* Re: [PATCH RFC] brcmfmac: sanitize DMI strings v2
From: Victor Bravo @ 2019-05-06 10:20 UTC (permalink / raw)
To: Hans de Goede
Cc: Arend Van Spriel, Franky Lin, Hante Meuleman, Chi-Hsien Lin,
Wright Feng, Kalle Valo, David S. Miller, linux-wireless,
brcm80211-dev-list.pdl, brcm80211-dev-list, linux-kernel
In-Reply-To: <70677dff-4336-28d5-7ab9-7ba7c3d74ebc@redhat.com>
On Mon, May 06, 2019 at 11:33:20AM +0200, Hans de Goede wrote:
> Hi,
>
> On 06-05-19 11:06, Victor Bravo wrote:
> > On Mon, May 06, 2019 at 10:13:38AM +0200, Hans de Goede wrote:
> > > Hi,
> >
> > Hi,
> >
> > > On 05-05-19 17:03, Victor Bravo wrote:
> > > > Sanitize DMI strings in brcmfmac driver to make resulting filenames
> > > > contain only safe characters. This version replaces all non-printable
> > > > characters incl. delete (0-31, 127-255), spaces and slashes with
> > > > underscores.
> > > >
> > > > This change breaks backward compatibility, but adds control over strings
> > > > passed to firmware loader and compatibility with CONFIG_EXTRA_FIRMWARE
> > > > which doesn't support spaces in filenames.
> > > >
> > > > Changes from v1: don't revert fresh commit by someone else
> > > >
> > > > Signed-off-by: Victor Bravo <1905@spmblk.com>
> > >
> > > Thank you for the patch, but I'm sorry to say this patch cannot go in as is,
> > > because it will break existing systems.
> > >
> > > If you look here:
> > >
> > > https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/brcm
> > >
> > > You will see a file named: "brcmfmac43430a0-sdio.ONDA-V80 PLUS.txt" there, which
> > > has a space in its name (and which works fine).
> >
> > Thanks for the updates. Spaces are actually a problem as files with spaces
> > don't work when built-in with CONFIG_EXTRA_FIRMWARE (which is used with
> > non-modular kernel containing brcmfmac driver).
> >
> > If the DMI string contains slashes, they will cause problems
> > for obvious reasons too.
>
> Right, as said I'm fine with sanitizing the names, so dropping e.g. / chars,
> but replacing space with _ will cause wifi to stop working on Onda V80 Plus devices and
> we have a clear no regressions policy in the kernel.
Please keep in mind that dropping slashes and non-printable characters
might be a regression too, as some users may already be using
intermediate directories or filenames with tabs etc. if their DMI
strings contain these. If such users exist, they will have to rename
their nvram files anyway.
On the other hand, removing just slashes and non-printable characters
(i.e. (*dst < 0x20) || (dst == '/') || (dst == 0x7f)) would allow
the sanitize function to the bit array and go with hardcoded conditions
(especially if those are final and won't need further adjustments in
set of allowed characters
> > > I'm fine with doing some sanitizing of the strings, but replacing spaces with _
> > > breaks existing use-cases (will cause a regression for them) and a space is absolutely
> > > a valid character in a filename and the firmware-loader can deal with this just fine.
> > >
> > > If the code for building firmwares into the kernel cannot deal with spaces then IMHO
> > > that code should be fixed instead. Have you looked into fixing that?
> >
> > Yes, but updating CONFIG_EXTRA_FIRMWARE to support spaces because of
> > this looks much like
>
> <snip off-topic remark>
>
> > Do you really think it's a good idea to propose that in
> > this case?
>
> I think expecting spaces in filenames to just work is quite reasonable, after all
> its been a long time since we've left DOS-es 8.3 filename limitations.
>
> Have you actually looked at how hard it would be to make filenames with spaces work
> with CONFIG_EXTRA_FIRMWARE ?
>
> No matter how you spin it, the space problem is a CONFIG_EXTRA_FIRMWARE bug, not an
> issue with the brcmfmac code.
Well CONFIG_EXTRA_FIRMWARE is defined to use space as filename
separator, so I don't really dare to call that a bug. Also brcmfmac
seems to be only driver requiring support for spaces etc. in firmware
file names, and this requirement seems to be quite fresh.
So to me the proper way to fix this via CONFIG_EXTRA_FIRMWARE seems to
be
1) wait some time until brcfmac's fw filenames with spaces become
de-facto standard and distros are literally full of these.
2) after this happens, propose update of CONFIG_EXTRA_FIRMWARE to
support spaces in filenames, arguing with status quo which coming from 1)
Unfortunately I feel more like a programmer and this seems more like
politics, so I can promise participation in the wait part only right
now.
> > > As for your T100HA example from earlier in this thread, the brcmfmac driver now
> > > also supports getting the firmware from a special EFI nvram variable, which the
> > > T100HA sets, so you do not need to provide a nvram file on the T100HA and things
> > > will still work.
> >
> > I don't really get this. Can you please suggest how do I make the driver
> > use something different than "brcmfmac43340-sdio.txt" or
> > "brcmfmac43340-sdio.ASUSTeK COMPUTER INC.-T100HAN.txt" on T100HAN?
>
> If you leave out either file, then with a recent kernel you should see this
> brcm_info trigger:
>
> brcmf_info("Using nvram EFI variable\n");
>
> So you should see this message when you do:
>
> dmesg | grep "Using nvram EFI variable"
>
> And the wifi on the T100HAN should just work, without needing to do any
> manual config / provide an nvram file in anyway.
>
> I always strive to make hardware just work with Linux and any UEFI x86 machine
> using brcmfmac which provides the necessary nvram EFI variable in its firmware
> should now just work when booting say a Fedora 30 livecd.
>
> The EFI nvram var support has been tested successfully on the following models:
>
> Acer Iconia Tab8 w1-8
> Acer One 10
> Asus T100CHI
> Asus T100HA
> Asus T100TA
> Asus T200TA
> Lenovo Mixx 2 8
> Lenovo Yoga2 tablet 10
Thanks! Wasn't aware of this, will try in the evening.
Regards,
v.
> Regards,
>
> Hans
>
>
>
> > > > diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/dmi.c b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/dmi.c
> > > > index 7535cb0d4ac0..84571e09b465 100644
> > > > --- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/dmi.c
> > > > +++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/dmi.c
> > > > @@ -23,6 +23,14 @@
> > > > /* The DMI data never changes so we can use a static buf for this */
> > > > static char dmi_board_type[128];
> > > > +/* Array of 128 bits representing 7-bit characters allowed in DMI strings. */
> > > > +static unsigned char brcmf_dmi_allowed_chars[] = {
> > > > + 0x00, 0x00, 0x00, 0x00, 0xfe, 0x7f, 0xff, 0xff,
> > > > + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x7f
> > > > +};
> > > > +
> > > > +#define BRCMF_DMI_SAFE_CHAR '_'
> > > > +
> > > > struct brcmf_dmi_data {
> > > > u32 chip;
> > > > u32 chiprev;
> > > > @@ -99,6 +107,15 @@ static const struct dmi_system_id dmi_platform_data[] = {
> > > > {}
> > > > };
> > > > +void brcmf_dmi_sanitize(char *dst, const unsigned char *allowed, char safe)
> > > > +{
> > > > + while (*dst) {
> > > > + if ((*dst < 0) || !(allowed[*dst / 8] & (1 << (*dst % 8))))
> > >
> > > At a first look I have no clue what this code is doing and I honestly do not feel
> > > like figuring it out, this is clever, but IMHO not readable.
> >
> > Understood. The cluless part actually checks corresponding bit
> > in allowed array, which is a bit mask describing what characters
> > are allowed or not.
> >
> > > Please just write this as if (*dst < 0x21 || (*dst > foo && < bar) || etc,
> > > so that a human can actually see in one look what the code is doing.
> > >
> > > You may want to wait for Arend to give his opinion before changing this though,
> > > maybe he likes the code as is.
> > >
> > > Also note that that should be < 0x20 of course, since we need to preserve spaces
> > > as is to avoid a regression.
> >
> > This has been already discussed, spaces are a problem. There even was an
> > opinion that adding the code that doesn't bother with spaces and slashes
> > might be a regression as well.
> >
> > Regards,
> >
> > v.
> >
> > > Regards,
> > >
> > > Hans
> > >
> > >
> > >
> > >
> > >
> > > > + *dst = safe;
> > > > + dst++;
> > > > + }
> > > > +}
> > > > +
> > > > void brcmf_dmi_probe(struct brcmf_mp_device *settings, u32 chip, u32 chiprev)
> > > > {
> > > > const struct dmi_system_id *match;
> > > > @@ -126,6 +143,9 @@ void brcmf_dmi_probe(struct brcmf_mp_device *settings, u32 chip, u32 chiprev)
> > > > if (sys_vendor && product_name) {
> > > > snprintf(dmi_board_type, sizeof(dmi_board_type), "%s-%s",
> > > > sys_vendor, product_name);
> > > > + brcmf_dmi_sanitize(dmi_board_type,
> > > > + brcmf_dmi_allowed_chars,
> > > > + BRCMF_DMI_SAFE_CHAR);
> > > > settings->board_type = dmi_board_type;
> > > > }
> > > > }
> > > >
> > >
>
^ permalink raw reply
* Re: [PATCH RFC] brcmfmac: sanitize DMI strings v2
From: Hans de Goede @ 2019-05-06 10:34 UTC (permalink / raw)
To: Victor Bravo
Cc: Arend Van Spriel, Franky Lin, Hante Meuleman, Chi-Hsien Lin,
Wright Feng, Kalle Valo, David S. Miller, linux-wireless,
brcm80211-dev-list.pdl, brcm80211-dev-list, linux-kernel
In-Reply-To: <20190506102032.3ximjecado4mz62j@localhost>
Hi,
On 06-05-19 12:20, Victor Bravo wrote:
> On Mon, May 06, 2019 at 11:33:20AM +0200, Hans de Goede wrote:
>> Hi,
>>
>> On 06-05-19 11:06, Victor Bravo wrote:
>>> On Mon, May 06, 2019 at 10:13:38AM +0200, Hans de Goede wrote:
>>>> Hi,
>>>
>>> Hi,
>>>
>>>> On 05-05-19 17:03, Victor Bravo wrote:
>>>>> Sanitize DMI strings in brcmfmac driver to make resulting filenames
>>>>> contain only safe characters. This version replaces all non-printable
>>>>> characters incl. delete (0-31, 127-255), spaces and slashes with
>>>>> underscores.
>>>>>
>>>>> This change breaks backward compatibility, but adds control over strings
>>>>> passed to firmware loader and compatibility with CONFIG_EXTRA_FIRMWARE
>>>>> which doesn't support spaces in filenames.
>>>>>
>>>>> Changes from v1: don't revert fresh commit by someone else
>>>>>
>>>>> Signed-off-by: Victor Bravo <1905@spmblk.com>
>>>>
>>>> Thank you for the patch, but I'm sorry to say this patch cannot go in as is,
>>>> because it will break existing systems.
>>>>
>>>> If you look here:
>>>>
>>>> https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/brcm
>>>>
>>>> You will see a file named: "brcmfmac43430a0-sdio.ONDA-V80 PLUS.txt" there, which
>>>> has a space in its name (and which works fine).
>>>
>>> Thanks for the updates. Spaces are actually a problem as files with spaces
>>> don't work when built-in with CONFIG_EXTRA_FIRMWARE (which is used with
>>> non-modular kernel containing brcmfmac driver).
>>>
>>> If the DMI string contains slashes, they will cause problems
>>> for obvious reasons too.
>>
>> Right, as said I'm fine with sanitizing the names, so dropping e.g. / chars,
>> but replacing space with _ will cause wifi to stop working on Onda V80 Plus devices and
>> we have a clear no regressions policy in the kernel.
>
> Please keep in mind that dropping slashes and non-printable characters
> might be a regression too, as some users may already be using
> intermediate directories or filenames with tabs etc. if their DMI
> strings contain these. If such users exist, they will have to rename
> their nvram files anyway.
I consider it very unlikely that such users exist OTOH the Onda V80 PLus
nvram file is actually in linux-firmware upstream, so that e.g. Fedora 30
has working wifi OOTB on the V80 Plus, if we go with your proposal to
also replace spaces, then users with a V80 Plus will all of a sudden have
their wifi stop working.
If Kalle and Arend are in favor of including space in the "bad character"
list then at a minimum we must first get a patch added to linux-firmware
adding an ONDA-V80_PLUS.txt symlink to the ONDA-V80 PLUS.txt file.
> On the other hand, removing just slashes and non-printable characters
> (i.e. (*dst < 0x20) || (dst == '/') || (dst == 0x7f)) would allow
> the sanitize function to the bit array and go with hardcoded conditions
> (especially if those are final and won't need further adjustments in
> set of allowed characters
If we're going to do some filtering, then I suggest we play it safe and also
disallow other chars which may be used as a separator somewhere, specifically
':' and ','.
Currently upstream linux-firmware has these files which rely on the DMI
matching:
brcmfmac4330-sdio.Prowise-PT301.txt
brcmfmac43430-sdio.Hampoo-D2D3_Vi8A1.txt
brcmfmac43430a0-sdio.ONDA-V80 PLUS.txt
The others are either part of the DMI override table for devices with unsuitable
DMI strings like "Default String"; or are device-tree based.
So as long as we don't break those 3 (or break the ONDA one but get a symlink
in place) we can sanitize a bit more then just non-printable and '/'.
Kalle, Arend, what is your opinion on this?
Note I do not expect the ONDA V80 Plus to have a lot of Linux users, but it definitely has some.
Regards,
Hans
>
>>>> I'm fine with doing some sanitizing of the strings, but replacing spaces with _
>>>> breaks existing use-cases (will cause a regression for them) and a space is absolutely
>>>> a valid character in a filename and the firmware-loader can deal with this just fine.
>>>>
>>>> If the code for building firmwares into the kernel cannot deal with spaces then IMHO
>>>> that code should be fixed instead. Have you looked into fixing that?
>>>
>>> Yes, but updating CONFIG_EXTRA_FIRMWARE to support spaces because of
>>> this looks much like
>>
>> <snip off-topic remark>
>>
>>> Do you really think it's a good idea to propose that in
>>> this case?
>>
>> I think expecting spaces in filenames to just work is quite reasonable, after all
>> its been a long time since we've left DOS-es 8.3 filename limitations.
>>
>> Have you actually looked at how hard it would be to make filenames with spaces work
>> with CONFIG_EXTRA_FIRMWARE ?
>>
>> No matter how you spin it, the space problem is a CONFIG_EXTRA_FIRMWARE bug, not an
>> issue with the brcmfmac code.
>
> Well CONFIG_EXTRA_FIRMWARE is defined to use space as filename
> separator, so I don't really dare to call that a bug. Also brcmfmac
> seems to be only driver requiring support for spaces etc. in firmware
> file names, and this requirement seems to be quite fresh.
>
> So to me the proper way to fix this via CONFIG_EXTRA_FIRMWARE seems to
> be
>
> 1) wait some time until brcfmac's fw filenames with spaces become
> de-facto standard and distros are literally full of these.
>
> 2) after this happens, propose update of CONFIG_EXTRA_FIRMWARE to
> support spaces in filenames, arguing with status quo which coming from 1)
>
> Unfortunately I feel more like a programmer and this seems more like
> politics, so I can promise participation in the wait part only right
> now.
>
>>>> As for your T100HA example from earlier in this thread, the brcmfmac driver now
>>>> also supports getting the firmware from a special EFI nvram variable, which the
>>>> T100HA sets, so you do not need to provide a nvram file on the T100HA and things
>>>> will still work.
>>>
>>> I don't really get this. Can you please suggest how do I make the driver
>>> use something different than "brcmfmac43340-sdio.txt" or
>>> "brcmfmac43340-sdio.ASUSTeK COMPUTER INC.-T100HAN.txt" on T100HAN?
>>
>> If you leave out either file, then with a recent kernel you should see this
>> brcm_info trigger:
>>
>> brcmf_info("Using nvram EFI variable\n");
>>
>> So you should see this message when you do:
>>
>> dmesg | grep "Using nvram EFI variable"
>>
>> And the wifi on the T100HAN should just work, without needing to do any
>> manual config / provide an nvram file in anyway.
>>
>> I always strive to make hardware just work with Linux and any UEFI x86 machine
>> using brcmfmac which provides the necessary nvram EFI variable in its firmware
>> should now just work when booting say a Fedora 30 livecd.
>>
>> The EFI nvram var support has been tested successfully on the following models:
>>
>> Acer Iconia Tab8 w1-8
>> Acer One 10
>> Asus T100CHI
>> Asus T100HA
>> Asus T100TA
>> Asus T200TA
>> Lenovo Mixx 2 8
>> Lenovo Yoga2 tablet 10
>
> Thanks! Wasn't aware of this, will try in the evening.
>
> Regards,
> v.
>
>> Regards,
>>
>> Hans
>>
>>
>>
>>>>> diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/dmi.c b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/dmi.c
>>>>> index 7535cb0d4ac0..84571e09b465 100644
>>>>> --- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/dmi.c
>>>>> +++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/dmi.c
>>>>> @@ -23,6 +23,14 @@
>>>>> /* The DMI data never changes so we can use a static buf for this */
>>>>> static char dmi_board_type[128];
>>>>> +/* Array of 128 bits representing 7-bit characters allowed in DMI strings. */
>>>>> +static unsigned char brcmf_dmi_allowed_chars[] = {
>>>>> + 0x00, 0x00, 0x00, 0x00, 0xfe, 0x7f, 0xff, 0xff,
>>>>> + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x7f
>>>>> +};
>>>>> +
>>>>> +#define BRCMF_DMI_SAFE_CHAR '_'
>>>>> +
>>>>> struct brcmf_dmi_data {
>>>>> u32 chip;
>>>>> u32 chiprev;
>>>>> @@ -99,6 +107,15 @@ static const struct dmi_system_id dmi_platform_data[] = {
>>>>> {}
>>>>> };
>>>>> +void brcmf_dmi_sanitize(char *dst, const unsigned char *allowed, char safe)
>>>>> +{
>>>>> + while (*dst) {
>>>>> + if ((*dst < 0) || !(allowed[*dst / 8] & (1 << (*dst % 8))))
>>>>
>>>> At a first look I have no clue what this code is doing and I honestly do not feel
>>>> like figuring it out, this is clever, but IMHO not readable.
>>>
>>> Understood. The cluless part actually checks corresponding bit
>>> in allowed array, which is a bit mask describing what characters
>>> are allowed or not.
>>>
>>>> Please just write this as if (*dst < 0x21 || (*dst > foo && < bar) || etc,
>>>> so that a human can actually see in one look what the code is doing.
>>>>
>>>> You may want to wait for Arend to give his opinion before changing this though,
>>>> maybe he likes the code as is.
>>>>
>>>> Also note that that should be < 0x20 of course, since we need to preserve spaces
>>>> as is to avoid a regression.
>>>
>>> This has been already discussed, spaces are a problem. There even was an
>>> opinion that adding the code that doesn't bother with spaces and slashes
>>> might be a regression as well.
>>>
>>> Regards,
>>>
>>> v.
>>>
>>>> Regards,
>>>>
>>>> Hans
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>> + *dst = safe;
>>>>> + dst++;
>>>>> + }
>>>>> +}
>>>>> +
>>>>> void brcmf_dmi_probe(struct brcmf_mp_device *settings, u32 chip, u32 chiprev)
>>>>> {
>>>>> const struct dmi_system_id *match;
>>>>> @@ -126,6 +143,9 @@ void brcmf_dmi_probe(struct brcmf_mp_device *settings, u32 chip, u32 chiprev)
>>>>> if (sys_vendor && product_name) {
>>>>> snprintf(dmi_board_type, sizeof(dmi_board_type), "%s-%s",
>>>>> sys_vendor, product_name);
>>>>> + brcmf_dmi_sanitize(dmi_board_type,
>>>>> + brcmf_dmi_allowed_chars,
>>>>> + BRCMF_DMI_SAFE_CHAR);
>>>>> settings->board_type = dmi_board_type;
>>>>> }
>>>>> }
>>>>>
>>>>
>>
^ permalink raw reply
* KASAN: slab-out-of-bounds Read in p54u_load_firmware_cb
From: syzbot @ 2019-05-06 11:16 UTC (permalink / raw)
To: andreyknvl, chunkeey, davem, kvalo, linux-kernel, linux-usb,
linux-wireless, netdev, syzkaller-bugs
Hello,
syzbot found the following crash on:
HEAD commit: 43151d6c usb-fuzzer: main usb gadget fuzzer driver
git tree: https://github.com/google/kasan.git usb-fuzzer
console output: https://syzkaller.appspot.com/x/log.txt?x=12c0112ca00000
kernel config: https://syzkaller.appspot.com/x/.config?x=4183eeef650d1234
dashboard link: https://syzkaller.appspot.com/bug?extid=6d237e74cdc13f036473
compiler: gcc (GCC) 9.0.0 20181231 (experimental)
Unfortunately, I don't have any reproducer for this crash yet.
IMPORTANT: if you fix the bug, please add the following tag to the commit:
Reported-by: syzbot+6d237e74cdc13f036473@syzkaller.appspotmail.com
cx231xx 2-1:0.1: New device @ 480 Mbps (2040:b111) with 1 interfaces
cx231xx 2-1:0.1: Not found matching IAD interface
usb 6-1: Direct firmware load for isl3887usb failed with error -2
usb 6-1: Firmware not found.
==================================================================
BUG: KASAN: slab-out-of-bounds in p54u_load_firmware_cb.cold+0x97/0x13a
drivers/net/wireless/intersil/p54/p54usb.c:936
Read of size 8 at addr ffff888099763588 by task kworker/0:1/12
CPU: 0 PID: 12 Comm: kworker/0:1 Not tainted 5.1.0-rc3-319004-g43151d6 #6
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS
Google 01/01/2011
Workqueue: events request_firmware_work_func
Call Trace:
__dump_stack lib/dump_stack.c:77 [inline]
dump_stack+0xe8/0x16e lib/dump_stack.c:113
print_address_description+0x6c/0x236 mm/kasan/report.c:187
kasan_report.cold+0x1a/0x3c mm/kasan/report.c:317
p54u_load_firmware_cb.cold+0x97/0x13a
drivers/net/wireless/intersil/p54/p54usb.c:936
request_firmware_work_func+0x12d/0x249
drivers/base/firmware_loader/main.c:785
process_one_work+0x90f/0x1580 kernel/workqueue.c:2269
worker_thread+0x9b/0xe20 kernel/workqueue.c:2415
kthread+0x313/0x420 kernel/kthread.c:253
ret_from_fork+0x3a/0x50 arch/x86/entry/entry_64.S:352
Allocated by task 7931:
set_track mm/kasan/common.c:87 [inline]
__kasan_kmalloc mm/kasan/common.c:497 [inline]
__kasan_kmalloc.constprop.0+0xbf/0xd0 mm/kasan/common.c:470
slab_post_alloc_hook mm/slab.h:437 [inline]
slab_alloc_node mm/slub.c:2756 [inline]
__kmalloc_node_track_caller+0xf3/0x320 mm/slub.c:4372
__kmalloc_reserve.isra.0+0x3e/0xf0 net/core/skbuff.c:140
__alloc_skb+0xf4/0x5a0 net/core/skbuff.c:208
alloc_skb include/linux/skbuff.h:1058 [inline]
alloc_uevent_skb+0x7b/0x210 lib/kobject_uevent.c:289
uevent_net_broadcast_untagged lib/kobject_uevent.c:325 [inline]
kobject_uevent_net_broadcast lib/kobject_uevent.c:408 [inline]
kobject_uevent_env+0x865/0x13d0 lib/kobject_uevent.c:589
udc_bind_to_driver+0x33e/0x5e0 drivers/usb/gadget/udc/core.c:1357
usb_gadget_probe_driver+0x23f/0x380 drivers/usb/gadget/udc/core.c:1408
fuzzer_ioctl_run drivers/usb/gadget/fuzzer/fuzzer.c:718 [inline]
fuzzer_ioctl+0x15be/0x1d90 drivers/usb/gadget/fuzzer/fuzzer.c:1130
full_proxy_unlocked_ioctl+0x11b/0x180 fs/debugfs/file.c:205
vfs_ioctl fs/ioctl.c:46 [inline]
file_ioctl fs/ioctl.c:509 [inline]
do_vfs_ioctl+0xced/0x12f0 fs/ioctl.c:696
ksys_ioctl+0xa0/0xc0 fs/ioctl.c:713
__do_sys_ioctl fs/ioctl.c:720 [inline]
__se_sys_ioctl fs/ioctl.c:718 [inline]
__x64_sys_ioctl+0x74/0xb0 fs/ioctl.c:718
do_syscall_64+0xcf/0x4f0 arch/x86/entry/common.c:290
entry_SYSCALL_64_after_hwframe+0x49/0xbe
Freed by task 805:
set_track mm/kasan/common.c:87 [inline]
__kasan_slab_free+0x130/0x180 mm/kasan/common.c:459
slab_free_hook mm/slub.c:1429 [inline]
slab_free_freelist_hook+0x5e/0x140 mm/slub.c:1456
slab_free mm/slub.c:3003 [inline]
kfree+0xce/0x290 mm/slub.c:3958
skb_free_head+0x90/0xb0 net/core/skbuff.c:557
skb_release_data+0x543/0x8b0 net/core/skbuff.c:577
skb_release_all+0x4b/0x60 net/core/skbuff.c:631
__kfree_skb net/core/skbuff.c:645 [inline]
consume_skb net/core/skbuff.c:705 [inline]
consume_skb+0xc5/0x2f0 net/core/skbuff.c:699
skb_free_datagram+0x1b/0xf0 net/core/datagram.c:329
netlink_recvmsg+0x663/0xea0 net/netlink/af_netlink.c:2004
sock_recvmsg_nosec net/socket.c:881 [inline]
sock_recvmsg net/socket.c:888 [inline]
sock_recvmsg+0xd1/0x110 net/socket.c:884
___sys_recvmsg+0x278/0x5a0 net/socket.c:2422
__sys_recvmsg+0xee/0x1b0 net/socket.c:2471
do_syscall_64+0xcf/0x4f0 arch/x86/entry/common.c:290
entry_SYSCALL_64_after_hwframe+0x49/0xbe
The buggy address belongs to the object at ffff888099763180
which belongs to the cache kmalloc-1k of size 1024
The buggy address is located 8 bytes to the right of
1024-byte region [ffff888099763180, ffff888099763580)
The buggy address belongs to the page:
page:ffffea000265d800 count:1 mapcount:0 mapping:ffff88812c3f4a00 index:0x0
compound_mapcount: 0
flags: 0xfff00000010200(slab|head)
raw: 00fff00000010200 dead000000000100 dead000000000200 ffff88812c3f4a00
raw: 0000000000000000 00000000000e000e 00000001ffffffff 0000000000000000
page dumped because: kasan: bad access detected
Memory state around the buggy address:
ffff888099763480: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
ffff888099763500: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
> ffff888099763580: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
^
ffff888099763600: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
ffff888099763680: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
==================================================================
---
This bug is generated by a bot. It may contain errors.
See https://goo.gl/tpsmEJ for more information about syzbot.
syzbot engineers can be reached at syzkaller@googlegroups.com.
syzbot will keep track of this bug report. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot.
^ permalink raw reply
* KASAN: use-after-free Read in p54u_load_firmware_cb
From: syzbot @ 2019-05-06 11:16 UTC (permalink / raw)
To: andreyknvl, chunkeey, davem, kvalo, linux-kernel, linux-usb,
linux-wireless, netdev, syzkaller-bugs
Hello,
syzbot found the following crash on:
HEAD commit: 43151d6c usb-fuzzer: main usb gadget fuzzer driver
git tree: https://github.com/google/kasan.git usb-fuzzer
console output: https://syzkaller.appspot.com/x/log.txt?x=142b312ca00000
kernel config: https://syzkaller.appspot.com/x/.config?x=4183eeef650d1234
dashboard link: https://syzkaller.appspot.com/bug?extid=200d4bb11b23d929335f
compiler: gcc (GCC) 9.0.0 20181231 (experimental)
Unfortunately, I don't have any reproducer for this crash yet.
IMPORTANT: if you fix the bug, please add the following tag to the commit:
Reported-by: syzbot+200d4bb11b23d929335f@syzkaller.appspotmail.com
usb 4-1: Direct firmware load for isl3887usb failed with error -2
usb 4-1: Firmware not found.
==================================================================
BUG: KASAN: use-after-free in p54u_load_firmware_cb.cold+0x97/0x13a
drivers/net/wireless/intersil/p54/p54usb.c:936
Read of size 8 at addr ffff888098bf3588 by task kworker/0:1/12
CPU: 0 PID: 12 Comm: kworker/0:1 Not tainted 5.1.0-rc3-319004-g43151d6 #6
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS
Google 01/01/2011
Workqueue: events request_firmware_work_func
Call Trace:
__dump_stack lib/dump_stack.c:77 [inline]
dump_stack+0xe8/0x16e lib/dump_stack.c:113
print_address_description+0x6c/0x236 mm/kasan/report.c:187
kasan_report.cold+0x1a/0x3c mm/kasan/report.c:317
p54u_load_firmware_cb.cold+0x97/0x13a
drivers/net/wireless/intersil/p54/p54usb.c:936
request_firmware_work_func+0x12d/0x249
drivers/base/firmware_loader/main.c:785
process_one_work+0x90f/0x1580 kernel/workqueue.c:2269
worker_thread+0x9b/0xe20 kernel/workqueue.c:2415
kthread+0x313/0x420 kernel/kthread.c:253
ret_from_fork+0x3a/0x50 arch/x86/entry/entry_64.S:352
The buggy address belongs to the page:
page:ffffea000262fcc0 count:0 mapcount:0 mapping:0000000000000000 index:0x0
flags: 0xfff00000000000()
raw: 00fff00000000000 0000000000000000 ffffffff02620101 0000000000000000
raw: 0000000000000000 0000000000000000 00000000ffffffff 0000000000000000
page dumped because: kasan: bad access detected
Memory state around the buggy address:
ffff888098bf3480: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
ffff888098bf3500: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
> ffff888098bf3580: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
^
ffff888098bf3600: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
ffff888098bf3680: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
==================================================================
---
This bug is generated by a bot. It may contain errors.
See https://goo.gl/tpsmEJ for more information about syzbot.
syzbot engineers can be reached at syzkaller@googlegroups.com.
syzbot will keep track of this bug report. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot.
^ permalink raw reply
* Re: [PATCH RFC] brcmfmac: sanitize DMI strings v2
From: Kalle Valo @ 2019-05-06 12:26 UTC (permalink / raw)
To: Hans de Goede
Cc: Victor Bravo, Arend Van Spriel, Franky Lin, Hante Meuleman,
Chi-Hsien Lin, Wright Feng, David S. Miller, linux-wireless,
brcm80211-dev-list.pdl, brcm80211-dev-list, linux-kernel
In-Reply-To: <fb07ae01-4cca-98e7-1c2d-dfdf44909900@redhat.com>
Hans de Goede <hdegoede@redhat.com> writes:
> If we're going to do some filtering, then I suggest we play it safe and also
> disallow other chars which may be used as a separator somewhere, specifically
> ':' and ','.
>
> Currently upstream linux-firmware has these files which rely on the DMI
> matching:
>
> brcmfmac4330-sdio.Prowise-PT301.txt
> brcmfmac43430-sdio.Hampoo-D2D3_Vi8A1.txt
> brcmfmac43430a0-sdio.ONDA-V80 PLUS.txt
>
> The others are either part of the DMI override table for devices with unsuitable
> DMI strings like "Default String"; or are device-tree based.
>
> So as long as we don't break those 3 (or break the ONDA one but get a symlink
> in place) we can sanitize a bit more then just non-printable and '/'.
>
> Kalle, Arend, what is your opinion on this?
>
> Note I do not expect the ONDA V80 Plus to have a lot of Linux users,
> but it definitely has some.
To me having spaces in filenames is a bad idea, but on the other hand we
do have the "don't break existing setups" rule, so it's not so simple. I
vote for not allowing spaces, I think that's the best for the long run,
but don't know what Arend thinks.
Maybe we could do some kind of fallback mechanism, like first trying the
sanitised filename and if that's not found then we try the old filename
with spaces? And if that old filename works we print a big fat warning
that the user should update the file and that the old "filename with
spaces" support is going away soon?
--
Kalle Valo
^ permalink raw reply
* Re: [PATCH RFC] brcmfmac: sanitize DMI strings v2
From: Kalle Valo @ 2019-05-06 12:29 UTC (permalink / raw)
To: Victor Bravo
Cc: Hans de Goede, Arend Van Spriel, Franky Lin, Hante Meuleman,
Chi-Hsien Lin, Wright Feng, David S. Miller, linux-wireless,
brcm80211-dev-list.pdl, brcm80211-dev-list, linux-kernel
In-Reply-To: <20190506091441.wqtccm4n6xxhxom2@localhost>
Victor Bravo <1905@spmblk.com> writes:
> On Mon, May 06, 2019 at 11:42:06AM +0300, Kalle Valo wrote:
>> Hans de Goede <hdegoede@redhat.com> writes:
>>
>> >> @@ -99,6 +107,15 @@ static const struct dmi_system_id dmi_platform_data[] = {
>> >> {}
>> >> };
>> >> +void brcmf_dmi_sanitize(char *dst, const unsigned char *allowed,
>> >> char safe)
>> >> +{
>> >> + while (*dst) {
>> >> + if ((*dst < 0) || !(allowed[*dst / 8] & (1 << (*dst % 8))))
>> >
>> > At a first look I have no clue what this code is doing and I honestly do not feel
>> > like figuring it out, this is clever, but IMHO not readable.
>> >
>> > Please just write this as if (*dst < 0x21 || (*dst > foo && < bar) || etc,
>> > so that a human can actually see in one look what the code is doing.
>>
>> Is there an existing function for sanitising filenames so that we don't
>> need to reinvent the wheel, maybe something like isalnum()?
>
> I would definitely prefer to use existing function, but I didn't find
> any suitable one. Suggestions are welcome.
I didn't find anything either, but hopefully someone knows.
> As for implementation details, the one I posted was optimized for both
> speed and size, and at least in my opinion this bit array driven
> parametric implementation is exactly what is needed here (using a string
> of allowed characters with strchr-style lookups would bring much worse
> complexity, and checking the characters using series of hardcoded if
> conditions could quickly grow to more than those 16 bytes used by the
> array).
But is this really something which should be optimised? This is driver
initialisation, not in some hot path, right? Can you even measure the
difference?
--
Kalle Valo
^ permalink raw reply
* Re: [PATCH v2 4/5] rtw88: fix unassigned rssi_level in rtw_sta_info
From: Kalle Valo @ 2019-05-06 12:35 UTC (permalink / raw)
To: Tony Chuang; +Cc: linux-wireless@vger.kernel.org
In-Reply-To: <F7CD281DE3E379468C6D07993EA72F84D17EB4C9@RTITMBSVM04.realtek.com.tw>
Tony Chuang <yhchuang@realtek.com> writes:
>> -----Original Message-----
>> From: Kalle Valo [mailto:kvalo@codeaurora.org]
>> Sent: Monday, May 06, 2019 4:49 PM
>> To: Tony Chuang
>> Cc: linux-wireless@vger.kernel.org
>> Subject: Re: [PATCH v2 4/5] rtw88: fix unassigned rssi_level in rtw_sta_info
>>
>> <yhchuang@realtek.com> writes:
>>
>> > From: Yan-Hsuan Chuang <yhchuang@realtek.com>
>> >
>> > The new rssi_level should be stored in si, otherwise the rssi_level will
>> > never be updated and get a wrong RA mask, which is calculated by the
>> > rssi level
>> >
>> > Signed-off-by: Yan-Hsuan Chuang <yhchuang@realtek.com>
>>
>> Stanislaw suggested that this should go to 5.2. So what breaks from
>> user's point of view if this is not applied?
>>
>
> If the rssi level remains unchanged, then we could choose wrong ra_mask.
> And some *bad rates* we be chosen by firmware.
> The most hurtful scene would be *noisy environment* such as office, or public.
> The latency would be high and overall throughput would be only half.
> (This was tested, such as 4x Mbps -> 1x Mbps)
Yeah, then this is definitely suitable for 5.2. Could you please resend
the patch and mention the symtomps in the commit log? And mark the patch
as "[PATCH 5.2 v3]" so that I can easily see it's for v5.2, please.
--
Kalle Valo
^ permalink raw reply
* [PATCH] iwlwifi: remove some unnecessary NULL checks
From: Dan Carpenter @ 2019-05-06 12:54 UTC (permalink / raw)
To: Johannes Berg
Cc: Emmanuel Grumbach, Luca Coelho, Intel Linux Wireless, Kalle Valo,
Naftali Goldstein, Sara Sharon, Shaul Triebitz,
Mordechay Goodstein, Liad Kaufman, Gregory Greenman,
linux-wireless, kernel-janitors
These pointers are an offset into the "sta" struct. They're assigned
like this:
const struct ieee80211_sta_vht_cap *vht_cap = &sta->vht_cap;
They're not the first member of the struct (->supp_rates[] is first) so
they can't be NULL.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
---
.../net/wireless/intel/iwlwifi/mvm/rs-fw.c | 23 +++++++++----------
1 file changed, 11 insertions(+), 12 deletions(-)
diff --git a/drivers/net/wireless/intel/iwlwifi/mvm/rs-fw.c b/drivers/net/wireless/intel/iwlwifi/mvm/rs-fw.c
index 659e21b2d4e7..b6fb670d249c 100644
--- a/drivers/net/wireless/intel/iwlwifi/mvm/rs-fw.c
+++ b/drivers/net/wireless/intel/iwlwifi/mvm/rs-fw.c
@@ -101,7 +101,7 @@ static u8 rs_fw_sgi_cw_support(struct ieee80211_sta *sta)
struct ieee80211_sta_he_cap *he_cap = &sta->he_cap;
u8 supp = 0;
- if (he_cap && he_cap->has_he)
+ if (he_cap->has_he)
return 0;
if (ht_cap->cap & IEEE80211_HT_CAP_SGI_20)
@@ -123,12 +123,12 @@ static u16 rs_fw_get_config_flags(struct iwl_mvm *mvm,
struct ieee80211_sta_ht_cap *ht_cap = &sta->ht_cap;
struct ieee80211_sta_vht_cap *vht_cap = &sta->vht_cap;
struct ieee80211_sta_he_cap *he_cap = &sta->he_cap;
- bool vht_ena = vht_cap && vht_cap->vht_supported;
+ bool vht_ena = vht_cap->vht_supported;
u16 flags = 0;
if (mvm->cfg->ht_params->stbc &&
(num_of_ant(iwl_mvm_get_valid_tx_ant(mvm)) > 1)) {
- if (he_cap && he_cap->has_he) {
+ if (he_cap->has_he) {
if (he_cap->he_cap_elem.phy_cap_info[2] &
IEEE80211_HE_PHY_CAP2_STBC_RX_UNDER_80MHZ)
flags |= IWL_TLC_MNG_CFG_FLAGS_STBC_MSK;
@@ -136,15 +136,14 @@ static u16 rs_fw_get_config_flags(struct iwl_mvm *mvm,
if (he_cap->he_cap_elem.phy_cap_info[7] &
IEEE80211_HE_PHY_CAP7_STBC_RX_ABOVE_80MHZ)
flags |= IWL_TLC_MNG_CFG_FLAGS_HE_STBC_160MHZ_MSK;
- } else if ((ht_cap &&
- (ht_cap->cap & IEEE80211_HT_CAP_RX_STBC)) ||
+ } else if ((ht_cap->cap & IEEE80211_HT_CAP_RX_STBC) ||
(vht_ena &&
(vht_cap->cap & IEEE80211_VHT_CAP_RXSTBC_MASK)))
flags |= IWL_TLC_MNG_CFG_FLAGS_STBC_MSK;
}
if (mvm->cfg->ht_params->ldpc &&
- ((ht_cap && (ht_cap->cap & IEEE80211_HT_CAP_LDPC_CODING)) ||
+ ((ht_cap->cap & IEEE80211_HT_CAP_LDPC_CODING) ||
(vht_ena && (vht_cap->cap & IEEE80211_VHT_CAP_RXLDPC))))
flags |= IWL_TLC_MNG_CFG_FLAGS_LDPC_MSK;
@@ -154,7 +153,7 @@ static u16 rs_fw_get_config_flags(struct iwl_mvm *mvm,
IEEE80211_HE_PHY_CAP1_LDPC_CODING_IN_PAYLOAD))
flags &= ~IWL_TLC_MNG_CFG_FLAGS_LDPC_MSK;
- if (he_cap && he_cap->has_he &&
+ if (he_cap->has_he &&
(he_cap->he_cap_elem.phy_cap_info[3] &
IEEE80211_HE_PHY_CAP3_DCM_MAX_CONST_RX_MASK))
flags |= IWL_TLC_MNG_CFG_FLAGS_HE_DCM_NSS_1_MSK;
@@ -293,13 +292,13 @@ static void rs_fw_set_supp_rates(struct ieee80211_sta *sta,
cmd->mode = IWL_TLC_MNG_MODE_NON_HT;
/* HT/VHT rates */
- if (he_cap && he_cap->has_he) {
+ if (he_cap->has_he) {
cmd->mode = IWL_TLC_MNG_MODE_HE;
rs_fw_he_set_enabled_rates(sta, sband, cmd);
- } else if (vht_cap && vht_cap->vht_supported) {
+ } else if (vht_cap->vht_supported) {
cmd->mode = IWL_TLC_MNG_MODE_VHT;
rs_fw_vht_set_enabled_rates(sta, vht_cap, cmd);
- } else if (ht_cap && ht_cap->ht_supported) {
+ } else if (ht_cap->ht_supported) {
cmd->mode = IWL_TLC_MNG_MODE_HT;
cmd->ht_rates[0][0] = cpu_to_le16(ht_cap->mcs.rx_mask[0]);
cmd->ht_rates[1][0] = cpu_to_le16(ht_cap->mcs.rx_mask[1]);
@@ -381,7 +380,7 @@ static u16 rs_fw_get_max_amsdu_len(struct ieee80211_sta *sta)
const struct ieee80211_sta_vht_cap *vht_cap = &sta->vht_cap;
const struct ieee80211_sta_ht_cap *ht_cap = &sta->ht_cap;
- if (vht_cap && vht_cap->vht_supported) {
+ if (vht_cap->vht_supported) {
switch (vht_cap->cap & IEEE80211_VHT_CAP_MAX_MPDU_MASK) {
case IEEE80211_VHT_CAP_MAX_MPDU_LENGTH_11454:
return IEEE80211_MAX_MPDU_LEN_VHT_11454;
@@ -391,7 +390,7 @@ static u16 rs_fw_get_max_amsdu_len(struct ieee80211_sta *sta)
return IEEE80211_MAX_MPDU_LEN_VHT_3895;
}
- } else if (ht_cap && ht_cap->ht_supported) {
+ } else if (ht_cap->ht_supported) {
if (ht_cap->cap & IEEE80211_HT_CAP_MAX_AMSDU)
/*
* agg is offloaded so we need to assume that agg
--
2.18.0
^ permalink raw reply related
* Re: [PATCH RFC] brcmfmac: sanitize DMI strings v2
From: Victor Bravo @ 2019-05-06 14:06 UTC (permalink / raw)
To: Kalle Valo
Cc: Hans de Goede, Arend Van Spriel, Franky Lin, Hante Meuleman,
Chi-Hsien Lin, Wright Feng, David S. Miller, linux-wireless,
brcm80211-dev-list.pdl, brcm80211-dev-list, linux-kernel
In-Reply-To: <878svjvk9q.fsf@codeaurora.org>
On Mon, May 06, 2019 at 03:29:21PM +0300, Kalle Valo wrote:
> Victor Bravo <1905@spmblk.com> writes:
>
> > On Mon, May 06, 2019 at 11:42:06AM +0300, Kalle Valo wrote:
> >> Hans de Goede <hdegoede@redhat.com> writes:
> >>
> >> >> @@ -99,6 +107,15 @@ static const struct dmi_system_id dmi_platform_data[] = {
> >> >> {}
> >> >> };
> >> >> +void brcmf_dmi_sanitize(char *dst, const unsigned char *allowed,
> >> >> char safe)
> >> >> +{
> >> >> + while (*dst) {
> >> >> + if ((*dst < 0) || !(allowed[*dst / 8] & (1 << (*dst % 8))))
> >> >
> >> > At a first look I have no clue what this code is doing and I honestly do not feel
> >> > like figuring it out, this is clever, but IMHO not readable.
> >> >
> >> > Please just write this as if (*dst < 0x21 || (*dst > foo && < bar) || etc,
> >> > so that a human can actually see in one look what the code is doing.
> >>
> >> Is there an existing function for sanitising filenames so that we don't
> >> need to reinvent the wheel, maybe something like isalnum()?
> >
> > I would definitely prefer to use existing function, but I didn't find
> > any suitable one. Suggestions are welcome.
>
> I didn't find anything either, but hopefully someone knows.
>
> > As for implementation details, the one I posted was optimized for both
> > speed and size, and at least in my opinion this bit array driven
> > parametric implementation is exactly what is needed here (using a string
> > of allowed characters with strchr-style lookups would bring much worse
> > complexity, and checking the characters using series of hardcoded if
> > conditions could quickly grow to more than those 16 bytes used by the
> > array).
>
> But is this really something which should be optimised? This is driver
> initialisation, not in some hot path, right? Can you even measure the
> difference?
I didn't measure it, but according to my understanding of "Linux Kernel
Contributor Covenant Code of Conduct Interpretation" this actually
seems to be right according to "everyone involved wants to see the best
possible solution for the overall success of Linux" (but it also depends
on defintion of "best possible" of course).
Honestly, my exact thought behind this was that switch() or if
conditions shouldn't be used here (except simple case handling few
characters) because of above rule.
I also definitely want to avoid reinventing the wheel or even having
duplicate code in the kernel, so I tried to write the routine in a way
that allows it to be used as kernel library function usable in any
driver unless we manage to find existing wheel of similar parameters
elsewhere. That's why allowed and safe are passed as arguments even if
they are available as global varible / macro.
The bitmask approach also still seems useful as it allows to easily
handle any set of allowed characters, while it seems that the set
of unwanted characters may be extended to even more than my patch
handles.
So *in case no suitable existing sanitizing function is found*, the
approach IMHO might be also
1) fine-tuning this kind of sanitizing routine and making it a kernel
"library" function outside the driver, possibly with some useful macros
- e.g. parametric macro to check whether a value is present in a set
(which could also help to make this check more readable as suggested by
Hans), and I would also really like a macro for compile-time initializon
of the bit array from string literal containing allowed characters
(which unfortunately doesn't seem possible in C yet).
2) having only brcmf_dmi_allowed_chars and BRCMF_DMI_SAFE_CHAR in the
driver and use these with the "library" function.
Unfortunately I'm new to kernel development and have no clue whether
it's a good idea to start with such a function inside driver and
eventually moving it later, or start with getting the function in final
location first and delay driver modificatios until it's ready.
Regards,
v.
^ permalink raw reply
* [PATCH v2 1/1] ath10k: Enable MSA region dump support for WCN3990
From: Govind Singh @ 2019-05-06 14:26 UTC (permalink / raw)
To: ath10k; +Cc: linux-wireless, Govind Singh
MSA memory region caries the hw descriptors information.
Dump MSA region in core dump as this is very helpful in debugging
hw issues.
Testing: Tested on WCN3990 HW
Tested FW: WLAN.HL.3.1-00959-QCAHLSWMTPLZ-1
Signed-off-by: Govind Singh <govinds@codeaurora.org>
---
drivers/net/wireless/ath/ath10k/coredump.c | 21 +++++++
drivers/net/wireless/ath/ath10k/coredump.h | 1 +
drivers/net/wireless/ath/ath10k/qmi.c | 6 ++
drivers/net/wireless/ath/ath10k/snoc.c | 67 ++++++++++++++++++++++
drivers/net/wireless/ath/ath10k/snoc.h | 1 +
5 files changed, 96 insertions(+)
diff --git a/drivers/net/wireless/ath/ath10k/coredump.c b/drivers/net/wireless/ath/ath10k/coredump.c
index eadae2f9206b..56d62035c988 100644
--- a/drivers/net/wireless/ath/ath10k/coredump.c
+++ b/drivers/net/wireless/ath/ath10k/coredump.c
@@ -962,6 +962,19 @@ static const struct ath10k_mem_region qca4019_hw10_mem_regions[] = {
},
};
+static const struct ath10k_mem_region wcn399x_hw10_mem_regions[] = {
+ {
+ /* MSA region start is not fixed, hence it is assigned at runtime */
+ .type = ATH10K_MEM_REGION_TYPE_MSA,
+ .len = 0x100000,
+ .name = "DRAM",
+ .section_table = {
+ .sections = NULL,
+ .size = 0,
+ },
+ },
+};
+
static const struct ath10k_hw_mem_layout hw_mem_layouts[] = {
{
.hw_id = QCA6174_HW_1_0_VERSION,
@@ -1059,6 +1072,14 @@ static const struct ath10k_hw_mem_layout hw_mem_layouts[] = {
.size = ARRAY_SIZE(qca4019_hw10_mem_regions),
},
},
+ {
+ .hw_id = WCN3990_HW_1_0_DEV_VERSION,
+ .hw_rev = ATH10K_HW_WCN3990,
+ .region_table = {
+ .regions = wcn399x_hw10_mem_regions,
+ .size = ARRAY_SIZE(wcn399x_hw10_mem_regions),
+ },
+ },
};
static u32 ath10k_coredump_get_ramdump_size(struct ath10k *ar)
diff --git a/drivers/net/wireless/ath/ath10k/coredump.h b/drivers/net/wireless/ath/ath10k/coredump.h
index 5dac653e1649..9802e90483f4 100644
--- a/drivers/net/wireless/ath/ath10k/coredump.h
+++ b/drivers/net/wireless/ath/ath10k/coredump.h
@@ -126,6 +126,7 @@ enum ath10k_mem_region_type {
ATH10K_MEM_REGION_TYPE_IRAM2 = 5,
ATH10K_MEM_REGION_TYPE_IOSRAM = 6,
ATH10K_MEM_REGION_TYPE_IOREG = 7,
+ ATH10K_MEM_REGION_TYPE_MSA = 8,
};
/* Define a section of the region which should be copied. As not all parts
diff --git a/drivers/net/wireless/ath/ath10k/qmi.c b/drivers/net/wireless/ath/ath10k/qmi.c
index ba8f5a8f83d1..47da492a326b 100644
--- a/drivers/net/wireless/ath/ath10k/qmi.c
+++ b/drivers/net/wireless/ath/ath10k/qmi.c
@@ -817,9 +817,15 @@ ath10k_qmi_driver_event_post(struct ath10k_qmi *qmi,
static void ath10k_qmi_event_server_exit(struct ath10k_qmi *qmi)
{
struct ath10k *ar = qmi->ar;
+ struct ath10k_snoc *ar_snoc = ath10k_snoc_priv(ar);
ath10k_qmi_remove_msa_permission(qmi);
ath10k_core_free_board_files(ar);
+ if (!test_bit(ATH10K_SNOC_FLAG_UNREGISTERING, &ar_snoc->flags)) {
+ ath10k_qmi_remove_msa_permission(qmi);
+ ath10k_snoc_fw_crashed_dump(ar);
+ ath10k_qmi_setup_msa_permissions(qmi);
+ }
ath10k_snoc_fw_indication(ar, ATH10K_QMI_EVENT_FW_DOWN_IND);
ath10k_dbg(ar, ATH10K_DBG_QMI, "wifi fw qmi service disconnected\n");
}
diff --git a/drivers/net/wireless/ath/ath10k/snoc.c b/drivers/net/wireless/ath/ath10k/snoc.c
index 0be12996beba..252dd4ee782d 100644
--- a/drivers/net/wireless/ath/ath10k/snoc.c
+++ b/drivers/net/wireless/ath/ath10k/snoc.c
@@ -24,6 +24,7 @@
#include <linux/regulator/consumer.h>
#include "ce.h"
+#include "coredump.h"
#include "debug.h"
#include "hif.h"
#include "htc.h"
@@ -1586,6 +1587,72 @@ static int ath10k_hw_power_off(struct ath10k *ar)
return ret;
}
+static void ath10k_msa_dump_memory(struct ath10k *ar,
+ struct ath10k_fw_crash_data *crash_data)
+{
+ struct ath10k_snoc *ar_snoc = ath10k_snoc_priv(ar);
+ const struct ath10k_hw_mem_layout *mem_layout;
+ const struct ath10k_mem_region *current_region;
+ struct ath10k_dump_ram_data_hdr *hdr;
+ size_t buf_len;
+ u8 *buf;
+
+ lockdep_assert_held(&ar->data_lock);
+
+ if (!crash_data && !crash_data->ramdump_buf)
+ return;
+
+ mem_layout = ath10k_coredump_get_mem_layout(ar);
+ if (!mem_layout)
+ return;
+
+ current_region = &mem_layout->region_table.regions[0];
+
+ buf = crash_data->ramdump_buf;
+ buf_len = crash_data->ramdump_buf_len;
+ memset(buf, 0, buf_len);
+
+ /* Reserve space for the header. */
+ hdr = (void *)buf;
+ buf += sizeof(*hdr);
+ buf_len -= sizeof(*hdr);
+
+ hdr->region_type = cpu_to_le32(current_region->type);
+ hdr->start = cpu_to_le32(ar_snoc->qmi->msa_va);
+ hdr->length = cpu_to_le32(ar_snoc->qmi->msa_mem_size);
+
+ if (current_region->len < ar_snoc->qmi->msa_mem_size) {
+ memcpy(buf, ar_snoc->qmi->msa_va, current_region->len);
+ ath10k_warn(ar, "msa dump length is less than msa size %x, %x\n",
+ current_region->len, ar_snoc->qmi->msa_mem_size);
+ } else {
+ memcpy(buf, ar_snoc->qmi->msa_va, ar_snoc->qmi->msa_mem_size);
+ }
+}
+
+void ath10k_snoc_fw_crashed_dump(struct ath10k *ar)
+{
+ struct ath10k_fw_crash_data *crash_data;
+ char guid[UUID_STRING_LEN + 1];
+
+ spin_lock_bh(&ar->data_lock);
+
+ ar->stats.fw_crash_counter++;
+
+ crash_data = ath10k_coredump_new(ar);
+
+ if (crash_data)
+ scnprintf(guid, sizeof(guid), "%pUl", &crash_data->guid);
+ else
+ scnprintf(guid, sizeof(guid), "n/a");
+
+ ath10k_err(ar, "firmware crashed! (guid %s)\n", guid);
+ ath10k_print_driver_info(ar);
+ ath10k_msa_dump_memory(ar, crash_data);
+
+ spin_unlock_bh(&ar->data_lock);
+}
+
static const struct of_device_id ath10k_snoc_dt_match[] = {
{ .compatible = "qcom,wcn3990-wifi",
.data = &drv_priv,
diff --git a/drivers/net/wireless/ath/ath10k/snoc.h b/drivers/net/wireless/ath/ath10k/snoc.h
index 25383de8f17d..6d28a6290a94 100644
--- a/drivers/net/wireless/ath/ath10k/snoc.h
+++ b/drivers/net/wireless/ath/ath10k/snoc.h
@@ -101,5 +101,6 @@ static inline struct ath10k_snoc *ath10k_snoc_priv(struct ath10k *ar)
}
int ath10k_snoc_fw_indication(struct ath10k *ar, u64 type);
+void ath10k_snoc_fw_crashed_dump(struct ath10k *ar);
#endif /* _SNOC_H_ */
--
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
a Linux Foundation Collaborative Project
^ permalink raw reply related
* Re: [PATCH RFC] brcmfmac: sanitize DMI strings v2
From: Victor Bravo @ 2019-05-06 15:24 UTC (permalink / raw)
To: Kalle Valo
Cc: Hans de Goede, Arend Van Spriel, Franky Lin, Hante Meuleman,
Chi-Hsien Lin, Wright Feng, David S. Miller, linux-wireless,
brcm80211-dev-list.pdl, brcm80211-dev-list, linux-kernel
In-Reply-To: <87d0kvvkej.fsf@codeaurora.org>
On Mon, May 06, 2019 at 03:26:28PM +0300, Kalle Valo wrote:
> Hans de Goede <hdegoede@redhat.com> writes:
>
> > If we're going to do some filtering, then I suggest we play it safe and also
> > disallow other chars which may be used as a separator somewhere, specifically
> > ':' and ','.
> >
> > Currently upstream linux-firmware has these files which rely on the DMI
> > matching:
> >
> > brcmfmac4330-sdio.Prowise-PT301.txt
> > brcmfmac43430-sdio.Hampoo-D2D3_Vi8A1.txt
> > brcmfmac43430a0-sdio.ONDA-V80 PLUS.txt
> >
> > The others are either part of the DMI override table for devices with unsuitable
> > DMI strings like "Default String"; or are device-tree based.
> >
> > So as long as we don't break those 3 (or break the ONDA one but get a symlink
> > in place) we can sanitize a bit more then just non-printable and '/'.
> >
> > Kalle, Arend, what is your opinion on this?
> >
> > Note I do not expect the ONDA V80 Plus to have a lot of Linux users,
> > but it definitely has some.
>
> To me having spaces in filenames is a bad idea, but on the other hand we
> do have the "don't break existing setups" rule, so it's not so simple. I
> vote for not allowing spaces, I think that's the best for the long run,
> but don't know what Arend thinks.
I have found a fresh judicate on this:
https://lkml.org/lkml/2018/12/22/221
It seems clear that we have to support at least spaces for some time
(maybe wih separate config option which will be deprecated but on by
defaut until old files are considered gone).
> Maybe we could do some kind of fallback mechanism, like first trying the
> sanitised filename and if that's not found then we try the old filename
> with spaces? And if that old filename works we print a big fat warning
> that the user should update the file and that the old "filename with
> spaces" support is going away soon?
In case of parametric sanitizing function, this might be achievable
by sanitizing using "final" character set first, and falling back
to "compatible" character set on file not found. So this may actually
bring another requirement on the sanitizing function.
Regards,
v.
^ permalink raw reply
* Re: [PATCH RFC] brcmfmac: sanitize DMI strings v2
From: Hans de Goede @ 2019-05-06 16:05 UTC (permalink / raw)
To: Victor Bravo, Kalle Valo
Cc: Arend Van Spriel, Franky Lin, Hante Meuleman, Chi-Hsien Lin,
Wright Feng, David S. Miller, linux-wireless,
brcm80211-dev-list.pdl, brcm80211-dev-list, linux-kernel
In-Reply-To: <20190506152441.ifjcdi73elxuq5it@localhost>
Hi,
On 06-05-19 17:24, Victor Bravo wrote:
> On Mon, May 06, 2019 at 03:26:28PM +0300, Kalle Valo wrote:
>> Hans de Goede <hdegoede@redhat.com> writes:
>>
>>> If we're going to do some filtering, then I suggest we play it safe and also
>>> disallow other chars which may be used as a separator somewhere, specifically
>>> ':' and ','.
>>>
>>> Currently upstream linux-firmware has these files which rely on the DMI
>>> matching:
>>>
>>> brcmfmac4330-sdio.Prowise-PT301.txt
>>> brcmfmac43430-sdio.Hampoo-D2D3_Vi8A1.txt
>>> brcmfmac43430a0-sdio.ONDA-V80 PLUS.txt
>>>
>>> The others are either part of the DMI override table for devices with unsuitable
>>> DMI strings like "Default String"; or are device-tree based.
>>>
>>> So as long as we don't break those 3 (or break the ONDA one but get a symlink
>>> in place) we can sanitize a bit more then just non-printable and '/'.
>>>
>>> Kalle, Arend, what is your opinion on this?
>>>
>>> Note I do not expect the ONDA V80 Plus to have a lot of Linux users,
>>> but it definitely has some.
>>
>> To me having spaces in filenames is a bad idea, but on the other hand we
>> do have the "don't break existing setups" rule, so it's not so simple. I
>> vote for not allowing spaces, I think that's the best for the long run,
>> but don't know what Arend thinks.
>
> I have found a fresh judicate on this:
> https://lkml.org/lkml/2018/12/22/221
>
> It seems clear that we have to support at least spaces for some time
> (maybe wih separate config option which will be deprecated but on by
> defaut until old files are considered gone).
Ah that issue, well that is not really comparable in that case a lot of
peoples setups were completely broken by that commit and it was a
quite surprising behavior change in a userspace facing API.
The nvram loading path already does 2 tries, I really don't want to
unnecessary complicate it with a third try.
The Onda V80 Plus is a X86 based Windows / Android dual boot tablet,
as said before I do not expect a ton of users to be running regular
Linux on it.
Given Kalle's clear preference for getting rid of the spaces lets
just do that. But first we must get a symlink added to linux-firmware
using the name with the _, newer kernels requiring a newer linux-firmware
to match is not unheard of AFAIK, so combined with the limited amount
of users I think this is a reasonable compromise.
Kalle, do you agree with getting the symlink added to linux-firmware
ASAP as a fix for the V80 Plus issue; or do you want to see a fallback
to the un-cleaned name as you suggested before ?
Regards,
Hans
^ permalink raw reply
* Re: [PATCH v2 1/1] ath10k: Enable MSA region dump support for WCN3990
From: Brian Norris @ 2019-05-06 16:31 UTC (permalink / raw)
To: Govind Singh; +Cc: ath10k, linux-wireless
In-Reply-To: <20190506142603.1746-1-govinds@codeaurora.org>
On Mon, May 6, 2019 at 7:26 AM Govind Singh <govinds@codeaurora.org> wrote:
> --- a/drivers/net/wireless/ath/ath10k/snoc.c
> +++ b/drivers/net/wireless/ath/ath10k/snoc.c
> @@ -1586,6 +1587,72 @@ static int ath10k_hw_power_off(struct ath10k *ar)
> return ret;
> }
>
> +static void ath10k_msa_dump_memory(struct ath10k *ar,
> + struct ath10k_fw_crash_data *crash_data)
> +{
> + struct ath10k_snoc *ar_snoc = ath10k_snoc_priv(ar);
> + const struct ath10k_hw_mem_layout *mem_layout;
> + const struct ath10k_mem_region *current_region;
> + struct ath10k_dump_ram_data_hdr *hdr;
> + size_t buf_len;
> + u8 *buf;
> +
> + lockdep_assert_held(&ar->data_lock);
I believe that's the wrong lock now. See below.
> +
> + if (!crash_data && !crash_data->ramdump_buf)
> + return;
> +
> + mem_layout = ath10k_coredump_get_mem_layout(ar);
> + if (!mem_layout)
> + return;
> +
> + current_region = &mem_layout->region_table.regions[0];
> +
> + buf = crash_data->ramdump_buf;
> + buf_len = crash_data->ramdump_buf_len;
> + memset(buf, 0, buf_len);
> +
> + /* Reserve space for the header. */
> + hdr = (void *)buf;
> + buf += sizeof(*hdr);
> + buf_len -= sizeof(*hdr);
> +
> + hdr->region_type = cpu_to_le32(current_region->type);
> + hdr->start = cpu_to_le32(ar_snoc->qmi->msa_va);
> + hdr->length = cpu_to_le32(ar_snoc->qmi->msa_mem_size);
> +
> + if (current_region->len < ar_snoc->qmi->msa_mem_size) {
> + memcpy(buf, ar_snoc->qmi->msa_va, current_region->len);
> + ath10k_warn(ar, "msa dump length is less than msa size %x, %x\n",
> + current_region->len, ar_snoc->qmi->msa_mem_size);
> + } else {
> + memcpy(buf, ar_snoc->qmi->msa_va, ar_snoc->qmi->msa_mem_size);
> + }
> +}
> +
> +void ath10k_snoc_fw_crashed_dump(struct ath10k *ar)
> +{
> + struct ath10k_fw_crash_data *crash_data;
> + char guid[UUID_STRING_LEN + 1];
> +
> + spin_lock_bh(&ar->data_lock);
> +
> + ar->stats.fw_crash_counter++;
> +
> + crash_data = ath10k_coredump_new(ar);
This will (for good reason) spit a lockdep warning after this, I think:
38faed150438 ath10k: perform crash dump collection in workqueue
You need to hold 'dump_mutex' now. I believe you only need to hold
'data_lock' for the sake of the crash counter.
Brian
> +
> + if (crash_data)
> + scnprintf(guid, sizeof(guid), "%pUl", &crash_data->guid);
> + else
> + scnprintf(guid, sizeof(guid), "n/a");
> +
> + ath10k_err(ar, "firmware crashed! (guid %s)\n", guid);
> + ath10k_print_driver_info(ar);
> + ath10k_msa_dump_memory(ar, crash_data);
> +
> + spin_unlock_bh(&ar->data_lock);
> +}
^ permalink raw reply
* [RFC] rtlwifi: rtl8821ae: Use inline routines rather than macros for descriptor word 0
From: Larry Finger @ 2019-05-06 17:39 UTC (permalink / raw)
To: kvalo; +Cc: linux-wireless, pkshih, Larry Finger
The driver uses complicated macros to set parts of word 0 of the TX and RX
descriptors. These are changed into inline routines.
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
---
Kalle,
Based on your comment on how much you dislike those "byte macros", I have
converted a few of them from rtl8821ae into static inline functions.
Is this what you had in mind, and do you consider these changes to
improve the code?
These routines still need to mask the value before the ! operation with
the masked original value.
Thanks,
Larry
---
.../wireless/realtek/rtlwifi/rtl8821ae/trx.c | 33 ++++----
.../wireless/realtek/rtlwifi/rtl8821ae/trx.h | 80 ++++++++++++++-----
2 files changed, 78 insertions(+), 35 deletions(-)
diff --git a/drivers/net/wireless/realtek/rtlwifi/rtl8821ae/trx.c b/drivers/net/wireless/realtek/rtlwifi/rtl8821ae/trx.c
index 7b6faf38e09c..2ad33cfb1656 100644
--- a/drivers/net/wireless/realtek/rtlwifi/rtl8821ae/trx.c
+++ b/drivers/net/wireless/realtek/rtlwifi/rtl8821ae/trx.c
@@ -703,8 +703,8 @@ void rtl8821ae_tx_fill_desc(struct ieee80211_hw *hw,
if (firstseg) {
if (rtlhal->earlymode_enable) {
SET_TX_DESC_PKT_OFFSET(pdesc, 1);
- SET_TX_DESC_OFFSET(pdesc, USB_HWDESC_HEADER_LEN +
- EM_HDR_LEN);
+ set_tx_desc_offset((__le32 *)pdesc,
+ USB_HWDESC_HEADER_LEN + EM_HDR_LEN);
if (ptcb_desc->empkt_num) {
RT_TRACE(rtlpriv, COMP_SEND, DBG_TRACE,
"Insert 8 byte.pTcb->EMPktNum:%d\n",
@@ -713,7 +713,8 @@ void rtl8821ae_tx_fill_desc(struct ieee80211_hw *hw,
(u8 *)(skb->data));
}
} else {
- SET_TX_DESC_OFFSET(pdesc, USB_HWDESC_HEADER_LEN);
+ set_tx_desc_offset((__le32 *)pdesc,
+ USB_HWDESC_HEADER_LEN);
}
@@ -752,8 +753,8 @@ void rtl8821ae_tx_fill_desc(struct ieee80211_hw *hw,
SET_TX_DESC_TX_SUB_CARRIER(pdesc,
rtl8821ae_sc_mapping(hw, ptcb_desc));
- SET_TX_DESC_LINIP(pdesc, 0);
- SET_TX_DESC_PKT_SIZE(pdesc, (u16)skb_len);
+ set_tx_desc_linip((__le32 *)pdesc, 0);
+ set_tx_desc_pkt_size((__le32 *)pdesc, (u16)skb_len);
if (sta) {
u8 ampdu_density = sta->ht_cap.ampdu_density;
@@ -789,15 +790,15 @@ void rtl8821ae_tx_fill_desc(struct ieee80211_hw *hw,
RT_TRACE(rtlpriv, COMP_SEND, DBG_TRACE,
"Enable RDG function.\n");
SET_TX_DESC_RDG_ENABLE(pdesc, 1);
- SET_TX_DESC_HTC(pdesc, 1);
+ set_tx_desc_htc((__le32 *)pdesc, 1);
}
}
/* tx report */
rtl_set_tx_report(ptcb_desc, pdesc, hw, tx_info);
}
- SET_TX_DESC_FIRST_SEG(pdesc, (firstseg ? 1 : 0));
- SET_TX_DESC_LAST_SEG(pdesc, (lastseg ? 1 : 0));
+ set_tx_desc_first_seg((__le32 *)pdesc, (firstseg ? 1 : 0));
+ set_tx_desc_last_seg((__le32 *)pdesc, (lastseg ? 1 : 0));
SET_TX_DESC_TX_BUFFER_SIZE(pdesc, (u16)buf_len);
SET_TX_DESC_TX_BUFFER_ADDRESS(pdesc, mapping);
/* if (rtlpriv->dm.useramask) { */
@@ -815,7 +816,7 @@ void rtl8821ae_tx_fill_desc(struct ieee80211_hw *hw,
SET_TX_DESC_MORE_FRAG(pdesc, (lastseg ? 0 : 1));
if (is_multicast_ether_addr(ieee80211_get_DA(hdr)) ||
is_broadcast_ether_addr(ieee80211_get_DA(hdr))) {
- SET_TX_DESC_BMC(pdesc, 1);
+ set_tx_desc_bmc((__le32 *)pdesc, 1);
}
rtl8821ae_dm_set_tx_ant_by_tx_info(hw, pdesc, ptcb_desc->mac_id);
@@ -841,12 +842,12 @@ void rtl8821ae_tx_fill_cmddesc(struct ieee80211_hw *hw,
}
CLEAR_PCI_TX_DESC_CONTENT(pdesc, TX_DESC_SIZE);
- SET_TX_DESC_FIRST_SEG(pdesc, 1);
- SET_TX_DESC_LAST_SEG(pdesc, 1);
+ set_tx_desc_first_seg((__le32 *)pdesc, 1);
+ set_tx_desc_last_seg((__le32 *)pdesc, 1);
- SET_TX_DESC_PKT_SIZE((u8 *)pdesc, (u16)(skb->len));
+ set_tx_desc_pkt_size((__le32 *)pdesc, (u16)(skb->len));
- SET_TX_DESC_OFFSET(pdesc, USB_HWDESC_HEADER_LEN);
+ set_tx_desc_offset((__le32 *)pdesc, USB_HWDESC_HEADER_LEN);
SET_TX_DESC_USE_RATE(pdesc, 1);
SET_TX_DESC_TX_RATE(pdesc, DESC_RATE1M);
@@ -864,7 +865,7 @@ void rtl8821ae_tx_fill_cmddesc(struct ieee80211_hw *hw,
SET_TX_DESC_MACID(pdesc, 0);
- SET_TX_DESC_OWN(pdesc, 1);
+ set_tx_desc_own((__le32 *)pdesc, 1);
RT_PRINT_DATA(rtlpriv, COMP_CMD, DBG_LOUD,
"H2C Tx Cmd Content\n",
@@ -877,7 +878,7 @@ void rtl8821ae_set_desc(struct ieee80211_hw *hw, u8 *pdesc,
if (istx) {
switch (desc_name) {
case HW_DESC_OWN:
- SET_TX_DESC_OWN(pdesc, 1);
+ set_tx_desc_own((__le32 *)pdesc, 1);
break;
case HW_DESC_TX_NEXTDESC_ADDR:
SET_TX_DESC_NEXT_DESC_ADDRESS(pdesc, *(u32 *)val);
@@ -919,7 +920,7 @@ u64 rtl8821ae_get_desc(struct ieee80211_hw *hw,
if (istx) {
switch (desc_name) {
case HW_DESC_OWN:
- ret = GET_TX_DESC_OWN(pdesc);
+ ret = get_tx_desc_own((__le32 *)pdesc);
break;
case HW_DESC_TXBUFF_ADDR:
ret = GET_TX_DESC_TX_BUFFER_ADDRESS(pdesc);
diff --git a/drivers/net/wireless/realtek/rtlwifi/rtl8821ae/trx.h b/drivers/net/wireless/realtek/rtlwifi/rtl8821ae/trx.h
index 861d78a24d05..64deaf4dab23 100644
--- a/drivers/net/wireless/realtek/rtlwifi/rtl8821ae/trx.h
+++ b/drivers/net/wireless/realtek/rtlwifi/rtl8821ae/trx.h
@@ -14,25 +14,67 @@
#define USB_HWDESC_HEADER_LEN 40
#define CRCLENGTH 4
-#define SET_TX_DESC_PKT_SIZE(__pdesc, __val) \
- SET_BITS_TO_LE_4BYTE(__pdesc, 0, 16, __val)
-#define SET_TX_DESC_OFFSET(__pdesc, __val) \
- SET_BITS_TO_LE_4BYTE(__pdesc, 16, 8, __val)
-#define SET_TX_DESC_BMC(__pdesc, __val) \
- SET_BITS_TO_LE_4BYTE(__pdesc, 24, 1, __val)
-#define SET_TX_DESC_HTC(__pdesc, __val) \
- SET_BITS_TO_LE_4BYTE(__pdesc, 25, 1, __val)
-#define SET_TX_DESC_LAST_SEG(__pdesc, __val) \
- SET_BITS_TO_LE_4BYTE(__pdesc, 26, 1, __val)
-#define SET_TX_DESC_FIRST_SEG(__pdesc, __val) \
- SET_BITS_TO_LE_4BYTE(__pdesc, 27, 1, __val)
-#define SET_TX_DESC_LINIP(__pdesc, __val) \
- SET_BITS_TO_LE_4BYTE(__pdesc, 28, 1, __val)
-#define SET_TX_DESC_OWN(__pdesc, __val) \
- SET_BITS_TO_LE_4BYTE(__pdesc, 31, 1, __val)
-
-#define GET_TX_DESC_OWN(__pdesc) \
- LE_BITS_TO_4BYTE(__pdesc, 31, 1)
+/* Set packet size (16 bits) in TX descriptor word 0 */
+static inline void set_tx_desc_pkt_size(__le32 *__pdesc, u16 __val)
+{
+ *__pdesc = cpu_to_le32((le32_to_cpu(*__pdesc) & ~GENMASK(0, 15)) |
+ __val);
+}
+
+/* Set offset (8 bits) in TX descriptor word 0 */
+static inline void set_tx_desc_offset(__le32 *__pdesc, u8 __val)
+{
+ *__pdesc = cpu_to_le32((le32_to_cpu(*__pdesc) & ~GENMASK(16, 23)) |
+ __val << 16);
+}
+
+/* Set bmc (1 bit) in TX descriptor word 0 */
+static inline void set_tx_desc_bmc(__le32 *__pdesc, u8 __val)
+{
+ *__pdesc = cpu_to_le32((le32_to_cpu(*__pdesc) & ~GENMASK(24, 24)) |
+ __val << 24);
+}
+
+/* Set htc (1 bit) in TX descriptor word 0 */
+static inline void set_tx_desc_htc(__le32 *__pdesc, u8 __val)
+{
+ *__pdesc = cpu_to_le32((le32_to_cpu(*__pdesc) & ~GENMASK(25, 25)) |
+ __val << 25);
+}
+
+/* Set last segment flag (1 bit) in TX descriptor word 0 */
+static inline void set_tx_desc_last_seg(__le32 *__pdesc, u8 __val)
+{
+ *__pdesc = cpu_to_le32((le32_to_cpu(*__pdesc) & ~GENMASK(26, 26)) |
+ __val << 26);
+}
+
+/* Set first segment iflag (1 bit) in TX descriptor word 0 */
+static inline void set_tx_desc_first_seg(__le32 *__pdesc, u8 __val)
+{
+ *__pdesc = cpu_to_le32((le32_to_cpu(*__pdesc) & ~GENMASK(27, 27)) |
+ __val << 27);
+}
+
+/* Set linip (1 bit) in TX descriptor word 0 */
+static inline void set_tx_desc_linip(__le32 *__pdesc, u8 __val)
+{
+ *__pdesc = cpu_to_le32((le32_to_cpu(*__pdesc) & ~GENMASK(28, 28)) |
+ __val << 18);
+}
+
+/* Set own flag (1 bit) in TX descriptor word 0 */
+static inline void set_tx_desc_own(__le32 *__pdesc, u8 __val)
+{
+ *__pdesc = cpu_to_le32((le32_to_cpu(*__pdesc) & ~GENMASK(31, 31)) |
+ __val << 31);
+}
+
+/* Get own flag (1 bit) from TX descriptor word 0 */
+static inline u8 get_tx_desc_own(__le32 *__pdesc)
+{
+ return (le32_to_cpu(*__pdesc) & ~GENMASK(31, 31)) >> 31;
+}
#define SET_TX_DESC_MACID(__pdesc, __val) \
SET_BITS_TO_LE_4BYTE(__pdesc+4, 0, 7, __val)
--
2.21.0
^ permalink raw reply related
* [PATCH 1/2] mac80211: AMPDU handling for Extended Key ID
From: Alexander Wetzel @ 2019-05-06 19:01 UTC (permalink / raw)
To: johannes; +Cc: linux-wireless, Alexander Wetzel
IEEE 802.11 - 2016 forbids mixing MPDUs with different keyIDs in one
A-MPDU. Drivers supporting A-MPDUs and Extended Key ID must actively
enforce that requirement due to the available two unicast keyIDs.
Allow driver to signal mac80211 that they will not check the keyID in
MPDUs when aggregating them and that they expect mac80211 to stop Tx
aggregation when rekeying a connection using Extended Key ID.
Signed-off-by: Alexander Wetzel <alexander@wetzel-home.de>
---
This is now just the simple and obvious solution for the A-MPDU mixing
problem. Just stopping Tx aggregation may not be very elegant, but for
drivers like iwlwifi - which can't control how many MPDUs can be
aggregated - the sane solution.
Once we have more cards supporting A-MPDU and Extended Key ID it may make
sense to add another mechanism, like the proposed Key ID border signal.
But then maybe not...
The patch is looking bigger than it really is, It's moving some code
around to clean up the logic and replace ieee80211_hw_key_replace() with
ieee80211_pairwise_rekey() which is then of course no longer installing
the keys to the HW.
That way we can have the "special cases" for unicast key crypto handling
in one function, for both HW and SW crypto.
(Which so far are PTK0 rekey and Extended Key ID A-MPDU keyID mixing.)
Not so nice is the use of WLAN_STA_BLOCK_BA. We only have to make sure
no new Tx aggregation sessions can be started but block both. It just
seems to be overkill to add a more granular control here.
WLAN_STA_BLOCK_BA is also used when suspending, so with some bad timing
Extended Key ID or PTK0 rekeys can unlock the lock from suspend.
It just looks pretty harmless, since a suspend will disconnect us anyhow.
I can of course add something like a lock counter, but that again seems
to be overkill for now. (After all this is intended to be the minimal
solution to get the job done:-)
I've of course ran some tests with the patch and can confirm that this
seems to solves the observed MPDU corruptions with mvm cards while not
causing undesired side effects.
I've also uploaded two captures taken on a STA connected to an AP and
uploaded them again to https://www.awhome.eu/index.php/s/AJJXBLsZmzHdxpX
iwlwifi-fixed.pcap.gz has been taken with an older, simpler and
incomplete patch, iwlwifi-fixed-final.pcap was then taken with the patch
provided here.
Note that I did not drop the unused variable "old" YueHaibing provided a
patch for on 29th of April. That is indeed a leftover from a older patch
without any function which somehow crept into the final patch.
I assume you'll apply his patch and prefer that I do not remove it here
also...
include/net/mac80211.h | 4 ++
net/mac80211/debugfs.c | 1 +
net/mac80211/key.c | 100 +++++++++++++++++++++++------------------
3 files changed, 61 insertions(+), 44 deletions(-)
diff --git a/include/net/mac80211.h b/include/net/mac80211.h
index 72080d9d617e..b7f0b56a09f4 100644
--- a/include/net/mac80211.h
+++ b/include/net/mac80211.h
@@ -2269,6 +2269,9 @@ struct ieee80211_txq {
* @IEEE80211_HW_EXT_KEY_ID_NATIVE: Driver and hardware are supporting Extended
* Key ID and can handle two unicast keys per station for Rx and Tx.
*
+ * @IEEE80211_HW_NO_AMPDU_KEYBORDER_SUPPORT: The card/driver can't handle
+ * active Tx A-MPDU sessions with Extended Key IDs during rekey.
+ *
* @NUM_IEEE80211_HW_FLAGS: number of hardware flags, used for sizing arrays
*/
enum ieee80211_hw_flags {
@@ -2321,6 +2324,7 @@ enum ieee80211_hw_flags {
IEEE80211_HW_SUPPORTS_MULTI_BSSID,
IEEE80211_HW_SUPPORTS_ONLY_HE_MULTI_BSSID,
IEEE80211_HW_EXT_KEY_ID_NATIVE,
+ IEEE80211_HW_NO_AMPDU_KEYBORDER_SUPPORT,
/* keep last, obviously */
NUM_IEEE80211_HW_FLAGS
diff --git a/net/mac80211/debugfs.c b/net/mac80211/debugfs.c
index 0d462206eef6..7c580010836e 100644
--- a/net/mac80211/debugfs.c
+++ b/net/mac80211/debugfs.c
@@ -274,6 +274,7 @@ static const char *hw_flag_names[] = {
FLAG(SUPPORTS_MULTI_BSSID),
FLAG(SUPPORTS_ONLY_HE_MULTI_BSSID),
FLAG(EXT_KEY_ID_NATIVE),
+ FLAG(NO_AMPDU_KEYBORDER_SUPPORT),
#undef FLAG
};
diff --git a/net/mac80211/key.c b/net/mac80211/key.c
index 20bf9db7a388..faadfeea7127 100644
--- a/net/mac80211/key.c
+++ b/net/mac80211/key.c
@@ -274,50 +274,61 @@ int ieee80211_set_tx_key(struct ieee80211_key *key)
old = key_mtx_dereference(local, sta->ptk[sta->ptk_idx]);
sta->ptk_idx = key->conf.keyidx;
+
+ if (ieee80211_hw_check(&local->hw, NO_AMPDU_KEYBORDER_SUPPORT))
+ clear_sta_flag(sta, WLAN_STA_BLOCK_BA);
ieee80211_check_fast_xmit(sta);
return 0;
}
-static int ieee80211_hw_key_replace(struct ieee80211_key *old_key,
- struct ieee80211_key *new_key,
- bool pairwise)
+static void ieee80211_pairwise_rekey(struct ieee80211_key *old,
+ struct ieee80211_key *new)
{
- struct ieee80211_sub_if_data *sdata;
- struct ieee80211_local *local;
- struct sta_info *sta;
- int ret;
-
- /* Aggregation sessions are OK when running on SW crypto.
- * A broken remote STA may cause issues not observed with HW
- * crypto, though.
- */
- if (!(old_key->flags & KEY_FLAG_UPLOADED_TO_HARDWARE))
- return 0;
+ struct ieee80211_local *local = new->local;
+ struct sta_info *sta = new->sta;
+ int i;
- assert_key_lock(old_key->local);
- sta = old_key->sta;
+ assert_key_lock(local);
- /* Unicast rekey without Extended Key ID needs special handling */
- if (new_key && sta && pairwise &&
- rcu_access_pointer(sta->ptk[sta->ptk_idx]) == old_key) {
- local = old_key->local;
- sdata = old_key->sdata;
+ if (new->conf.flags & IEEE80211_KEY_FLAG_NO_AUTO_TX) {
+ /* Extended Key ID key install, initial one or rekey */
+
+ if (sta->ptk_idx != INVALID_PTK_KEYIDX &&
+ ieee80211_hw_check(&local->hw,
+ NO_AMPDU_KEYBORDER_SUPPORT)) {
+ /* Aggregation Sessions with Extended Key ID must not
+ * mix MPDUs with different keyIDs within one A-MPDU.
+ * Tear down any running Tx aggregation and all new
+ * Rx/Tx aggregation request during rekey if the driver
+ * asks us to do so. (Blocking Tx only would be
+ * sufficient but WLAN_STA_BLOCK_BA gets the job done
+ * for the few ms we need it.)
+ */
+ set_sta_flag(sta, WLAN_STA_BLOCK_BA);
+ mutex_lock(&sta->ampdu_mlme.mtx);
+ for (i = 0; i < IEEE80211_NUM_TIDS; i++)
+ ___ieee80211_stop_tx_ba_session(sta, i,
+ AGG_STOP_LOCAL_REQUEST);
+ mutex_unlock(&sta->ampdu_mlme.mtx);
+ }
+ } else if (old) {
+ /* Rekey without Extended Key ID.
+ * Aggregation sessions are OK when running on SW crypto.
+ * A broken remote STA may cause issues not observed with HW
+ * crypto, though.
+ */
+ if (!(old->flags & KEY_FLAG_UPLOADED_TO_HARDWARE))
+ return;
- /* Stop TX till we are on the new key */
- old_key->flags |= KEY_FLAG_TAINTED;
+ /* Stop Tx till we are on the new key */
+ old->flags |= KEY_FLAG_TAINTED;
ieee80211_clear_fast_xmit(sta);
-
- /* Aggregation sessions during rekey are complicated due to the
- * reorder buffer and retransmits. Side step that by blocking
- * aggregation during rekey and tear down running sessions.
- */
if (ieee80211_hw_check(&local->hw, AMPDU_AGGREGATION)) {
set_sta_flag(sta, WLAN_STA_BLOCK_BA);
ieee80211_sta_tear_down_BA_sessions(sta,
AGG_STOP_LOCAL_REQUEST);
}
-
if (!wiphy_ext_feature_isset(local->hw.wiphy,
NL80211_EXT_FEATURE_CAN_REPLACE_PTK0)) {
pr_warn_ratelimited("Rekeying PTK for STA %pM but driver can't safely do that.",
@@ -325,18 +336,9 @@ static int ieee80211_hw_key_replace(struct ieee80211_key *old_key,
/* Flushing the driver queues *may* help prevent
* the clear text leaks and freezes.
*/
- ieee80211_flush_queues(local, sdata, false);
+ ieee80211_flush_queues(local, old->sdata, false);
}
}
-
- ieee80211_key_disable_hw_accel(old_key);
-
- if (new_key)
- ret = ieee80211_key_enable_hw_accel(new_key);
- else
- ret = 0;
-
- return ret;
}
static void __ieee80211_set_default_key(struct ieee80211_sub_if_data *sdata,
@@ -394,7 +396,6 @@ void ieee80211_set_default_mgmt_key(struct ieee80211_sub_if_data *sdata,
mutex_unlock(&sdata->local->key_mtx);
}
-
static int ieee80211_key_replace(struct ieee80211_sub_if_data *sdata,
struct sta_info *sta,
bool pairwise,
@@ -402,7 +403,7 @@ static int ieee80211_key_replace(struct ieee80211_sub_if_data *sdata,
struct ieee80211_key *new)
{
int idx;
- int ret;
+ int ret = 0;
bool defunikey, defmultikey, defmgmtkey;
/* caller must provide at least one old/new */
@@ -414,16 +415,27 @@ static int ieee80211_key_replace(struct ieee80211_sub_if_data *sdata,
WARN_ON(new && old && new->conf.keyidx != old->conf.keyidx);
+ if (new && sta && pairwise) {
+ /* Unicast rekey needs special handling. With Extended Key ID
+ * old is still NULL for the first rekey.
+ */
+ ieee80211_pairwise_rekey(old, new);
+ }
+
if (old) {
idx = old->conf.keyidx;
- ret = ieee80211_hw_key_replace(old, new, pairwise);
+
+ if (old->flags & KEY_FLAG_UPLOADED_TO_HARDWARE) {
+ ieee80211_key_disable_hw_accel(old);
+
+ if (new)
+ ret = ieee80211_key_enable_hw_accel(new);
+ }
} else {
/* new must be provided in case old is not */
idx = new->conf.keyidx;
if (!new->local->wowlan)
ret = ieee80211_key_enable_hw_accel(new);
- else
- ret = 0;
}
if (ret)
--
2.21.0
^ permalink raw reply related
* Re: [PATCH RFC] brcmfmac: sanitize DMI strings v2
From: Arend Van Spriel @ 2019-05-06 19:30 UTC (permalink / raw)
To: Hans de Goede, Victor Bravo, Kalle Valo
Cc: Franky Lin, Hante Meuleman, Chi-Hsien Lin, Wright Feng,
David S. Miller, linux-wireless, brcm80211-dev-list.pdl,
brcm80211-dev-list, linux-kernel, Luis Chamberlain
In-Reply-To: <3f3cca6e-50b7-c61d-4a62-26ce508af9e7@redhat.com>
+ Luis (for real this time)
On 5/6/2019 6:05 PM, Hans de Goede wrote:
> Hi,
>
> On 06-05-19 17:24, Victor Bravo wrote:
>> On Mon, May 06, 2019 at 03:26:28PM +0300, Kalle Valo wrote:
>>> Hans de Goede <hdegoede@redhat.com> writes:
>>>
>>>> If we're going to do some filtering, then I suggest we play it safe
>>>> and also
>>>> disallow other chars which may be used as a separator somewhere,
>>>> specifically
>>>> ':' and ','.
>>>>
>>>> Currently upstream linux-firmware has these files which rely on the DMI
>>>> matching:
>>>>
>>>> brcmfmac4330-sdio.Prowise-PT301.txt
>>>> brcmfmac43430-sdio.Hampoo-D2D3_Vi8A1.txt
>>>> brcmfmac43430a0-sdio.ONDA-V80 PLUS.txt
>>>>
>>>> The others are either part of the DMI override table for devices
>>>> with unsuitable
>>>> DMI strings like "Default String"; or are device-tree based.
>>>>
>>>> So as long as we don't break those 3 (or break the ONDA one but get
>>>> a symlink
>>>> in place) we can sanitize a bit more then just non-printable and '/'.
>>>>
>>>> Kalle, Arend, what is your opinion on this?
>>>>
>>>> Note I do not expect the ONDA V80 Plus to have a lot of Linux users,
>>>> but it definitely has some.
>>>
>>> To me having spaces in filenames is a bad idea, but on the other hand we
>>> do have the "don't break existing setups" rule, so it's not so simple. I
>>> vote for not allowing spaces, I think that's the best for the long run,
>>> but don't know what Arend thinks.
Hi,
Had a day off today so I did see some of the discussion, but was not
able to chime in until now.
To be honest I always disliked spaces in filenames, but that does not
necessarily make it a bad idea. What I would like to know is why
built-in firmware can not deal with spaces in the firmware file names. I
think Hans mentioned it in the thread and it crossed my mind as well
last night. From driver perspective, being brcmfmac or any other for
that matter, there is only one API to request firmware and in my opinion
it should behave the same no matter where the firmware is coming from. I
would prefer to fix that for built-in firmware, but we need to
understand where this limitation is coming from. Hopefully Luis can
elaborate on that.
>>
>> I have found a fresh judicate on this:
>> https://lkml.org/lkml/2018/12/22/221
>>
>> It seems clear that we have to support at least spaces for some time
>> (maybe wih separate config option which will be deprecated but on by
>> defaut until old files are considered gone).
>
> Ah that issue, well that is not really comparable in that case a lot of
> peoples setups were completely broken by that commit and it was a
> quite surprising behavior change in a userspace facing API.
>
> The nvram loading path already does 2 tries, I really don't want to
> unnecessary complicate it with a third try.
>
> The Onda V80 Plus is a X86 based Windows / Android dual boot tablet,
> as said before I do not expect a ton of users to be running regular
> Linux on it.
>
> Given Kalle's clear preference for getting rid of the spaces lets
> just do that. But first we must get a symlink added to linux-firmware
> using the name with the _, newer kernels requiring a newer linux-firmware
> to match is not unheard of AFAIK, so combined with the limited amount
> of users I think this is a reasonable compromise.
Right. In the brcm folder we have bcm4329-fullmac-4.bin for older
kernels and brcmfmac4329-sdio.bin for later kernels when we switched to
stricter firmware naming convention.
> Kalle, do you agree with getting the symlink added to linux-firmware
> ASAP as a fix for the V80 Plus issue; or do you want to see a fallback
> to the un-cleaned name as you suggested before ?
How many releases have an issue and how many V80 Plus users running
regular linux are actually using built-in firmware. And is it really a
regression for them? Not saying it does not require fixing. However, as
stated above I would prefer to fix the built-in firmware limitation if
possible and backport that fix if it is only a few kernel releases
provided stable allows such a backport.
Regards,
Arend
^ permalink raw reply
* [PATCH 2/2] iwlwifi: Extended Key ID support
From: Alexander Wetzel @ 2019-05-06 19:01 UTC (permalink / raw)
To: johannes; +Cc: linux-wireless, Alexander Wetzel
In-Reply-To: <20190506190149.3624-1-alexander@wetzel-home.de>
iwlwifi cards are all able to support Extended Key ID as long as
mac80211 stops A-MPDU aggregation when replacing the unicast key.
Enable Extended Key ID support and tell mac80211 that it has to tear
down any running Tx aggregation sessions during rekey when using
Extended Key ID.
Signed-off-by: Alexander Wetzel <alexander@wetzel-home.de>
---
drivers/net/wireless/intel/iwlwifi/dvm/mac80211.c | 2 ++
drivers/net/wireless/intel/iwlwifi/mvm/mac80211.c | 2 ++
2 files changed, 4 insertions(+)
diff --git a/drivers/net/wireless/intel/iwlwifi/dvm/mac80211.c b/drivers/net/wireless/intel/iwlwifi/dvm/mac80211.c
index 54b759cec8b3..160c3fd04399 100644
--- a/drivers/net/wireless/intel/iwlwifi/dvm/mac80211.c
+++ b/drivers/net/wireless/intel/iwlwifi/dvm/mac80211.c
@@ -111,6 +111,8 @@ int iwlagn_mac_setup_register(struct iwl_priv *priv,
ieee80211_hw_set(hw, SUPPORTS_DYNAMIC_PS);
ieee80211_hw_set(hw, SUPPORT_FAST_XMIT);
ieee80211_hw_set(hw, WANT_MONITOR_VIF);
+ ieee80211_hw_set(hw, EXT_KEY_ID_NATIVE);
+ ieee80211_hw_set(hw, NO_AMPDU_KEYBORDER_SUPPORT);
if (priv->trans->max_skb_frags)
hw->netdev_features = NETIF_F_HIGHDMA | NETIF_F_SG;
diff --git a/drivers/net/wireless/intel/iwlwifi/mvm/mac80211.c b/drivers/net/wireless/intel/iwlwifi/mvm/mac80211.c
index 5c52469288be..6475adbe60e9 100644
--- a/drivers/net/wireless/intel/iwlwifi/mvm/mac80211.c
+++ b/drivers/net/wireless/intel/iwlwifi/mvm/mac80211.c
@@ -470,6 +470,8 @@ int iwl_mvm_mac_setup_register(struct iwl_mvm *mvm)
ieee80211_hw_set(hw, STA_MMPDU_TXQ);
ieee80211_hw_set(hw, TX_AMSDU);
ieee80211_hw_set(hw, TX_FRAG_LIST);
+ ieee80211_hw_set(hw, EXT_KEY_ID_NATIVE);
+ ieee80211_hw_set(hw, NO_AMPDU_KEYBORDER_SUPPORT);
if (iwl_mvm_has_tlc_offload(mvm)) {
ieee80211_hw_set(hw, TX_AMPDU_SETUP_IN_HW);
--
2.21.0
^ permalink raw reply related
* Re: [PATCH] rsi: Properly initialize data in rsi_sdio_ta_reset
From: Nathan Chancellor @ 2019-05-06 20:51 UTC (permalink / raw)
To: Kalle Valo
Cc: Nick Desaulniers, Amitkumar Karwar, Siva Rebbagondla,
linux-wireless, netdev, LKML, clang-built-linux
In-Reply-To: <87h8ackv8j.fsf@kamboji.qca.qualcomm.com>
On Fri, May 03, 2019 at 07:38:52AM +0300, Kalle Valo wrote:
> Nick Desaulniers <ndesaulniers@google.com> writes:
>
> > On Thu, May 2, 2019 at 8:16 AM Nathan Chancellor
> > <natechancellor@gmail.com> wrote:
> >>
> >> When building with -Wuninitialized, Clang warns:
> >>
> >> drivers/net/wireless/rsi/rsi_91x_sdio.c:940:43: warning: variable 'data'
> >> is uninitialized when used here [-Wuninitialized]
> >> put_unaligned_le32(TA_HOLD_THREAD_VALUE, data);
> >> ^~~~
> >> drivers/net/wireless/rsi/rsi_91x_sdio.c:930:10: note: initialize the
> >> variable 'data' to silence this warning
> >> u8 *data;
> >> ^
> >> = NULL
> >> 1 warning generated.
> >>
> >> Using Clang's suggestion of initializing data to NULL wouldn't work out
> >> because data will be dereferenced by put_unaligned_le32. Use kzalloc to
> >> properly initialize data, which matches a couple of other places in this
> >> driver.
> >>
> >> Fixes: e5a1ecc97e5f ("rsi: add firmware loading for 9116 device")
> >> Link: https://github.com/ClangBuiltLinux/linux/issues/464
> >> Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
> >> ---
> >> drivers/net/wireless/rsi/rsi_91x_sdio.c | 21 ++++++++++++++-------
> >> 1 file changed, 14 insertions(+), 7 deletions(-)
> >>
> >> diff --git a/drivers/net/wireless/rsi/rsi_91x_sdio.c b/drivers/net/wireless/rsi/rsi_91x_sdio.c
> >> index f9c67ed473d1..b35728564c7b 100644
> >> --- a/drivers/net/wireless/rsi/rsi_91x_sdio.c
> >> +++ b/drivers/net/wireless/rsi/rsi_91x_sdio.c
> >> @@ -929,11 +929,15 @@ static int rsi_sdio_ta_reset(struct rsi_hw *adapter)
> >> u32 addr;
> >> u8 *data;
> >>
> >> + data = kzalloc(sizeof(u32), GFP_KERNEL);
> >
> > Something fishy is going on here. We allocate 4 B but declare data as
> > a u8* (pointer to individual bytes)? In general, dynamically
> > allocating that few bytes is a code smell; either you meant to just
> > use the stack, or this memory's lifetime extends past the lifetime of
> > this stackframe, at which point you probably just meant to stack
> > allocate space in a higher parent frame and pass this preallocated
> > memory down to the child frame to get filled in.
> >
> > Reading through this code, I don't think that the memory is meant to
> > outlive the stack frame. Is there a reason why we can't just declare
> > data as:
> >
> > u8 data [4];
> >
> > then use ARRAY_SIZE(data) or RSI_9116_REG_SIZE in rsi_reset_chip(),
> > getting rid of the kzalloc/kfree?
>
> I haven't checked the details but AFAIK stack variables are not supposed
> to be used with DMA. So in that case I think it's ok alloc four bytes,
> unless the DMA rules have changed of course. But I didn't check if rsi
> is using DMA here, just a general comment.
>
> --
> Kalle Valo
I don't think it is using the DMA API but it might be the same thing for
SDIO. If passing that around on the stack is okay, great but we don't
want what commit f700546682a6 ("rsi: fix nommu_map_sg overflow kernel
panic") fixes to happen here.
I can't answer that for sure though since I am not at all familiar with
this driver or the SDIO APIs.
Cheers,
Nathan
^ permalink raw reply
* Re: [PATCH v8 13/15] x86/split_lock: Enable split lock detection by default
From: Fenghua Yu @ 2019-05-06 21:39 UTC (permalink / raw)
To: Thomas Gleixner
Cc: Ingo Molnar, Borislav Petkov, H Peter Anvin, Paolo Bonzini,
Dave Hansen, Ashok Raj, Peter Zijlstra, Ravi V Shankar,
Xiaoyao Li, Christopherson Sean J, Kalle Valo, Michael Chan,
linux-kernel, x86, kvm, netdev, linux-wireless
In-Reply-To: <alpine.DEB.2.21.1904250943160.1762@nanos.tec.linutronix.de>
On Thu, Apr 25, 2019 at 09:50:20AM +0200, Thomas Gleixner wrote:
> On Wed, 24 Apr 2019, Fenghua Yu wrote:
> >
> > +static void split_lock_update_msr(void)
> > +{
> > + /* Enable split lock detection */
> > + msr_set_bit(MSR_TEST_CTL, TEST_CTL_SPLIT_LOCK_DETECT_SHIFT);
> > + this_cpu_or(msr_test_ctl_cache, TEST_CTL_SPLIT_LOCK_DETECT);
>
> I'm pretty sure, that I told you to utilize the cache proper. Again:
>
> > > Nothing in this file initializes msr_test_ctl_cache explicitely. Register
> > > caching always requires to read the register and store it in the cache
> > > before doing anything with it. Nothing guarantees that all bits in that MSR
> > > are 0 by default forever.
> > >
> > > And once you do that _before_ calling split_lock_update_msr() then you can
> > > spare the RMW in that function.
>
> So you managed to fix the initializaiton part, but then you still do a
> pointless RMW.
Ok. I see. msr_set_bit() is a RMW operation.
So is the following the right code to update msr and cache variable?
+static void split_lock_update_msr(void)
+{
+ /* Enable split lock detection */
+ this_cpu_or(msr_test_ctl_cache, TEST_CTL_SPLIT_LOCK_DETECT);
+ wrmsrl(MSR_TEST_CTL, msr_test_ctl_cache);
Thanks.
-Fenghua
^ permalink raw reply
* [PATCH RFC v3] brcmfmac: sanitize DMI strings
From: Victor Bravo @ 2019-05-06 22:04 UTC (permalink / raw)
To: Arend Van Spriel
Cc: Hans de Goede, Kalle Valo, Franky Lin, Hante Meuleman,
Chi-Hsien Lin, Wright Feng, David S. Miller, linux-wireless,
brcm80211-dev-list.pdl, brcm80211-dev-list, linux-kernel,
Luis Chamberlain
Hi,
as I expect to be busy tomorrow, I'm sending polished, hopefully more
readable version of the patch, and also some more implementation details
which can be useful in the search for proper sanitizing function.
Patch description:
Sanitize DMI strings in brcmfmac driver to make resulting filenames
contain only safe characters. This version replaces all non-printable
characters incl. delete (1-31, 127-255), spaces and slashes with
underscores.
This change breaks backward compatibility, but adds control over strings
passed to firmware loader and compatibility with CONFIG_EXTRA_FIRMWARE
which doesn't support spaces in filenames.
Implementation details:
The algorithm for sanitizing function was selected after following
choices were considered (C = number of allowed characters , N = number
of characters in sanitized strings):
ALGORITHM TIME COMPL. DATA DISADVANTAGES
bitmask-based O(N) 16 no macro to make mask from string
hardcoded-check O(N) 0 not universal, big / prone to growth
runtime-mask O(C+N) 16+C C+N and 16+C don't seem "best" enough
string-based O(C*N) C slow, big, far from "best possible"
(runtime-mask means bitmask-based algorithm with runtime mask
initialization from string of allowed characters)
Bitmask-based implementation won due to results summarized above.
It's only 7-bit clear, because this allows to save 16 bytes of data
at the price of single (char < 0) comparison, and C char is kind
of 7-bit anyway.
Known issues:
Function like brcmf_dmi_sanitize() shouldn't be part of driver, as
this practice leads to duplicate code, which is not acceptable for many
good reasons. An existing function from lib/ should be used instead,
or implemented there if no suitable replacement is found.
---
Changes from v1: don't revert fresh commit by someone else
Changes from v2: outfactor check to BRCMF_DMI_SANE_CHAR in an attempt to
make the code more readable, improve comments, +Luis
Signed-off-by: Victor Bravo <1905@spmblk.com>
diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/dmi.c b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/dmi.c
index 7535cb0d4ac0..c8cb9c0b3f6e 100644
--- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/dmi.c
+++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/dmi.c
@@ -23,6 +23,18 @@
/* The DMI data never changes so we can use a static buf for this */
static char dmi_board_type[128];
+/*
+ * Array of 128 bits representing 7-bit characters allowed in DMI strings,
+ * byte-based to avoid endianess issues.
+ */
+static unsigned char brcmf_dmi_allowed_chars[] = {
+ 0x00, 0x00, 0x00, 0x00, 0xfe, 0x7f, 0xff, 0xff,
+ 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x7f
+};
+
+/* Safe character to replace disallowed ones. */
+#define BRCMF_DMI_SAFE_CHAR '_'
+
struct brcmf_dmi_data {
u32 chip;
u32 chiprev;
@@ -99,6 +111,19 @@ static const struct dmi_system_id dmi_platform_data[] = {
{}
};
+/* Checks character value agains bitmask of allowed characters */
+#define BRCMF_DMI_SANE_CHAR(mask, value) \
+ (mask[value / 8] & (1 << (value % 8)))
+
+void brcmf_dmi_sanitize(char *dest, const unsigned char *allowed, char safe)
+{
+ while (*dest) {
+ if ((*dest < 0) || !BRCMF_DMI_SANE_CHAR(allowed, *dest))
+ *dest = safe;
+ dest++;
+ }
+}
+
void brcmf_dmi_probe(struct brcmf_mp_device *settings, u32 chip, u32 chiprev)
{
const struct dmi_system_id *match;
@@ -126,6 +151,9 @@ void brcmf_dmi_probe(struct brcmf_mp_device *settings, u32 chip, u32 chiprev)
if (sys_vendor && product_name) {
snprintf(dmi_board_type, sizeof(dmi_board_type), "%s-%s",
sys_vendor, product_name);
+ brcmf_dmi_sanitize(dmi_board_type,
+ brcmf_dmi_allowed_chars,
+ BRCMF_DMI_SAFE_CHAR);
settings->board_type = dmi_board_type;
}
}
^ permalink raw reply related
* [PATCH 5.2 v3] rtw88: fix unassigned rssi_level in rtw_sta_info
From: yhchuang @ 2019-05-07 2:28 UTC (permalink / raw)
To: kvalo; +Cc: linux-wireless
From: Yan-Hsuan Chuang <yhchuang@realtek.com>
The new rssi_level should be stored in si, otherwise the rssi_level will
never be updated and get a wrong RA mask, which is calculated by the
rssi level
If a wrong RA mask is chosen, the firmware will pick some *bad rates*.
The most hurtful scene will be in *noisy environment*, such as office or
public area with many APs and users.
The latency would be high and the overall throughput would be only half
or less.
Tested in 2.4G in office area, with this patch the throughput increased
from such as "1x Mbps -> 4x Mbps".
Signed-off-by: Yan-Hsuan Chuang <yhchuang@realtek.com>
---
drivers/net/wireless/realtek/rtw88/phy.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/wireless/realtek/rtw88/phy.c b/drivers/net/wireless/realtek/rtw88/phy.c
index 4381b36..7f437e2 100644
--- a/drivers/net/wireless/realtek/rtw88/phy.c
+++ b/drivers/net/wireless/realtek/rtw88/phy.c
@@ -144,10 +144,10 @@ static void rtw_phy_stat_rssi_iter(void *data, struct ieee80211_sta *sta)
struct rtw_phy_stat_iter_data *iter_data = data;
struct rtw_dev *rtwdev = iter_data->rtwdev;
struct rtw_sta_info *si = (struct rtw_sta_info *)sta->drv_priv;
- u8 rssi, rssi_level;
+ u8 rssi;
rssi = ewma_rssi_read(&si->avg_rssi);
- rssi_level = rtw_phy_get_rssi_level(si->rssi_level, rssi);
+ si->rssi_level = rtw_phy_get_rssi_level(si->rssi_level, rssi);
rtw_fw_send_rssi_info(rtwdev, si);
--
2.7.4
^ permalink raw reply related
* RE: [PATCH] ath10k: remove mmc_hw_reset while hif power down
From: Wen Gong @ 2019-05-07 5:05 UTC (permalink / raw)
To: Grant Grundler, Wen Gong
Cc: linux-wireless@vger.kernel.org, ath10k@lists.infradead.org
In-Reply-To: <CANEJEGug_YregN5vZq_R=ppUV5ptb4WUq_TB+JApzQ4DEO_AfQ@mail.gmail.com>
> -----Original Message-----
> From: ath10k <ath10k-bounces@lists.infradead.org> On Behalf Of Grant
> Grundler
> Sent: Saturday, May 4, 2019 2:01 AM
> To: Wen Gong <wgong@codeaurora.org>
> Cc: linux-wireless@vger.kernel.org; ath10k@lists.infradead.org
> Subject: [EXT] Re: [PATCH] ath10k: remove mmc_hw_reset while hif power
> down
>
> [repeating comments I made in the gerrit review for Chrome OS :
> https://chromium-
> review.googlesource.com/c/chromiumos/third_party/kernel/+/1585667
> ]
>
> On Sat, Apr 27, 2019 at 7:17 PM Wen Gong <wgong@codeaurora.org> wrote:
> >
> > For sdio 3.0 chip, the clock will drop from 200M Hz to 50M Hz after load
> > ath10k driver, it is because mmc_hw_reset will reset the sdio's power,
> > then mmc will consider it as sdio 2.0 and drop the clock.
>
> Wen,
> 5468e784c0600551ca03263f5255a375c05f88e7 commit message gives
> reasons
> for adding the mmc_hw_reset() call. The commit message for removing
> gives different reason for removal. Both are good but second one is
> incomplete.
>
> The commit message for removal should ALSO explain why adding this
> call wasn't necessary in the first place OR move the call to a
> different code path.
>
> > Remove mmc_hw_reset will avoid the drop of clock.
>
> This commit message makes it clear the original patch introduced a new
> problem. But the original patch fixed a different problem and that
> this proposed change seems likely to re-introduce and the commit
> message should explain why that isn't true (or how the original was
> fixed differently)
Hi Grant,
The mmc_hw_reset's effect depends on the hardware layout/configure
software's behavior, recently it will effect the clock of sdio for the platform
I used. And it will still work well without mmc_hw_reset for the platform I
Used currently.
If sdio cannot work on other platform, I think it can add flag in
ath10k_hw_params_list for the platform to call the mmc_hw_reset depends
on the flag.
>
> cheers,
> grant
^ permalink raw reply
* [PATCH AUTOSEL 5.0 83/99] spi: ST ST95HF NFC: declare missing of table
From: Sasha Levin @ 2019-05-07 5:32 UTC (permalink / raw)
To: linux-kernel, stable
Cc: Daniel Gomez, Javier Martinez Canillas, David S . Miller,
Sasha Levin, linux-wireless
In-Reply-To: <20190507053235.29900-1-sashal@kernel.org>
From: Daniel Gomez <dagmcr@gmail.com>
[ Upstream commit d04830531d0c4a99c897a44038e5da3d23331d2f ]
Add missing <of_device_id> table for SPI driver relying on SPI
device match since compatible is in a DT binding or in a DTS.
Before this patch:
modinfo drivers/nfc/st95hf/st95hf.ko | grep alias
alias: spi:st95hf
After this patch:
modinfo drivers/nfc/st95hf/st95hf.ko | grep alias
alias: spi:st95hf
alias: of:N*T*Cst,st95hfC*
alias: of:N*T*Cst,st95hf
Reported-by: Javier Martinez Canillas <javier@dowhile0.org>
Signed-off-by: Daniel Gomez <dagmcr@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
drivers/nfc/st95hf/core.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/drivers/nfc/st95hf/core.c b/drivers/nfc/st95hf/core.c
index 2b26f762fbc3..01acb6e53365 100644
--- a/drivers/nfc/st95hf/core.c
+++ b/drivers/nfc/st95hf/core.c
@@ -1074,6 +1074,12 @@ static const struct spi_device_id st95hf_id[] = {
};
MODULE_DEVICE_TABLE(spi, st95hf_id);
+static const struct of_device_id st95hf_spi_of_match[] = {
+ { .compatible = "st,st95hf" },
+ { },
+};
+MODULE_DEVICE_TABLE(of, st95hf_spi_of_match);
+
static int st95hf_probe(struct spi_device *nfc_spi_dev)
{
int ret;
@@ -1260,6 +1266,7 @@ static struct spi_driver st95hf_driver = {
.driver = {
.name = "st95hf",
.owner = THIS_MODULE,
+ .of_match_table = of_match_ptr(st95hf_spi_of_match),
},
.id_table = st95hf_id,
.probe = st95hf_probe,
--
2.20.1
^ permalink raw reply related
* [PATCH AUTOSEL 4.19 13/81] cfg80211: Handle WMM rules in regulatory domain intersection
From: Sasha Levin @ 2019-05-07 5:34 UTC (permalink / raw)
To: linux-kernel, stable
Cc: Ilan Peer, Luca Coelho, Johannes Berg, Sasha Levin,
linux-wireless, netdev
In-Reply-To: <20190507053554.30848-1-sashal@kernel.org>
From: Ilan Peer <ilan.peer@intel.com>
[ Upstream commit 08a75a887ee46828b54600f4bb7068d872a5edd5 ]
The support added for regulatory WMM rules did not handle
the case of regulatory domain intersections. Fix it.
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Fixes: 230ebaa189af ("cfg80211: read wmm rules from regulatory database")
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
net/wireless/reg.c | 39 +++++++++++++++++++++++++++++++++++++++
1 file changed, 39 insertions(+)
diff --git a/net/wireless/reg.c b/net/wireless/reg.c
index 8002ace7c9f6..8a47297ff206 100644
--- a/net/wireless/reg.c
+++ b/net/wireless/reg.c
@@ -1287,6 +1287,16 @@ reg_intersect_dfs_region(const enum nl80211_dfs_regions dfs_region1,
return dfs_region1;
}
+static void reg_wmm_rules_intersect(const struct ieee80211_wmm_ac *wmm_ac1,
+ const struct ieee80211_wmm_ac *wmm_ac2,
+ struct ieee80211_wmm_ac *intersect)
+{
+ intersect->cw_min = max_t(u16, wmm_ac1->cw_min, wmm_ac2->cw_min);
+ intersect->cw_max = max_t(u16, wmm_ac1->cw_max, wmm_ac2->cw_max);
+ intersect->cot = min_t(u16, wmm_ac1->cot, wmm_ac2->cot);
+ intersect->aifsn = max_t(u8, wmm_ac1->aifsn, wmm_ac2->aifsn);
+}
+
/*
* Helper for regdom_intersect(), this does the real
* mathematical intersection fun
@@ -1301,6 +1311,8 @@ static int reg_rules_intersect(const struct ieee80211_regdomain *rd1,
struct ieee80211_freq_range *freq_range;
const struct ieee80211_power_rule *power_rule1, *power_rule2;
struct ieee80211_power_rule *power_rule;
+ const struct ieee80211_wmm_rule *wmm_rule1, *wmm_rule2;
+ struct ieee80211_wmm_rule *wmm_rule;
u32 freq_diff, max_bandwidth1, max_bandwidth2;
freq_range1 = &rule1->freq_range;
@@ -1311,6 +1323,10 @@ static int reg_rules_intersect(const struct ieee80211_regdomain *rd1,
power_rule2 = &rule2->power_rule;
power_rule = &intersected_rule->power_rule;
+ wmm_rule1 = &rule1->wmm_rule;
+ wmm_rule2 = &rule2->wmm_rule;
+ wmm_rule = &intersected_rule->wmm_rule;
+
freq_range->start_freq_khz = max(freq_range1->start_freq_khz,
freq_range2->start_freq_khz);
freq_range->end_freq_khz = min(freq_range1->end_freq_khz,
@@ -1354,6 +1370,29 @@ static int reg_rules_intersect(const struct ieee80211_regdomain *rd1,
intersected_rule->dfs_cac_ms = max(rule1->dfs_cac_ms,
rule2->dfs_cac_ms);
+ if (rule1->has_wmm && rule2->has_wmm) {
+ u8 ac;
+
+ for (ac = 0; ac < IEEE80211_NUM_ACS; ac++) {
+ reg_wmm_rules_intersect(&wmm_rule1->client[ac],
+ &wmm_rule2->client[ac],
+ &wmm_rule->client[ac]);
+ reg_wmm_rules_intersect(&wmm_rule1->ap[ac],
+ &wmm_rule2->ap[ac],
+ &wmm_rule->ap[ac]);
+ }
+
+ intersected_rule->has_wmm = true;
+ } else if (rule1->has_wmm) {
+ *wmm_rule = *wmm_rule1;
+ intersected_rule->has_wmm = true;
+ } else if (rule2->has_wmm) {
+ *wmm_rule = *wmm_rule2;
+ intersected_rule->has_wmm = true;
+ } else {
+ intersected_rule->has_wmm = false;
+ }
+
if (!is_valid_reg_rule(intersected_rule))
return -EINVAL;
--
2.20.1
^ permalink raw reply related
* [PATCH AUTOSEL 4.19 15/81] nl80211: Add NL80211_FLAG_CLEAR_SKB flag for other NL commands
From: Sasha Levin @ 2019-05-07 5:34 UTC (permalink / raw)
To: linux-kernel, stable
Cc: Sunil Dutt, Johannes Berg, Sasha Levin, linux-wireless, netdev
In-Reply-To: <20190507053554.30848-1-sashal@kernel.org>
From: Sunil Dutt <usdutt@codeaurora.org>
[ Upstream commit d6db02a88a4aaa1cd7105137c67ddec7f3bdbc05 ]
This commit adds NL80211_FLAG_CLEAR_SKB flag to other NL commands
that carry key data to ensure they do not stick around on heap
after the SKB is freed.
Also introduced this flag for NL80211_CMD_VENDOR as there are sub
commands which configure the keys.
Signed-off-by: Sunil Dutt <usdutt@codeaurora.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
net/wireless/nl80211.c | 18 ++++++++++++------
1 file changed, 12 insertions(+), 6 deletions(-)
diff --git a/net/wireless/nl80211.c b/net/wireless/nl80211.c
index 295cd8d5554f..048e004ed0ee 100644
--- a/net/wireless/nl80211.c
+++ b/net/wireless/nl80211.c
@@ -13392,7 +13392,8 @@ static const struct genl_ops nl80211_ops[] = {
.policy = nl80211_policy,
.flags = GENL_UNS_ADMIN_PERM,
.internal_flags = NL80211_FLAG_NEED_NETDEV_UP |
- NL80211_FLAG_NEED_RTNL,
+ NL80211_FLAG_NEED_RTNL |
+ NL80211_FLAG_CLEAR_SKB,
},
{
.cmd = NL80211_CMD_DEAUTHENTICATE,
@@ -13443,7 +13444,8 @@ static const struct genl_ops nl80211_ops[] = {
.policy = nl80211_policy,
.flags = GENL_UNS_ADMIN_PERM,
.internal_flags = NL80211_FLAG_NEED_NETDEV_UP |
- NL80211_FLAG_NEED_RTNL,
+ NL80211_FLAG_NEED_RTNL |
+ NL80211_FLAG_CLEAR_SKB,
},
{
.cmd = NL80211_CMD_UPDATE_CONNECT_PARAMS,
@@ -13451,7 +13453,8 @@ static const struct genl_ops nl80211_ops[] = {
.policy = nl80211_policy,
.flags = GENL_ADMIN_PERM,
.internal_flags = NL80211_FLAG_NEED_NETDEV_UP |
- NL80211_FLAG_NEED_RTNL,
+ NL80211_FLAG_NEED_RTNL |
+ NL80211_FLAG_CLEAR_SKB,
},
{
.cmd = NL80211_CMD_DISCONNECT,
@@ -13480,7 +13483,8 @@ static const struct genl_ops nl80211_ops[] = {
.policy = nl80211_policy,
.flags = GENL_UNS_ADMIN_PERM,
.internal_flags = NL80211_FLAG_NEED_NETDEV_UP |
- NL80211_FLAG_NEED_RTNL,
+ NL80211_FLAG_NEED_RTNL |
+ NL80211_FLAG_CLEAR_SKB,
},
{
.cmd = NL80211_CMD_DEL_PMKSA,
@@ -13832,7 +13836,8 @@ static const struct genl_ops nl80211_ops[] = {
.policy = nl80211_policy,
.flags = GENL_UNS_ADMIN_PERM,
.internal_flags = NL80211_FLAG_NEED_WIPHY |
- NL80211_FLAG_NEED_RTNL,
+ NL80211_FLAG_NEED_RTNL |
+ NL80211_FLAG_CLEAR_SKB,
},
{
.cmd = NL80211_CMD_SET_QOS_MAP,
@@ -13887,7 +13892,8 @@ static const struct genl_ops nl80211_ops[] = {
.doit = nl80211_set_pmk,
.policy = nl80211_policy,
.internal_flags = NL80211_FLAG_NEED_NETDEV_UP |
- NL80211_FLAG_NEED_RTNL,
+ NL80211_FLAG_NEED_RTNL |
+ NL80211_FLAG_CLEAR_SKB,
},
{
.cmd = NL80211_CMD_DEL_PMK,
--
2.20.1
^ permalink raw reply related
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox