| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <4b7cdd3e-f244-8458-f5a2-1f1da79fa4a3@gmail.com>
Date: Fri, 7 Feb 2020 00:56:22 +0200
From: Lauri Jakku <ljakku77@...il.com>
To: Heiner Kallweit <hkallweit1@...il.com>, Lauri Jakku <lja@....fi>,
nic_swsd@...ltek.com
Cc: netdev@...r.kernel.org
Subject: Re: [PATCH] NET: Realtek depency chain r8169 -> realtec -> libphy
fixed.
Hi,
Toughts while compiling:
On 2020-02-07 00:39, Lauri Jakku wrote:
> Hi,
>
> On 2020-02-07 00:36, Heiner Kallweit wrote:
>> On 06.02.2020 23:20, Lauri Jakku wrote:
>>> On 2020-02-06 23:56, Heiner Kallweit wrote:
>>>> On 06.02.2020 19:51, Lauri Jakku wrote:
>>>>> * Added soft depency from realtec phy to libphy.
>>>>>
>>>>> [ 39.953438] Generic PHY r8169-200:00: attached PHY driver [Generic PHY] (mii_bus:phy_addr=r8169-200:00, irq=IGNORE)
>>>>> [ 39.957413] ------------[ cut here ]------------
>>>>> [ 39.957414] read_page callback not available, PHY driver not loaded?
>>>>> [ 39.957458] WARNING: CPU: 3 PID: 3896 at drivers/net/phy/phy-core.c:700 __phy_read_page+0x3f/0x50 [libphy]
>>>>> [ 39.957459] Modules linked in: cmac algif_hash algif_skcipher af_alg bnep nls_iso8859_1 nls_cp437 vfat fat squashfs loop videobuf2_vmalloc videobuf2_memops snd_usb_audio videobuf2_v4l2 amdgpu videobuf2_common snd_usbmidi_lib videodev snd_rawmidi snd_seq_device mc btusb btrtl btbcm btintel mousedev input_leds joydev
>>>>> bluetooth gpu_sched snd_hda_codec_realtek i2c_algo_bit ttm ecdh_generic snd_hda_codec_generic snd_hda_codec_hdmi rfkill ecc drm_kms_helper ledtrig_audio snd_hda_intel drm snd_intel_dspcfg snd_hda_codec agpgart snd_hda_core syscopyarea sysfillrect sysimgblt snd_hwdep fb_sys_fops snd_pcm snd_timer r8169 snd soundcore eda
>>>>> c_mce_amd sp5100_tco kvm_amd i2c_piix4 realtek libphy ccp wmi_bmof ppdev rng_core k10temp kvm irqbypass parport_pc evdev parport mac_hid wmi pcspkr acpi_cpufreq uinput crypto_user ip_tables x_tables ext4 crc32c_generic crc16 mbcache jbd2 hid_generic usbhid hid sr_mod cdrom sd_mod ohci_pci pata_atiixp ata_generic pata_acpi firewire_ohci ahci pata_jmicron
>>>>> [ 39.957483] firewire_core libahci crc_itu_t libata scsi_mod ehci_pci ehci_hcd ohci_hcd floppy
>>>>> [ 39.957488] CPU: 3 PID: 3896 Comm: NetworkManager Not tainted 5.5.0-2-MANJARO-usb-mod-v4 #1
>>>>> [ 39.957489] Hardware name: Gigabyte Technology Co., Ltd. GA-MA790FXT-UD5P/GA-MA790FXT-UD5P, BIOS F8l 07/15/2010
>>>>> [ 39.957494] RIP: 0010:__phy_read_page+0x3f/0x50 [libphy]
>>>>> [ 39.957496] Code: c0 74 05 e9 33 77 3d e9 80 3d cd e3 00 00 00 74 06 b8 a1 ff ff ff c3 48 c7 c7 50 0c 63 c0 c6 05 b7 e3 00 00 01 e8 33 70 86 e8 <0f> 0b eb e3 66 66 2e 0f 1f 84 00 00 00 00 00 66 90 0f 1f 44 00 00
>>>>> [ 39.957497] RSP: 0018:ffffa459ca3fb3b0 EFLAGS: 00010282
>>>>> [ 39.957498] RAX: 0000000000000000 RBX: 0000000000006662 RCX: 0000000000000000
>>>>> [ 39.957499] RDX: 0000000000000001 RSI: 0000000000000092 RDI: 00000000ffffffff
>>>>> [ 39.957499] RBP: ffff9c91b46c3800 R08: 000000000000047a R09: 0000000000000001
>>>>> [ 39.957500] R10: 0000000000000000 R11: 0000000000000001 R12: ffff9c91b5a8a8c0
>>>>> [ 39.957500] R13: 0000000000000002 R14: 0000000000000001 R15: 0000000000000000
>>>>> [ 39.957501] FS: 00007ff199d38d80(0000) GS:ffff9c91b7cc0000(0000) knlGS:0000000000000000
>>>>> [ 39.957502] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>>>>> [ 39.957503] CR2: 00007f907f428ff8 CR3: 00000001ed122000 CR4: 00000000000006e0
>>>>> [ 39.957503] Call Trace:
>>>>> [ 39.957511] phy_select_page+0x28/0x50 [libphy]
>>>>> [ 39.957518] phy_write_paged+0x18/0x50 [libphy]
>>>>> [ 39.957523] rtl8168d_1_hw_phy_config+0x1c8/0x1f0 [r8169]
>>>>> [ 39.957526] rtl8169_init_phy+0x2c/0xb0 [r8169]
>>>>> [ 39.957529] rtl_open+0x3b2/0x570 [r8169]
>>>>> [ 39.957533] __dev_open+0xe0/0x170
>>>>> [ 39.957535] __dev_change_flags+0x188/0x1e0
>>>>> [ 39.957537] dev_change_flags+0x21/0x60
>>>>> [ 39.957539] do_setlink+0x78a/0xf90
>>>>> [ 39.957544] ? kernel_init_free_pages+0x6d/0x90
>>>>> [ 39.957546] ? prep_new_page+0x46/0xd0
>>>>> [ 39.957548] ? cpumask_next+0x16/0x20
>>>>> [ 39.957550] ? __snmp6_fill_stats64.isra.0+0x66/0x110
>>>>> [ 39.957553] __rtnl_newlink+0x5d1/0x9a0
>>>>> [ 39.957563] rtnl_newlink+0x44/0x70
>>>>> [ 39.957564] rtnetlink_rcv_msg+0x137/0x3c0
>>>>> [ 39.957566] ? rtnl_calcit.isra.0+0x120/0x120
>>>>> [ 39.957568] netlink_rcv_skb+0x75/0x140
>>>>> [ 39.957570] netlink_unicast+0x199/0x240
>>>>> [ 39.957572] netlink_sendmsg+0x243/0x480
>>>>> [ 39.957575] sock_sendmsg+0x5e/0x60
>>>>> [ 39.957576] ____sys_sendmsg+0x21b/0x290
>>>>> [ 39.957577] ? copy_msghdr_from_user+0xe1/0x160
>>>>> [ 39.957580] ___sys_sendmsg+0x9e/0xe0
>>>>> [ 39.957583] ? addrconf_sysctl_forward+0x12b/0x270
>>>>> [ 39.957585] __sys_sendmsg+0x81/0xd0
>>>>> [ 39.957588] do_syscall_64+0x4e/0x150
>>>>> [ 39.957591] entry_SYSCALL_64_after_hwframe+0x44/0xa9
>>>>> [ 39.957593] RIP: 0033:0x7ff19af247ed
>>>>> [ 39.957594] Code: 28 89 54 24 1c 48 89 74 24 10 89 7c 24 08 e8 4a 53 f8 ff 8b 54 24 1c 48 8b 74 24 10 41 89 c0 8b 7c 24 08 b8 2e 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 2f 44 89 c7 48 89 44 24 08 e8 7e 53 f8 ff 48
>>>>> [ 39.957595] RSP: 002b:00007ffd570ac710 EFLAGS: 00000293 ORIG_RAX: 000000000000002e
>>>>> [ 39.957596] RAX: ffffffffffffffda RBX: 000000000000000c RCX: 00007ff19af247ed
>>>>> [ 39.957596] RDX: 0000000000000000 RSI: 00007ffd570ac750 RDI: 000000000000000c
>>>>> [ 39.957597] RBP: 0000562f3d390090 R08: 0000000000000000 R09: 0000000000000000
>>>>> [ 39.957597] R10: 0000000000000000 R11: 0000000000000293 R12: 0000000000000000
>>>>> [ 39.957598] R13: 00007ffd570ac8b0 R14: 00007ffd570ac8ac R15: 0000000000000000
>>>>> [ 39.957601] ---[ end trace f2cccff3f7fdfb28 ]---
>>>>>
>>>>> Signed-off-by: Lauri Jakku <lja@....fi>
>>>> This patch is not correct for several reasons, most important one being that realtek.ko
>>>> has a hard dependency on libphy already. In your case supposedly r8169.ko is in
>>>> initramfs but realtek.ko is not. This needs to be changed, then the error should be gone.
>>> Yeah, I tried to make a depency chain that would load realtek.ko before probing
>>>
>>> at r8169 driver.
>>>
>> This soft dependency is included in r8169 already. Just if r8169.ko is in initramfs then it
>> can't load realtek.ko if it's not in initramfs.
>>> Should i make realtek.ko as build-in to kernel, or how one enforces module to initramfs ?
>>>
>> Check mkinitcpio.conf. Some distributions may have own tools for configuring initramfs.
>> Ubuntu has lsinitramfs for checking what's included in a particular initramfs file.
> Ok, will do.
I saw that there is no MODULE_SOFTDEP done, witch I think triggers the mechanism
to load realtek.ko (i put in in initramfs now, hopefully). I'm not sure what does the
code at r8169_main.c:
/* Some tools for creating an initramfs don't consider softdeps, then
* r8169.ko may be in initramfs, but realtek.ko not. Then the generic
* PHY driver is used that doesn't work with most chip versions.
*/
if (!driver_find("RTL8201CP Ethernet", &mdio_bus_type)) {
dev_err(&pdev->dev, "realtek.ko not loaded, maybe it needs to be added to initramfs?\n");
return -ENOENT;
}
Trigger module loading without this patch ?
>>>>> ---
>>>>> drivers/net/phy/realtek.c | 1 +
>>>>> 1 file changed, 1 insertion(+)
>>>>>
>>>>> diff --git a/drivers/net/phy/realtek.c b/drivers/net/phy/realtek.c
>>>>> index f5fa2fff3ddc..4a1d4342c71e 100644
>>>>> --- a/drivers/net/phy/realtek.c
>>>>> +++ b/drivers/net/phy/realtek.c
>>>>> @@ -54,6 +54,7 @@
>>>>> MODULE_DESCRIPTION("Realtek PHY driver");
>>>>> MODULE_AUTHOR("Johnson Leung");
>>>>> MODULE_LICENSE("GPL");
>>>>> +MODULE_SOFTDEP("pre: libphy");
>>>>>
>>>>> static int rtl821x_read_page(struct phy_device *phydev)
>>>>> {
>>>>>
Powered by blists - more mailing lists