdt->fd[fd] != NULL) {
>>> +        printk(KERN_WARNING "get_unused_fd: slot %d not NULL!\n", fd);
>>> +        fdt->fd[fd] = NULL;
>>> +    }
>>> +
>>> +    error = fd;
>>> +out_unlock:
>>> +    spin_unlock(&files->file_lock);
>>> +out:
>>> +    return error;
>>> +}
>>> +
>>> +/*
>>>   * Find an empty file descriptor entry, and mark it busy.
>>>   */
>>>  int get_unused_fd_flags(int flags)
>>> @@ -1088,7 +1141,14 @@ long do_sys_open(int dfd, const char __u
>>>      int fd = PTR_ERR(tmp);
>>>
>>>      if (!IS_ERR(tmp)) {
>>> -        fd = get_unused_fd_flags(flags);
>>> +        if (unlikely(next_data_set(current))) {
>>> +            int next_fd = get_next_data(current);
>>> +
>>> +            fd = get_predefined_fd_flags(next_fd, flags);
>>> +            reset_next_syscall_data(current);
>>> +        } else
>>> +            fd = get_unused_fd_flags(flags);
>>> +
>>>          if (fd >= 0) {
>>>              struct file *f = do_filp_open(dfd, tmp, flags, mode);
>>>              if (IS_ERR(f)) {
>>>
>>> -- 
>>>     
>>
>>
>>   
> 
> 
> 
> 


-- 
===============================================================
Name.......... Nadia DERBEY
Organization.. BULL/DT/OSwR&D/Linux
---------------------------------------------------------------
Email......... mailto:Nadia.Derbey@bull.net
Address....... BULL, B.P. 208, 38432 Echirolles Cedex, France
Tel........... (33) 76 29 77 62 [Internal Bull: (229) 77 62]
Telex,Fax..... 980648 F - (33) 76 29 76 00
Internal Bull. Mail: FREC-B1208
===============================================================

_______________________________________________
Containers mailing list
Containers@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers
Re: [RFC PATCH 5/5] use next syscall data to predefine the file descriptor value [message #31799 is a reply to message #31789] Thu, 10 July 2008 06:25 Go to previous messageGo to next message
Nadia Derbey is currently offline  Nadia Derbey
Messages: 114
Registered: January 2008
Senior Member
From: openvz.org
Eric W. Biederman wrote:
> Nadia.Derbey@bull.net writes:
> 
> 
>>[PATCH 05/05]
>>
>>This patch uses the value written into the next_syscall_data proc file
>>as a target file descriptor for the next file to be opened.
>>
>>This makes it easy to restart a process with the same fds as the ones it was
>>using during the checkpoint phase, instead of 1. opening the file, 2. dup2'ing
>>the open file descriptor.
> 
> 
> As it happens the behavior of open is deterministic.  So if you open
> the files in the right order you should not need this.  dup2 is only needed
> if there is a gap in the fds used.
> 

This covers the case where you're checkpointing a process that has
1. opened, say 3 files (fds x, x+1, and x+2)
2. closed fd x+1
--> checkpoint occurs at that point.

During restart, you'll have to only recreate fds x and x+2.

But I'm realizing that this might be what you're calling a gap in the 
fds ;-)

Regards,
Nadia


_______________________________________________
Containers mailing list
Containers@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers
Re: [RFC PATCH 5/5] use next syscall data to predefine the file descriptor value [message #31907 is a reply to message #31798] Mon, 14 July 2008 04:57 Go to previous message
kathys is currently offline  kathys
Messages: 3
Registered: June 2008
Junior Member
From: openvz.org
Nadia Derbey wrote:
> kathys wrote:
>> Hi Nadia,
>>
>> I am trying with great difficulty to incorporate these patches into 
>> the existing lxc-tree on 2.6.26-rc8-mm1-lxc1, they are conflicting 
>> with a number of other patches from checkpoint/.
>
> Kathy,
>
> Is it the same problem as the one we have solved by private e-mail?
>
> Regards,
> Nadia
Hi Nadia, thanks, I think the confusion was that I was working my way 
through and sent a number of emails in the threads telling you what I 
was going to do.   So yes, this is the same issue.   Thankyou for the 
information.   I will re apply the patches and remove the old ones.

Thanks,

Kathy
>
>> Serge has asked me to include them in the next lxc release so I need 
>> to know how to make them fit.
>>
>> I will put out 2.6.26-rc8-mm1-lxc1 without your patches because its 
>> taking me too long, I will endeavor to include them in the 
>> 2.6.26-rc8-mm1-lxc2, so if you could have a look at them against the 
>> next release of lxc which I hope to get out by tomorrow (Thursday) 
>> afternoon.
>>
>> Thanks,
>>
>> Kathy
>>
>> Serge E. Hallyn wrote:
>>
>>> Quoting Nadia.Derbey@bull.net (Nadia.Derbey@bull.net):
>>>  
>>>
>>>> [PATCH 05/05]
>>>>
>>>> This patch uses the value written into the next_syscall_data proc file
>>>> as a target file descriptor for the next file to be opened.
>>>>
>>>> This makes it easy to restart a process with the same fds as the 
>>>> ones it was
>>>> using during the checkpoint phase, instead of 1. opening the file, 
>>>> 2. dup2'ing
>>>> the open file descriptor.
>>>>
>>>> The following syscalls are impacted if next_syscall_data is set:
>>>> . open()
>>>> . openat()
>>>>     
>>>
>>>
>>> Oh, neat, I somehow missed the fact that you had this in your previous
>>> posting  :)
>>>
>>>  
>>>
>>>> Signed-off-by: Nadia Derbey <Nadia.Derbey@bull.net>
>>>>     
>>>
>>>
>>> It'd be nice if the get_predefined_fd_flags() could share a helper
>>> with get_unused_fd_flags() (in particular because the "/* snaity 
>>> check */
>>> at the end is between a '#if 1' which sounds like it may one day be
>>> removed), but I'm not sure offhand the best way to do that.  So for now
>>>
>>> Acked-by: Serge Hallyn <serue@us.ibm.com>
>>>
>>> Thanks, Nadia.
>>>
>>> Kathy, I'd love to see a -lxc release with this patchset so we can test
>>> it with cryo.
>>>
>>> Suka, the open with specified id here might help your simplify your 
>>> pipe
>>> c/r patches for cryo?
>>>
>>> -serge
>>>
>>>  
>>>
>>>> ---
>>>>  fs/open.c |   62 
>>>> +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++-
>>>>  1 file changed, 61 insertions(+), 1 deletion(-)
>>>>
>>>> Index: linux-2.6.26-rc8-mm1/fs/open.c
>>>> ===================================================================
>>>> --- linux-2.6.26-rc8-mm1.orig/fs/open.c    2008-07-08 
>>>> 12:12:34.000000000 +0200
>>>> +++ linux-2.6.26-rc8-mm1/fs/open.c    2008-07-08 13:23:03.000000000 
>>>> +0200
>>>> @@ -974,6 +974,59 @@ struct file *dentry_open(struct dentry *
>>>>  EXPORT_SYMBOL(dentry_open);
>>>>
>>>>  /*
>>>> + * Marks a given file descriptor entry as busy (should not be busy 
>>>> when this
>>>> + * routine is called.
>>>> + *
>>>> + * files->next_fd is not updated: this lets the potentially 
>>>> created hole be
>>>> + * filled up on next calls to get_unused_fd_flags.
>>>> + *
>>>> + * Returns the specified fd if successful, -errno else.
>>>> + */
>>>> +static int get_predefined_fd_flags(int fd, int flags)
>>>> +{
>>>> +    struct files_struct *files = current->files;
>>>> +    int error;
>>>> +    struct fdtable *fdt;
>>>> +
>>>> +    error = -EINVAL;
>>>> +    if (fd < 0)
>>>> +        goto out;
>>>> +
>>>> +    error = -EMFILE;
>>>> +    if (fd >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur)
>>>> +        goto out;
>>>> +
>>>> +    spin_lock(&files->file_lock);
>>>> +    fdt = files_fdtable(files);
>>>> +
>>>> +    error = expand_files(files, fd);
>>>> +    if (error < 0)
>>>> +        goto out_unlock;
>>>> +
>>>> +    error = -EBUSY;
>>>> +    if (FD_ISSET(fd, fdt->open_fds))
>>>> +        goto out_unlock;
>>>> +
>>>> +    FD_SET(fd, fdt->open_fds);
>>>> +    if (flags & O_CLOEXEC)
>>>> +        FD_SET(fd, fdt->close_on_exec);
>>>> +    else
>>>> +        FD_CLR(fd, fdt->close_on_exec);
>>>> +
>>>> +    /* Sanity check */
>>>> +    if (fdt->fd[fd] != NULL) {
>>>> +        printk(KERN_WARNING "get_unused_fd: slot %d not NULL!\n", 
>>>> fd);
>>>> +        fdt->fd[fd] = NULL;
>>>> +    }
>>>> +
>>>> +    error = fd;
>>>> +out_unlock:
>>>> +    spin_unlock(&files->file_lock);
>>>> +out:
>>>> +    return error;
>>>> +}
>>>> +
>>>> +/*
>>>>   * Find an empty file descriptor entry, and mark it busy.
>>>>   */
>>>>  int get_unused_fd_flags(int flags)
>>>> @@ -1088,7 +1141,14 @@ long do_sys_open(int dfd, const char __u
>>>>      int fd = PTR_ERR(tmp);
>>>>
>>>>      if (!IS_ERR(tmp)) {
>>>> -        fd = get_unused_fd_flags(flags);
>>>> +        if (unlikely(next_data_set(current))) {
>>>> +            int next_fd = get_next_data(current);
>>>> +
>>>> +            fd = get_predefined_fd_flags(next_fd, flags);
>>>> +            reset_next_syscall_data(current);
>>>> +        } else
>>>> +            fd = get_unused_fd_flags(flags);
>>>> +
>>>>          if (fd >= 0) {
>>>>              struct file *f = do_filp_open(dfd, tmp, flags, mode);
>>>>              if (IS_ERR(f)) {
>>>>
>>>> -- 
>>>>     
>>>
>>>
>>>   
>>
>>
>>
>>
>
>

_______________________________________________
Containers mailing list
Containers@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers
Previous Topic: [PATCH -mm 3/3] i/o accounting and control
Next Topic: mini-summit conf#
Goto Forum:
  


Current Time: Sat Sep 21 04:44:50 GMT 2019