[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH v5 9/9] drivers/char: fix handling cable re-plug in XHCI console driver
- To: Andrew Cooper <Andrew.Cooper3@xxxxxxxxxx>
- From: Jan Beulich <jbeulich@xxxxxxxx>
- Date: Tue, 6 Sep 2022 08:52:25 +0200
- Arc-authentication-results: i=1; mx.microsoft.com 1; spf=pass smtp.mailfrom=suse.com; dmarc=pass action=none header.from=suse.com; dkim=pass header.d=suse.com; arc=none
- Arc-message-signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=microsoft.com; s=arcselector9901; h=From:Date:Subject:Message-ID:Content-Type:MIME-Version:X-MS-Exchange-AntiSpam-MessageData-ChunkCount:X-MS-Exchange-AntiSpam-MessageData-0:X-MS-Exchange-AntiSpam-MessageData-1; bh=ytRg+7SBo4U3Tgcagqli6pR6ip8yI/MJujbsIYjbjEA=; b=m2g4oBWcx0izBJ9UoRQe4Vdgl29Zwkqs4qTN3D8qQEkKBYWNmmghnJomEcAtZ8YX3TlhRwAtpx1TvC6TRhoUuLUmBvgkfweac9Iurjnyf79PKHSmU9p8m6Ig9hG18WmmkOPZdPV17M8vfvHJI2OgEn/xjABSe621jQKAcd0VVK7/8HHFVrgM34fdoyvEFnjx7bweC0HbAutGV5RfkliYs2DYooZ0THZoHkFju++luLrDJG8m7n/n8Akb6gR63Y6DVyZCVoDmSkhRmh5aqsH82UpMgydQu+fP7S2HUSGOOmOIOwn7aC2rNLJur+h70C9/b6SaQFZFVk5RA+oZKHDBEQ==
- Arc-seal: i=1; a=rsa-sha256; s=arcselector9901; d=microsoft.com; cv=none; b=HdmDYn0PyVMUeuEMq6Rvkw5GKAIXQ93uWZK9sc9/oN2bjmoCHi5/rC0qdc61b2nC8pBcywrqDE1NxBMH0nW/1fxs1xAUDN/2Fv5/IKiKdv6BksEX4eSPx1V8HJVipQSvNkHJO31VyRD9c2FVfADd2MmJZFxaRCkwE0d1WUahWLi58dhKWeVsZTYULlCPVjaP7NlxQTk+8ESsm+hWpcACvgZY2P8SihAJWMCDgjfx6leweayua1uFT9LUdSiMs+HxT6JddJDo/wUYKF3euws7FxMRXCbNyHAnOqhziDwtdO/OnPBdl7N4sft4Lbs+8NfaPQmI+V6BssQ3D1mrLLScEw==
- Authentication-results: dkim=none (message not signed) header.d=none;dmarc=none action=none header.from=suse.com;
- Cc: George Dunlap <George.Dunlap@xxxxxxxxxx>, Julien Grall <julien@xxxxxxx>, Stefano Stabellini <sstabellini@xxxxxxxxxx>, Wei Liu <wl@xxxxxxx>, "xen-devel@xxxxxxxxxxxxxxxxxxxx" <xen-devel@xxxxxxxxxxxxxxxxxxxx>, Marek Marczykowski-Górecki <marmarek@xxxxxxxxxxxxxxxxxxxxxx>
- Delivery-date: Tue, 06 Sep 2022 06:52:24 +0000
- List-id: Xen developer discussion <xen-devel.lists.xenproject.org>
On 26.08.2022 17:44, Andrew Cooper wrote:
> On 26/08/2022 15:50, Jan Beulich wrote:
>> On 22.08.2022 17:27, Marek Marczykowski-Górecki wrote:
>>> @@ -1023,6 +1032,8 @@ static bool dbc_ensure_running(struct dbc *dbc)
>>> writel(ctrl | (1U << DBC_CTRL_DRC), ®->ctrl);
>>> writel(readl(®->portsc) | (1U << DBC_PSC_PED), ®->portsc);
>>> wmb();
>>> + dbc_ring_doorbell(dbc, dbc->dbc_iring.db);
>>> + dbc_ring_doorbell(dbc, dbc->dbc_oring.db);
>>> }
>> You retain the wmb() here, but ...
>>
>>> @@ -1066,8 +1073,7 @@ static void dbc_flush(struct dbc *dbc, struct
>>> xhci_trb_ring *trb,
>>> }
>>> }
>>>
>>> - wmb();
>>> - writel(db, ®->db);
>>> + dbc_ring_doorbell(dbc, trb->db);
>>> }
>> ... you drop it here. Why the difference?
>
> As a tangent, every single barrier in this file is buggy. Should be
> smp_*() variants, not mandatory variants.
>
> All (interesting) data is in plain WB cached memory, and the few BAR
> registers which are configured have a UC mapping which orders properly
> WRT other writes on x86.
But such drivers shouldn't be x86-specific when it comes to their use
of barriers. For this reason I specifically did not complain about any
of the barrier uses throughout the series (with the further thinking
of "better one too many than one too few").
Jan
|
