From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail-lj1-f179.google.com (mail-lj1-f179.google.com [209.85.208.179]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 0A83772 for ; Tue, 24 Aug 2021 08:53:38 +0000 (UTC) Received: by mail-lj1-f179.google.com with SMTP id q21so36248946ljj.6 for ; Tue, 24 Aug 2021 01:53:38 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=subject:from:to:cc:references:message-id:date:user-agent :mime-version:in-reply-to:content-language:content-transfer-encoding; bh=2r4ngFjLjfq9YSn0jkH9oDjdsB7Zq43pzXKudghtmFc=; b=pfS31p19wSJYrzNTPwffVtwFHGcHwSbE703NfJkMZH5TEeMFBPQkz4U8U5NVDBcgud nelWuK3WRt3rFouau4tArlPm+lEYxJzuGkVRnbdDm+UvreOGbvUO/55dsiJkycjVM+VS nLe7P1dmc7V2VWWKcllGBVZreQFTPLpuzcSRu3y1FAIMNXWj/vsdEC7RKmby80DpzQ+i 7v5yFX/4MhHpNzf0AkE01cJNiGXz0dmqDXk9AfYZHusHy+5jbo1X08oz6soEylDWipBm XaFMYU0beRWzYecyLRODrzOanqI0kxI07xNb9Y46WshIKKaj5vaG1AHZqvxVX19XFO6X /j5A== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:subject:from:to:cc:references:message-id:date :user-agent:mime-version:in-reply-to:content-language :content-transfer-encoding; bh=2r4ngFjLjfq9YSn0jkH9oDjdsB7Zq43pzXKudghtmFc=; b=SJxY3+YHGOAD8EPysXR5fBEGUHOyCl6KcB3F+92PAQWs30yKOtQPz4w6iVLslAPmWo QDsPIloBeCKIFdwctWYFi3Q4mC0pONFi90CIUfTp5fuwvNp6Cu0twBgzWVIan4vdbLQa GpTWg+04+NYD0Tbgy5C/19HFN1+Du+bXypQPVki9S20QrKIvHwgfinus0YZEKq/bM8Ma HcM2L2dxJ+yPbthkfQIy+CGgMWMMeVWZiod8JnKf1wvrLcFK+vIbqzHC44FXii1Z3O5w TZK0YRDrvhyfI7BknFSEdD6NySTeVIooM8Ugm/icBTf2Duh8Q+bfPwQpTQ1Jgxdg9+eh d+5A== X-Gm-Message-State: AOAM530DBt0PZnY2EBkxJnC0V9CTAdpyiugTNf/4KO27yjv5CWYb5LIZ 9OhfGiFHfpILisob83nzM+Q= X-Google-Smtp-Source: ABdhPJyz3LJTJ8WmlkjVQKtHRaQseRiwNI2uP++yp4ZBLERUO8lQSi1N4M9St5jdzKtYjGLY/1CL7w== X-Received: by 2002:a05:651c:1695:: with SMTP id bd21mr31444075ljb.5.1629795217106; Tue, 24 Aug 2021 01:53:37 -0700 (PDT) Received: from [192.168.1.11] ([46.235.66.127]) by smtp.gmail.com with ESMTPSA id u2sm1715045lji.82.2021.08.24.01.53.36 (version=TLS1_3 cipher=TLS_AES_128_GCM_SHA256 bits=128/128); Tue, 24 Aug 2021 01:53:36 -0700 (PDT) Subject: Re: [PATCH RFC v2 5/6] staging: r8188eu: add error handling of rtw_read32 From: Pavel Skripkin To: "Fabio M. De Francesco" , Phillip Potter Cc: Larry Finger , Greg KH , Michael Straube , "open list:STAGING SUBSYSTEM" , Linux Kernel Mailing List References: <2424268.2DCLMNk6Lg@localhost.localdomain> <2355365e-d75e-8d1b-aad8-6979b70f33ec@gmail.com> <3419359.n4KOUNuIW1@localhost.localdomain> <77da7c63-a5b1-a09d-39ec-32c5485b8eac@gmail.com> Message-ID: <96e3703e-a5e2-3c6d-ea3c-b5d3892849b2@gmail.com> Date: Tue, 24 Aug 2021 11:53:35 +0300 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101 Thunderbird/78.13.0 Precedence: bulk X-Mailing-List: linux-staging@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 In-Reply-To: <77da7c63-a5b1-a09d-39ec-32c5485b8eac@gmail.com> Content-Type: text/plain; charset=utf-8; format=flowed Content-Language: en-US Content-Transfer-Encoding: 7bit On 8/24/21 11:47 AM, Pavel Skripkin wrote: > On 8/24/21 11:38 AM, Fabio M. De Francesco wrote: >> On Tuesday, August 24, 2021 8:40:18 AM CEST Pavel Skripkin wrote: >>> On 8/24/21 3:10 AM, Fabio M. De Francesco wrote: >>> > On Tuesday, August 24, 2021 1:33:46 AM CEST Phillip Potter wrote: >>> >> On Sun, 22 Aug 2021 at 15:36, Pavel Skripkin wrote: >>> >> > -static u32 usb_read32(struct intf_hdl *pintfhdl, u32 addr) >>> >> > +static int usb_read32(struct intf_hdl *pintfhdl, u32 addr, u32 *data) >>> >> > { >>> >> > u8 requesttype; >>> >> > u16 wvalue; >>> >> > u16 len; >>> >> > - __le32 data; >>> >> > + int res; >>> >> > + __le32 tmp; >>> >> > + >>> >> > + if (WARN_ON(unlikely(!data))) >>> >> > + return -EINVAL; >>> >> > >>> >> > requesttype = 0x01;/* read_in */ >>> >> > >>> >> > wvalue = (u16)(addr & 0x0000ffff); >>> >> > len = 4; >>> >> > >>> >> > - usbctrl_vendorreq(pintfhdl, wvalue, &data, len, requesttype); >>> >> > + res = usbctrl_vendorreq(pintfhdl, wvalue, &data, len, requesttype); >>> >> > + if (res < 0) { >>> >> > + dev_err(dvobj_to_dev(pintfhdl->pintf_dev), "Failed to read 32 bytes: %d\n", res); >>> >> > + } else { >>> >> > + /* Noone cares about positive return value */ >>> >> > + *data = le32_to_cpu(tmp); >>> >> > + res = 0; >>> >> > + } >>> >> > >>> >> > - return le32_to_cpu(data); >>> >> > + return res; >>> >> > } >>> >> >>> >> Dear Pavel, >>> >> >>> >> OK, found the issue with decoded stack trace after reviewing this >>> >> usb_read32 function. Your line: >>> >> res = usbctrl_vendorreq(pintfhdl, wvalue, &data, len, requesttype); >>> >> >>> >> should read: >>> >> res = usbctrl_vendorreq(pintfhdl, wvalue, &tmp, len, requesttype); >>> > >>> > Dear Philip, >>> > >>> > No, it should read: >>> > >>> > res = usbctrl_vendorreq(pintfhdl, wvalue, data, len, requesttype); >>> > >>> > I suspect that Pavel didn't notice he was reusing a line of the old code >>> > wth no due changes. >>> > >>> >> With this change, the driver runs fine with no crashes/oopses. I will >>> >> explain the issue but you can probably see already, so I hope I'm not >>> >> coming across as patronising, just trying to be helpful :-) >>> >> >>> >> Essentially, you are taking the address of the data function parameter >>> >> on this line with &data, a pointer to u32, which is giving you a >>> >> pointer to a pointer to u32 (u32 **) for this function parameter >>> >> variable. When passed to usbctrl_vendorreq, it is being passed to >>> >> memcpy inside this function as a void *, meaning that memcpy >>> >> subsequently overwrites the value of the memory address inside data to >>> >> point to a different location, which is problem when it is later >>> >> deferenced at: >>> >> *data = le32_to_cpu(tmp); >>> >> causing the OOPS >>> >> >>> >> Also, as written, you can probably see that tmp is uninitialised. This >>> >> looks like a typo, so guessing this wasn't your intention. Anyhow, >>> >> with that small change, usbctrl_vendorreq reads into tmp, which is >>> >> then passed to le32_to_cpu whose return value is stored via the >>> >> deferenced data ptr (which now has its original address within and not >>> >> inadvertently modified). Hope this helps, and I'd be happy to Ack the >>> >> series if you want to resend this patch. Many thanks. >>> > >>> > I think that another typo is having 'tmp', because that variable is unnecessary >>> > and "*data = le32_to_cpu(tmp);" is wrong too. >>> > >>> > Now I also see that also usb_read16() is wrong, while usb_read8() (the one that >>> > I had read yesterday) is the only correct function of the three usb_read*(). >>> > >>> >>> Hi, guys! >>> >>> >>> Sorry for breaking your system, Phillip. This code was part of "last >>> minute" changes and yes, it's broken :) >>> >>> I get what Phillip said, because I _should_ read into tmp variable >>> instead of directly to data, but I don't get Fabio's idea, sorry. >> >> Hi Pavel, >> >> I (wrongly?) assumed from the prototype of usb_read32() that u32 *data is in native >> endianness. So, I didn't see the necessity of using _le32 tmp and then convert that tmp >> with le32_to_cpu(). >> >> I simply thought that data could be passed to usbctrl_vendorreq as it-is. >> >>> Data from chip comes in little-endian, so we _should_ convert it to >>> cpu's endian. Temp variable is needed to make smatch and all other >>> static anylis tools happy about this code. >> >> Now that you explained that "Data from chip comes in little-endian", obviously >> I must agree with you that the code needs tmp and that tmp must be >> swapped by le32_to_cpu(), ahead of assigning it to *data. >> >> Just a curiosity... Since I was not able to see that *data is returned in little endian, >> can you please point me where in the code you found out that it is? There must >> be some place in the code that I'm unable to find and see that *data is LE. >> >> Thanks in advance, >> >> Fabio > > Hi, Fabio! > > previous usb_read16() realization, which is 100% right: > > > static u16 usb_read16(struct intf_hdl *pintfhdl, u32 addr) > { > u8 requesttype; > u16 wvalue; > u16 len; > __le32 data; > > requesttype = 0x01;/* read_in */ > wvalue = (u16)(addr & 0x0000ffff); > len = 2; > usbctrl_vendorreq(pintfhdl, wvalue, &data, len, requesttype); > > return (u16)(le32_to_cpu(data) & 0xffff); > } > > > Bases on this code, I think, it's oblivious, that data comes in > little-endian. That's why I leaved temp variable for casting le32 to > cpu's endianess. > > I could just read into u{16,32} * and then make smth like > > *data = le32_to_cpu(*data) > > but static analysis tools will complain about wrong data type passed to > le32_to_cpu() > > + Phillip tested fixed v2 version and it worked well for him. I guess, > Phillip was able to spot weird driver behavior, if this cast is wrong. > ^^^^^& I am wrong with this statement, I guess. Most likely, Phillip is testing on smth like x64 and this arch is le, so... With regards, Pavel Skripkin