Message ID | Pine.LNX.4.44L0.1912101622030.1647-100000@iolanthe.rowland.org (mailing list archive) |
---|---|
State | Mainlined |
Commit | 8ec321e96e056de84022c032ffea253431a83c3c |
Delegated to: | Jiri Kosina |
Headers | show |
Series | HID: Fix slab-out-of-bounds read in hid_field_extract | expand |
On Tue, 10 Dec 2019, Alan Stern wrote: > The syzbot fuzzer found a slab-out-of-bounds bug in the HID report > handler. The bug was caused by a report descriptor which included a > field with size 12 bits and count 4899, for a total size of 7349 > bytes. > > The usbhid driver uses at most a single-page 4-KB buffer for reports. > In the test there wasn't any problem about overflowing the buffer, > since only one byte was received from the device. Rather, the bug > occurred when the HID core tried to extract the data from the report > fields, which caused it to try reading data beyond the end of the > allocated buffer. > > This patch fixes the problem by rejecting any report whose total > length exceeds the HID_MAX_BUFFER_SIZE limit (minus one byte to allow > for a possible report index). In theory a device could have a report > longer than that, but if there was such a thing we wouldn't handle it > correctly anyway. > > Reported-and-tested-by: syzbot+09ef48aa58261464b621@syzkaller.appspotmail.com > Signed-off-by: Alan Stern <stern@rowland.harvard.edu> > CC: <stable@vger.kernel.org> Thanks for hunting this down Alan. Applied.
On Wed, 11 Dec 2019, Jiri Kosina wrote: > On Tue, 10 Dec 2019, Alan Stern wrote: > > > The syzbot fuzzer found a slab-out-of-bounds bug in the HID report > > handler. The bug was caused by a report descriptor which included a > > field with size 12 bits and count 4899, for a total size of 7349 > > bytes. > > > > The usbhid driver uses at most a single-page 4-KB buffer for reports. > > In the test there wasn't any problem about overflowing the buffer, > > since only one byte was received from the device. Rather, the bug > > occurred when the HID core tried to extract the data from the report > > fields, which caused it to try reading data beyond the end of the > > allocated buffer. > > > > This patch fixes the problem by rejecting any report whose total > > length exceeds the HID_MAX_BUFFER_SIZE limit (minus one byte to allow > > for a possible report index). In theory a device could have a report > > longer than that, but if there was such a thing we wouldn't handle it > > correctly anyway. > > > > Reported-and-tested-by: syzbot+09ef48aa58261464b621@syzkaller.appspotmail.com > > Signed-off-by: Alan Stern <stern@rowland.harvard.edu> > > CC: <stable@vger.kernel.org> > > Thanks for hunting this down Alan. Applied. I just noticed this code: u8 *hid_alloc_report_buf(struct hid_report *report, gfp_t flags) { /* * 7 extra bytes are necessary to achieve proper functionality * of implement() working on 8 byte chunks */ u32 len = hid_report_len(report) + 7; return kmalloc(len, flags); } Does this indicate that the upper limit on a report length should really be HID_MAX_BUFFER_SIZE - 8 instead of HID_MAX_BUFFER_SIZE - 1? Alan Stern
On Wed, 11 Dec 2019, Alan Stern wrote: > > > The syzbot fuzzer found a slab-out-of-bounds bug in the HID report > > > handler. The bug was caused by a report descriptor which included a > > > field with size 12 bits and count 4899, for a total size of 7349 > > > bytes. > > > > > > The usbhid driver uses at most a single-page 4-KB buffer for reports. > > > In the test there wasn't any problem about overflowing the buffer, > > > since only one byte was received from the device. Rather, the bug > > > occurred when the HID core tried to extract the data from the report > > > fields, which caused it to try reading data beyond the end of the > > > allocated buffer. > > > > > > This patch fixes the problem by rejecting any report whose total > > > length exceeds the HID_MAX_BUFFER_SIZE limit (minus one byte to allow > > > for a possible report index). In theory a device could have a report > > > longer than that, but if there was such a thing we wouldn't handle it > > > correctly anyway. > > > > > > Reported-and-tested-by: syzbot+09ef48aa58261464b621@syzkaller.appspotmail.com > > > Signed-off-by: Alan Stern <stern@rowland.harvard.edu> > > > CC: <stable@vger.kernel.org> > > > > Thanks for hunting this down Alan. Applied. > > I just noticed this code: > > u8 *hid_alloc_report_buf(struct hid_report *report, gfp_t flags) > { > /* > * 7 extra bytes are necessary to achieve proper functionality > * of implement() working on 8 byte chunks > */ > > u32 len = hid_report_len(report) + 7; > > return kmalloc(len, flags); > } > > Does this indicate that the upper limit on a report length should > really be HID_MAX_BUFFER_SIZE - 8 instead of HID_MAX_BUFFER_SIZE - 1? As far as I remember, this is just very lousy way of properly rounding the size up (see 27ce405039bfe). So I believe HID_MAX_BUFFER_SIZE -1 is still functionally correct. Thanks,
Index: usb-devel/drivers/hid/hid-core.c =================================================================== --- usb-devel.orig/drivers/hid/hid-core.c +++ usb-devel/drivers/hid/hid-core.c @@ -268,6 +268,12 @@ static int hid_add_field(struct hid_pars offset = report->size; report->size += parser->global.report_size * parser->global.report_count; + /* Total size check: Allow for possible report index byte */ + if (report->size > (HID_MAX_BUFFER_SIZE - 1) << 3) { + hid_err(parser->device, "report is too long\n"); + return -1; + } + if (!parser->local.usage_index) /* Ignore padding fields */ return 0;
The syzbot fuzzer found a slab-out-of-bounds bug in the HID report handler. The bug was caused by a report descriptor which included a field with size 12 bits and count 4899, for a total size of 7349 bytes. The usbhid driver uses at most a single-page 4-KB buffer for reports. In the test there wasn't any problem about overflowing the buffer, since only one byte was received from the device. Rather, the bug occurred when the HID core tried to extract the data from the report fields, which caused it to try reading data beyond the end of the allocated buffer. This patch fixes the problem by rejecting any report whose total length exceeds the HID_MAX_BUFFER_SIZE limit (minus one byte to allow for a possible report index). In theory a device could have a report longer than that, but if there was such a thing we wouldn't handle it correctly anyway. Reported-and-tested-by: syzbot+09ef48aa58261464b621@syzkaller.appspotmail.com Signed-off-by: Alan Stern <stern@rowland.harvard.edu> CC: <stable@vger.kernel.org> --- [as1926] drivers/hid/hid-core.c | 6 ++++++ 1 file changed, 6 insertions(+)