[PATCH] scsi: sg: protect accesses to 'reserved' page array

[PATCH] scsi: sg: protect accesses to 'reserved' page array

[Todd Poynor]

From: Hannes Reinecke <hare@suse.de>

The 'reserved' page array is used as a short-cut for mapping data,
saving us to allocate pages per request. However, the 'reserved' array
is only capable of holding one request, so this patch introduces a mutex
for protect 'sg_fd' against concurrent accesses.

[toddpoynor@google.com: backport to 3.18-4.9,  fixup for bad ioctl
SG_SET_FORCE_LOW_DMA code removed in later versions and not modified by
the original patch.]

commit 1bc0eb0446158cc76562176b80623aa119afee5b upstream.

Signed-off-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Tested-by: Johannes Thumshirn <jthumshirn@suse.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Todd Poynor <toddpoynor@google.com>
---
Suggested for 4.9, 4.4, and 3.18.  This version includes a trivial
fixup for some bad ioctl code that has been removed.  The fix is
already in 4.12.

Fixes a KASAN use-after-free bug found during multithreaded fuzz testing
of SCSI char device syscalls, where ioctl SG_SET_RESERVED_SIZE frees a
pre-allocated "reserve buffer" in a race with an I/O request that uses
the buffer.

BUG: KASAN: use-after-free in bio_copy_user_iov+xx block/bio.c:1205
Call Trace:
...
 [<ffffffff81e051cf>] bio_copy_user_iov+0xcdf/0xe50 block/bio.c:1205
 [<ffffffff81e3665f>] __blk_rq_map_user_iov block/blk-map.c:56 [inline]
 [<ffffffff81e3665f>] blk_rq_map_user_iov+0x22f/0x770 block/blk-map.c:133
 [<ffffffff81e36ca9>] blk_rq_map_user+0x109/0x180 block/blk-map.c:163
 [<ffffffff8278f982>] sg_start_req drivers/scsi/sg.c:1766 [inline]
 [<ffffffff8278f982>] sg_common_write.isra.21+0xc12/0x17a0 drivers/scsi/sg.c:775
 [<ffffffff82793d1b>] sg_write+0x68b/0xb10 drivers/scsi/sg.c:678

 drivers/scsi/sg.c | 47 ++++++++++++++++++++++++++---------------------
 1 file changed, 26 insertions(+), 21 deletions(-)

diff --git a/drivers/scsi/sg.c b/drivers/scsi/sg.c
index f753df25ba34..cf7e5f096988 100644
--- a/drivers/scsi/sg.c
+++ b/drivers/scsi/sg.c
@@ -142,6 +142,7 @@ typedef struct sg_fd {		/* holds the state of a file descriptor */
 	struct sg_device *parentdp;	/* owning device */
 	wait_queue_head_t read_wait;	/* queue read until command done */
 	rwlock_t rq_list_lock;	/* protect access to list in req_arr */
+	struct mutex f_mutex;	/* protect against changes in this fd */
 	int timeout;		/* defaults to SG_DEFAULT_TIMEOUT      */
 	int timeout_user;	/* defaults to SG_DEFAULT_TIMEOUT_USER */
 	Sg_scatter_hold reserve;	/* buffer held for this file descriptor */
@@ -155,6 +156,7 @@ typedef struct sg_fd {		/* holds the state of a file descriptor */
 	unsigned char next_cmd_len; /* 0: automatic, >0: use on next write() */
 	char keep_orphan;	/* 0 -> drop orphan (def), 1 -> keep for read() */
 	char mmap_called;	/* 0 -> mmap() never called on this fd */
+	char res_in_use;	/* 1 -> 'reserve' array in use */
 	struct kref f_ref;
 	struct execute_work ew;
 } Sg_fd;
@@ -198,7 +200,6 @@ static void sg_remove_sfp(struct kref *);
 static Sg_request *sg_get_rq_mark(Sg_fd * sfp, int pack_id);
 static Sg_request *sg_add_request(Sg_fd * sfp);
 static int sg_remove_request(Sg_fd * sfp, Sg_request * srp);
-static int sg_res_in_use(Sg_fd * sfp);
 static Sg_device *sg_get_dev(int dev);
 static void sg_device_destroy(struct kref *kref);
 
@@ -614,6 +615,7 @@ sg_write(struct file *filp, const char __user *buf, size_t count, loff_t * ppos)
 	}
 	buf += SZ_SG_HEADER;
 	__get_user(opcode, buf);
+	mutex_lock(&sfp->f_mutex);
 	if (sfp->next_cmd_len > 0) {
 		cmd_size = sfp->next_cmd_len;
 		sfp->next_cmd_len = 0;	/* reset so only this write() effected */
@@ -622,6 +624,7 @@ sg_write(struct file *filp, const char __user *buf, size_t count, loff_t * ppos)
 		if ((opcode >= 0xc0) && old_hdr.twelve_byte)
 			cmd_size = 12;
 	}
+	mutex_unlock(&sfp->f_mutex);
 	SCSI_LOG_TIMEOUT(4, sg_printk(KERN_INFO, sdp,
 		"sg_write:   scsi opcode=0x%02x, cmd_size=%d\n", (int) opcode, cmd_size));
 /* Determine buffer size.  */
@@ -721,7 +724,7 @@ sg_new_write(Sg_fd *sfp, struct file *file, const char __user *buf,
 			sg_remove_request(sfp, srp);
 			return -EINVAL;	/* either MMAP_IO or DIRECT_IO (not both) */
 		}
-		if (sg_res_in_use(sfp)) {
+		if (sfp->res_in_use) {
 			sg_remove_request(sfp, srp);
 			return -EBUSY;	/* reserve buffer already being used */
 		}
@@ -892,7 +895,7 @@ sg_ioctl(struct file *filp, unsigned int cmd_in, unsigned long arg)
 			return result;
 		if (val) {
 			sfp->low_dma = 1;
-			if ((0 == sfp->low_dma) && (0 == sg_res_in_use(sfp))) {
+			if ((0 == sfp->low_dma) && !sfp->res_in_use) {
 				val = (int) sfp->reserve.bufflen;
 				sg_remove_scat(sfp, &sfp->reserve);
 				sg_build_reserve(sfp, val);
@@ -967,12 +970,18 @@ sg_ioctl(struct file *filp, unsigned int cmd_in, unsigned long arg)
                         return -EINVAL;
 		val = min_t(int, val,
 			    max_sectors_bytes(sdp->device->request_queue));
+		mutex_lock(&sfp->f_mutex);
 		if (val != sfp->reserve.bufflen) {
-			if (sg_res_in_use(sfp) || sfp->mmap_called)
+			if (sfp->mmap_called ||
+			    sfp->res_in_use) {
+				mutex_unlock(&sfp->f_mutex);
 				return -EBUSY;
+			}
+
 			sg_remove_scat(sfp, &sfp->reserve);
 			sg_build_reserve(sfp, val);
 		}
+		mutex_unlock(&sfp->f_mutex);
 		return 0;
 	case SG_GET_RESERVED_SIZE:
 		val = min_t(int, sfp->reserve.bufflen,
@@ -1727,13 +1736,22 @@ sg_start_req(Sg_request *srp, unsigned char *cmd)
 		md = &map_data;
 
 	if (md) {
-		if (!sg_res_in_use(sfp) && dxfer_len <= rsv_schp->bufflen)
+		mutex_lock(&sfp->f_mutex);
+		if (dxfer_len <= rsv_schp->bufflen &&
+		    !sfp->res_in_use) {
+			sfp->res_in_use = 1;
 			sg_link_reserve(sfp, srp, dxfer_len);
-		else {
+		} else if ((hp->flags & SG_FLAG_MMAP_IO) && sfp->res_in_use) {
+			mutex_unlock(&sfp->f_mutex);
+			return -EBUSY;
+		} else {
 			res = sg_build_indirect(req_schp, sfp, dxfer_len);
-			if (res)
+			if (res) {
+				mutex_unlock(&sfp->f_mutex);
 				return res;
+			}
 		}
+		mutex_unlock(&sfp->f_mutex);
 
 		md->pages = req_schp->pages;
 		md->page_order = req_schp->page_order;
@@ -2135,6 +2153,7 @@ sg_add_sfp(Sg_device * sdp)
 	rwlock_init(&sfp->rq_list_lock);
 
 	kref_init(&sfp->f_ref);
+	mutex_init(&sfp->f_mutex);
 	sfp->timeout = SG_DEFAULT_TIMEOUT;
 	sfp->timeout_user = SG_DEFAULT_TIMEOUT_USER;
 	sfp->force_packid = SG_DEF_FORCE_PACK_ID;
@@ -2210,20 +2229,6 @@ sg_remove_sfp(struct kref *kref)
 	schedule_work(&sfp->ew.work);
 }
 
-static int
-sg_res_in_use(Sg_fd * sfp)
-{
-	const Sg_request *srp;
-	unsigned long iflags;
-
-	read_lock_irqsave(&sfp->rq_list_lock, iflags);
-	for (srp = sfp->headrp; srp; srp = srp->nextrp)
-		if (srp->res_used)
-			break;
-	read_unlock_irqrestore(&sfp->rq_list_lock, iflags);
-	return srp ? 1 : 0;
-}
-
 #ifdef CONFIG_SCSI_PROC_FS
 static int
 sg_idr_max_id(int id, void *p, void *data)
-- 
2.14.1.480.gb18f417b89-goog

Re: [PATCH] scsi: sg: protect accesses to 'reserved' page array

[Todd Poynor]

Ah, this patch also requires another follow-on fix:
>
> commit e791ce27c3f6a1d3c746fd6a8f8e36c9540ec6f9
> Author: Hannes Reinecke <hare@suse.de>
> Date:   Mon Apr 24 10:26:36 2017 +0200
>
>     scsi: sg: reset 'res_in_use' after unlinking reserved array
>
>
> On Tue, Aug 15, 2017 at 8:09 PM, Todd Poynor <toddpoynor@google.com> wrote:
>>
>> From: Hannes Reinecke <hare@suse.de>
>>
>> The 'reserved' page array is used as a short-cut for mapping data,
>> saving us to allocate pages per request. However, the 'reserved' array
>> is only capable of holding one request, so this patch introduces a mutex
>> for protect 'sg_fd' against concurrent accesses.
>>
>> [toddpoynor@google.com: backport to 3.18-4.9,  fixup for bad ioctl
>> SG_SET_FORCE_LOW_DMA code removed in later versions and not modified by
>> the original patch.]
>>
>> commit 1bc0eb0446158cc76562176b80623aa119afee5b upstream.
>>
>> Signed-off-by: Hannes Reinecke <hare@suse.com>
>> Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
>> Tested-by: Johannes Thumshirn <jthumshirn@suse.de>
>> Reviewed-by: Christoph Hellwig <hch@lst.de>
>> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
>> Signed-off-by: Todd Poynor <toddpoynor@google.com>
>> ---
>> Suggested for 4.9, 4.4, and 3.18.  This version includes a trivial
>> fixup for some bad ioctl code that has been removed.  The fix is
>> already in 4.12.
>>
>> Fixes a KASAN use-after-free bug found during multithreaded fuzz testing
>> of SCSI char device syscalls, where ioctl SG_SET_RESERVED_SIZE frees a
>> pre-allocated "reserve buffer" in a race with an I/O request that uses
>> the buffer.
>>
>> BUG: KASAN: use-after-free in bio_copy_user_iov+xx block/bio.c:1205
>> Call Trace:
>> ...
>>  [<ffffffff81e051cf>] bio_copy_user_iov+0xcdf/0xe50 block/bio.c:1205
>>  [<ffffffff81e3665f>] __blk_rq_map_user_iov block/blk-map.c:56 [inline]
>>  [<ffffffff81e3665f>] blk_rq_map_user_iov+0x22f/0x770 block/blk-map.c:133
>>  [<ffffffff81e36ca9>] blk_rq_map_user+0x109/0x180 block/blk-map.c:163
>>  [<ffffffff8278f982>] sg_start_req drivers/scsi/sg.c:1766 [inline]
>>  [<ffffffff8278f982>] sg_common_write.isra.21+0xc12/0x17a0
>> drivers/scsi/sg.c:775
>>  [<ffffffff82793d1b>] sg_write+0x68b/0xb10 drivers/scsi/sg.c:678
>>
>>  drivers/scsi/sg.c | 47 ++++++++++++++++++++++++++---------------------
>>  1 file changed, 26 insertions(+), 21 deletions(-)
>>
>> diff --git a/drivers/scsi/sg.c b/drivers/scsi/sg.c
>> index f753df25ba34..cf7e5f096988 100644
>> --- a/drivers/scsi/sg.c
>> +++ b/drivers/scsi/sg.c
>> @@ -142,6 +142,7 @@ typedef struct sg_fd {              /* holds the state
>> of a file descriptor */
>>         struct sg_device *parentdp;     /* owning device */
>>         wait_queue_head_t read_wait;    /* queue read until command done
>> */
>>         rwlock_t rq_list_lock;  /* protect access to list in req_arr */
>> +       struct mutex f_mutex;   /* protect against changes in this fd */
>>         int timeout;            /* defaults to SG_DEFAULT_TIMEOUT      */
>>         int timeout_user;       /* defaults to SG_DEFAULT_TIMEOUT_USER */
>>         Sg_scatter_hold reserve;        /* buffer held for this file
>> descriptor */
>> @@ -155,6 +156,7 @@ typedef struct sg_fd {              /* holds the state
>> of a file descriptor */
>>         unsigned char next_cmd_len; /* 0: automatic, >0: use on next
>> write() */
>>         char keep_orphan;       /* 0 -> drop orphan (def), 1 -> keep for
>> read() */
>>         char mmap_called;       /* 0 -> mmap() never called on this fd */
>> +       char res_in_use;        /* 1 -> 'reserve' array in use */
>>         struct kref f_ref;
>>         struct execute_work ew;
>>  } Sg_fd;
>> @@ -198,7 +200,6 @@ static void sg_remove_sfp(struct kref *);
>>  static Sg_request *sg_get_rq_mark(Sg_fd * sfp, int pack_id);
>>  static Sg_request *sg_add_request(Sg_fd * sfp);
>>  static int sg_remove_request(Sg_fd * sfp, Sg_request * srp);
>> -static int sg_res_in_use(Sg_fd * sfp);
>>  static Sg_device *sg_get_dev(int dev);
>>  static void sg_device_destroy(struct kref *kref);
>>
>> @@ -614,6 +615,7 @@ sg_write(struct file *filp, const char __user *buf,
>> size_t count, loff_t * ppos)
>>         }
>>         buf += SZ_SG_HEADER;
>>         __get_user(opcode, buf);
>> +       mutex_lock(&sfp->f_mutex);
>>         if (sfp->next_cmd_len > 0) {
>>                 cmd_size = sfp->next_cmd_len;
>>                 sfp->next_cmd_len = 0;  /* reset so only this write()
>> effected */
>> @@ -622,6 +624,7 @@ sg_write(struct file *filp, const char __user *buf,
>> size_t count, loff_t * ppos)
>>                 if ((opcode >= 0xc0) && old_hdr.twelve_byte)
>>                         cmd_size = 12;
>>         }
>> +       mutex_unlock(&sfp->f_mutex);
>>         SCSI_LOG_TIMEOUT(4, sg_printk(KERN_INFO, sdp,
>>                 "sg_write:   scsi opcode=0x%02x, cmd_size=%d\n", (int)
>> opcode, cmd_size));
>>  /* Determine buffer size.  */
>> @@ -721,7 +724,7 @@ sg_new_write(Sg_fd *sfp, struct file *file, const char
>> __user *buf,
>>                         sg_remove_request(sfp, srp);
>>                         return -EINVAL; /* either MMAP_IO or DIRECT_IO
>> (not both) */
>>                 }
>> -               if (sg_res_in_use(sfp)) {
>> +               if (sfp->res_in_use) {
>>                         sg_remove_request(sfp, srp);
>>                         return -EBUSY;  /* reserve buffer already being
>> used */
>>                 }
>> @@ -892,7 +895,7 @@ sg_ioctl(struct file *filp, unsigned int cmd_in,
>> unsigned long arg)
>>                         return result;
>>                 if (val) {
>>                         sfp->low_dma = 1;
>> -                       if ((0 == sfp->low_dma) && (0 ==
>> sg_res_in_use(sfp))) {
>> +                       if ((0 == sfp->low_dma) && !sfp->res_in_use) {
>>                                 val = (int) sfp->reserve.bufflen;
>>                                 sg_remove_scat(sfp, &sfp->reserve);
>>                                 sg_build_reserve(sfp, val);
>> @@ -967,12 +970,18 @@ sg_ioctl(struct file *filp, unsigned int cmd_in,
>> unsigned long arg)
>>                          return -EINVAL;
>>                 val = min_t(int, val,
>>
>> max_sectors_bytes(sdp->device->request_queue));
>> +               mutex_lock(&sfp->f_mutex);
>>                 if (val != sfp->reserve.bufflen) {
>> -                       if (sg_res_in_use(sfp) || sfp->mmap_called)
>> +                       if (sfp->mmap_called ||
>> +                           sfp->res_in_use) {
>> +                               mutex_unlock(&sfp->f_mutex);
>>                                 return -EBUSY;
>> +                       }
>> +
>>                         sg_remove_scat(sfp, &sfp->reserve);
>>                         sg_build_reserve(sfp, val);
>>                 }
>> +               mutex_unlock(&sfp->f_mutex);
>>                 return 0;
>>         case SG_GET_RESERVED_SIZE:
>>                 val = min_t(int, sfp->reserve.bufflen,
>> @@ -1727,13 +1736,22 @@ sg_start_req(Sg_request *srp, unsigned char *cmd)
>>                 md = &map_data;
>>
>>         if (md) {
>> -               if (!sg_res_in_use(sfp) && dxfer_len <= rsv_schp->bufflen)
>> +               mutex_lock(&sfp->f_mutex);
>> +               if (dxfer_len <= rsv_schp->bufflen &&
>> +                   !sfp->res_in_use) {
>> +                       sfp->res_in_use = 1;
>>                         sg_link_reserve(sfp, srp, dxfer_len);
>> -               else {
>> +               } else if ((hp->flags & SG_FLAG_MMAP_IO) &&
>> sfp->res_in_use) {
>> +                       mutex_unlock(&sfp->f_mutex);
>> +                       return -EBUSY;
>> +               } else {
>>                         res = sg_build_indirect(req_schp, sfp, dxfer_len);
>> -                       if (res)
>> +                       if (res) {
>> +                               mutex_unlock(&sfp->f_mutex);
>>                                 return res;
>> +                       }
>>                 }
>> +               mutex_unlock(&sfp->f_mutex);
>>
>>                 md->pages = req_schp->pages;
>>                 md->page_order = req_schp->page_order;
>> @@ -2135,6 +2153,7 @@ sg_add_sfp(Sg_device * sdp)
>>         rwlock_init(&sfp->rq_list_lock);
>>
>>         kref_init(&sfp->f_ref);
>> +       mutex_init(&sfp->f_mutex);
>>         sfp->timeout = SG_DEFAULT_TIMEOUT;
>>         sfp->timeout_user = SG_DEFAULT_TIMEOUT_USER;
>>         sfp->force_packid = SG_DEF_FORCE_PACK_ID;
>> @@ -2210,20 +2229,6 @@ sg_remove_sfp(struct kref *kref)
>>         schedule_work(&sfp->ew.work);
>>  }
>>
>> -static int
>> -sg_res_in_use(Sg_fd * sfp)
>> -{
>> -       const Sg_request *srp;
>> -       unsigned long iflags;
>> -
>> -       read_lock_irqsave(&sfp->rq_list_lock, iflags);
>> -       for (srp = sfp->headrp; srp; srp = srp->nextrp)
>> -               if (srp->res_used)
>> -                       break;
>> -       read_unlock_irqrestore(&sfp->rq_list_lock, iflags);
>> -       return srp ? 1 : 0;
>> -}
>> -
>>  #ifdef CONFIG_SCSI_PROC_FS
>>  static int
>>  sg_idr_max_id(int id, void *p, void *data)
>> --
>> 2.14.1.480.gb18f417b89-goog
>>
>

Re: [PATCH] scsi: sg: protect accesses to 'reserved' page array

[Greg KH]

On Tue, Aug 15, 2017 at 08:32:06PM -0700, Todd Poynor wrote:
> Ah, this patch also requires another follow-on fix:
> >
> > commit e791ce27c3f6a1d3c746fd6a8f8e36c9540ec6f9
> > Author: Hannes Reinecke <hare@suse.de>
> > Date:   Mon Apr 24 10:26:36 2017 +0200
> >
> >     scsi: sg: reset 'res_in_use' after unlinking reserved array
> >

Thanks, both now queued up.

greg k-h

Re: [PATCH] scsi: sg: protect accesses to 'reserved' page array

[Todd Poynor]

Thanks!  A couple more follow-on fixes for sg race conditions are now
in the linux-next tree (via linux-scsi), in case it's appropriate to
also queue these for linux-stable trees now or after 4.14 release:

commit 8d26f491116feaa0b16de370b6a7ba40a40fa0b4
Author: Todd Poynor <toddpoynor@google.com>
Date:   Tue Aug 15 21:48:43 2017 -0700

    scsi: sg: recheck MMAP_IO request length with lock held

commit 6a8dadcca81fceff9976e8828cceb072873b7bd5
Author: Todd Poynor <toddpoynor@google.com>
Date:   Tue Aug 15 22:41:08 2017 -0700

    scsi: sg: protect against races between mmap() and SG_SET_RESERVED_SIZE


Todd

Re: [PATCH] scsi: sg: protect accesses to 'reserved' page array

[Greg KH]

On Thu, Aug 31, 2017 at 01:11:18PM -0700, Todd Poynor wrote:
> Thanks!  A couple more follow-on fixes for sg race conditions are now
> in the linux-next tree (via linux-scsi), in case it's appropriate to
> also queue these for linux-stable trees now or after 4.14 release:
> 
> commit 8d26f491116feaa0b16de370b6a7ba40a40fa0b4
> Author: Todd Poynor <toddpoynor@google.com>
> Date:   Tue Aug 15 21:48:43 2017 -0700
> 
>     scsi: sg: recheck MMAP_IO request length with lock held
> 
> commit 6a8dadcca81fceff9976e8828cceb072873b7bd5
> Author: Todd Poynor <toddpoynor@google.com>
> Date:   Tue Aug 15 22:41:08 2017 -0700
> 
>     scsi: sg: protect against races between mmap() and SG_SET_RESERVED_SIZE
> 

As per:
    https://www.kernel.org/doc/html/latest/process/stable-kernel-rules.html

I have to wait until they hit Linus's tree, and ideally, show up in a
-rc release, before I can add them to a stable tree.

thanks,

greg k-h

Re: [PATCH] scsi: sg: protect accesses to 'reserved' page array

[Greg KH]

On Fri, Sep 01, 2017 at 07:03:55AM +0200, Greg KH wrote:
> On Thu, Aug 31, 2017 at 01:11:18PM -0700, Todd Poynor wrote:
> > Thanks!  A couple more follow-on fixes for sg race conditions are now
> > in the linux-next tree (via linux-scsi), in case it's appropriate to
> > also queue these for linux-stable trees now or after 4.14 release:
> > 
> > commit 8d26f491116feaa0b16de370b6a7ba40a40fa0b4
> > Author: Todd Poynor <toddpoynor@google.com>
> > Date:   Tue Aug 15 21:48:43 2017 -0700
> > 
> >     scsi: sg: recheck MMAP_IO request length with lock held
> > 
> > commit 6a8dadcca81fceff9976e8828cceb072873b7bd5
> > Author: Todd Poynor <toddpoynor@google.com>
> > Date:   Tue Aug 15 22:41:08 2017 -0700
> > 
> >     scsi: sg: protect against races between mmap() and SG_SET_RESERVED_SIZE
> > 
> 
> As per:
>     https://www.kernel.org/doc/html/latest/process/stable-kernel-rules.html
> 
> I have to wait until they hit Linus's tree, and ideally, show up in a
> -rc release, before I can add them to a stable tree.

Now all queued up.

thanks,

greg k-h