* Fix data corruption bug
@ 2022-11-17 17:27 Ammar Faizi
2022-11-18 2:30 ` Alviro Iskandar Setiawan
0 siblings, 1 reply; 3+ messages in thread
From: Ammar Faizi @ 2022-11-17 17:27 UTC (permalink / raw)
To: Gilang Fachrezi
Cc: Alviro Iskandar Setiawan, Muhammad Rizki, Dina Maulina,
Kernel Team, Network Integration Team, GNU/Weeb Mailing List,
Ammar Faizi
From: Ammar Faizi <[email protected]>
Subject: [PATCH] chnet: Fix data corruption when a read operation is pending
When a read operation is pending, the @read_ret_ variable has to be
set to -EINPROGRESS. The current CHNetDelegate::__Read() method
doesn't do that. This situation makes the caller abort the request
because it doesn't see -EINPROGRESS from the return value. This bug
is reproducible on a slow network connection where __Read()
operation often results in a pending state.
This bug is introduced in commit 77bbcc903899 ("chnet: Completely
refactor again").
Fixes: 77bbcc903899 ("chnet: Completely refactor again")
Signed-off-by: Ammar Faizi <[email protected]>
---
chnet/chnet.cc | 9 +++++++--
1 file changed, 7 insertions(+), 2 deletions(-)
diff --git a/chnet/chnet.cc b/chnet/chnet.cc
index cd27810..fdbbdc6 100644
--- a/chnet/chnet.cc
+++ b/chnet/chnet.cc
@@ -713,9 +713,14 @@ int CHNetDelegate::__Read(int size)
read_buf_ = base::MakeRefCounted<net::IOBufferWithSize>(size);
ret = url_req_->Read(read_buf_.get(), size);
- read_ret_.store(ret, std::memory_order_release);
- if (ret != net::ERR_IO_PENDING && callback_.on_read_completed_)
+ if (ret == net::ERR_IO_PENDING) {
+ read_ret_.store(-EINPROGRESS, std::memory_order_release);
+ return ret;
+ }
+
+ read_ret_.store(ret, std::memory_order_release);
+ if (callback_.on_read_completed_)
std::move(callback_.on_read_completed_)(url_req_.get(), ret);
return ret;
--
Ammar Faizi
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: Fix data corruption bug
2022-11-17 17:27 Fix data corruption bug Ammar Faizi
@ 2022-11-18 2:30 ` Alviro Iskandar Setiawan
2022-11-18 16:09 ` Kyoko Senpai
0 siblings, 1 reply; 3+ messages in thread
From: Alviro Iskandar Setiawan @ 2022-11-18 2:30 UTC (permalink / raw)
To: Ammar Faizi
Cc: Gilang Fachrezi, Muhammad Rizki, Dina Maulina, Kernel Team,
Network Integration Team, GNU/Weeb Mailing List
On Fri, Nov 18, 2022 at 12:27 AM Ammar Faizi wrote:
> From: Ammar Faizi <[email protected]>
> Subject: [PATCH] chnet: Fix data corruption when a read operation is pending
>
> When a read operation is pending, the @read_ret_ variable has to be
> set to -EINPROGRESS. The current CHNetDelegate::__Read() method
> doesn't do that. This situation makes the caller abort the request
> because it doesn't see -EINPROGRESS from the return value. This bug
> is reproducible on a slow network connection where __Read()
> operation often results in a pending state.
>
> This bug is introduced in commit 77bbcc903899 ("chnet: Completely
> refactor again").
>
> Fixes: 77bbcc903899 ("chnet: Completely refactor again")
> Signed-off-by: Ammar Faizi <[email protected]>
Reviewed-by: Alviro Iskandar Setiawan <[email protected]>
tq
-- Viro
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: Fix data corruption bug
2022-11-18 2:30 ` Alviro Iskandar Setiawan
@ 2022-11-18 16:09 ` Kyoko Senpai
0 siblings, 0 replies; 3+ messages in thread
From: Kyoko Senpai @ 2022-11-18 16:09 UTC (permalink / raw)
To: Alviro Iskandar Setiawan
Cc: Ammar Faizi, Muhammad Rizki, Dina Maulina, GNU/Weeb Mailing List
On Fri, Nov 18, 2022 at 9:31 AM Alviro Iskandar Setiawan wrote:
> On Fri, Nov 18, 2022 at 12:27 AM Ammar Faizi wrote:
> > From: Ammar Faizi <[email protected]>
> > Subject: [PATCH] chnet: Fix data corruption when a read operation is pending
> >
> > When a read operation is pending, the @read_ret_ variable has to be
> > set to -EINPROGRESS. The current CHNetDelegate::__Read() method
> > doesn't do that. This situation makes the caller abort the request
> > because it doesn't see -EINPROGRESS from the return value. This bug
> > is reproducible on a slow network connection where __Read()
> > operation often results in a pending state.
> >
> > This bug is introduced in commit 77bbcc903899 ("chnet: Completely
> > refactor again").
> >
> > Fixes: 77bbcc903899 ("chnet: Completely refactor again")
> > Signed-off-by: Ammar Faizi <[email protected]>
>
> Reviewed-by: Alviro Iskandar Setiawan <[email protected]>
Reviewed-by: Gilang Fachrezy <[email protected]>
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2022-11-18 16:09 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-11-17 17:27 Fix data corruption bug Ammar Faizi
2022-11-18 2:30 ` Alviro Iskandar Setiawan
2022-11-18 16:09 ` Kyoko Senpai
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox